1、WAVE V|2023DATA-POWERED INNOVATION REVIEW FOREWORDN I R A J PA R I H A RCEO,Insights&Data,CapgeminiBeing caught between a rock and a hard place is usually not seen as a distinct pleasure.But I feel quite different,very positive vibes when looking at the visual theme of this fifth edition of our Data

2、-powered Innovation Review.It somehow unleashes the explorer in me,looking for new,unchartered ways to innovate with data.The trail towards innovation can be intricate and mysterious.The still mountains hold thousands of years of knowledge,seemingly rock-solid and immovable yet pounded by ever-chang

3、ing conditions.They reflect sustainability like nothing else.And within,there is a strong core and an innovative layer that adapts,changes,and creates new evolutions.Navigating this time-tested environment,finding new paths towards unknown destinations is an exciting challenge.Thats why I am happy t

4、o invite you to join us in our journey of elevating data&AI and serving sustainability with the latest release of our innovation review magazine.Both main topics of this issue carried by a great collection of articles are near to my heart.Becoming a true data master is a matter of leveraging data to

5、 the full extent,activating it as close as possible to business operations,building more value on top of proper data foundations and uplifted data behaviors.It is about looking beyond the usual surroundings and be inspired by new ways to elevate data&AI.Just see the articles about serendipity,data m

6、esh,human augmentation and even playing the data poker game to get a flavor of it.Then there is the quest for sustainability and a better society higher on the corporate agenda than ever before.Data plays an absolutely pivotal role in addressing some of the biggest challenges of our time.A variety o

7、f articles in this magazine such as the ones about data and AI for Net Zero,the circular economy,battling river blindness,feeding the world,and the UNs 17 sustainable development goals are all testament to this.Lets go and explore.I sincerely hope you find this edition of the Data-powered Innovation

8、 Review rocks like always.Enjoy it!2Data-powered Innovation Review|Wave V 2022 Capgemini EDITORS NOTERO N TO L I D OCTO,Insights&Data,CapgeminiTo Capgeminis data-powered innovation movers&shakers this good-looking magazine is not the 42 answer to all questions of life.But it sure is a great catalyst

9、 to come together and brainstorm innovative topics,reach out to relevant top experts and technology partners,create unique content and then after publication surf a long wave of follow-up posts and events.Now in its 5th edition,the Data-powered Innovation Review has supplied a steadily growing repos

10、itory of innovative ideas and best practices,all geared around the breakthrough business potential of data,analytics,and AI.Looking at 2 years of exploring data-powered innovation,we see certain topics rapidly entering mainstream,sometimes at blistering speed.Just look at generative and creative AI

11、right from the first edition of this magazine identified as a likely disruptive technology trend.It now,in just a few months,has become commonly available and used through the likes of chatGPT,Dall-e,Stable Diffusion,and Midjourney.Also,an early trend such as data mesh has rapidly grown from early,e

12、xploratory posts to dozens of implementations we are currently involved in across the globe.And what will be next?We have 19 articles for you to get a taste of it.In the first section of the magazine Data&AI Elevated we bring yet again inspiring perspectives on how to activate data for on-the-ground

13、,real-life business purposes.From innovative ways to rapidly start benefiting from cloud and edge computing,via best practices and use cases in the energy&utilities sector,all the way up to serendipity and augmentation as the drivers to next-generation user experiences.As we are entering 2023,there

14、are even some new years resolutions for you.And what exactly has playing poker to do with becoming a data master?Whatever economic,technological,and geopolitical scenarios pan out in the new year;we know that sustainability tops the agenda of most organizations.Hence the second section of the public

15、ation Sustainability Served brings a flurry of innovative ideas on how data,analytics and AI can be used to create a better planet;and not only in terms of the environment,but also in terms of battling diseases,feeding the world and even addressing the full scope of the United Nations 17 sustainable

16、 development goals.Doing good things with data turns out to be a journey with benefits:building data mastery along the way sure makes organizations doing well themselves too.The articles have been crafted by leading Capgemini experts,a bit of generative AI and in collaboration with key partners such

17、 as AWS,Denodo,Databricks and Dataiku.Dont hesitate to contact the contributors.Theyll be most happy to help you on your innovation journey.Also,look out for a series of follow-up activities,such as additional in-depth articles and live events featuring contributors and guest panelists.3Data-powered

18、 Innovation Review|Wave V 2022 Capgemini 06Honey,I shrunk my cloud from cloud native to edge nativeNedyalka Delistoyanova&Arne Romann,Capgemini11Human augmentation the next frontier in AI Rajashree Das&Vrushali Joshi,Capgemini17Serendipity systems:Building world class personalization teams Neerav Vy

19、as&Chloe Cheau,Capgemini22Three strategies for building an adaptable data architecture Inessa Gerber&Felix Liao,Denodo26Your 2023 data strategy in four resolutionsSabina Shaikh,Databricks29Fly faster into the cloudAurobindo Saha,Capgemini34Everyday AI:next-generation self-service analyticsBridget Sh

20、ea,Dataiku38The revitalization of data meshBeverley Coy&Neil Parker,Capgemini43Play data like pokerShi Kai,Capgemini48Emerging technologies are transforming the energy and utitilies sectorIsabelle Tachet&Caroline Ball,Capgemini52Putting AI in sustainability Tijana Nikolic&Robert Engels,Capgemini57Fi

21、nding the code for a cureAnne Laure Thibauld,Capgemini&Mike Miller,AWS61Creating a circular economythrough AI Faizan Pratyasha,Soumitra Upadhyay&Pratyasha Shishodia,frog66Data sharing is caring.Or is it a waste of investment?Daniela Rittmeier,Capgemini71Three data-foundation strategies for Scope 3Ro

22、osa Sntti,Capgemini76AI roundtable on sustainabilityRajeswaran Viswanathan,Aishwarya Jagtap,&Karan Kajrolkar,Capgemini80The paradigm shift of enterprise sustainabilityUmamaheswari Kathirvel&Yashowardhan Sowale,Capgemini84Feeding the world better with Project ENHANCEMaartje de Laat&Marijn Markus,Capg

23、emini88Sust-AI-UN-ability:AI meets UN Sustainable Development GoalsNiharika Kalvagunta,CapgeminiDATA&AI ELEVATEDSUSTAINABILITY SERVEDData-powered Innovation Review|Wave V 2022 CapgeminiDATA&AI ELEVATED06Honey,I shrunk my cloud from cloud native to edge nativeNedyalka Delistoyanova&Arne Romann,Capgem

24、ini11Human augmentation the next frontier in AI Rajashree Das&Vrushali Joshi,Capgemini17Serendipity systems:Building world class personalization teams Neerav Vyas&Chloe Cheau,Capgemini22Three strategies for building an adaptable data architecture Inessa Gerber&Felix Liao,Denodo26Your 2023 data strat

25、egy in four resolutionsSabina Shaikh,Databricks29Fly faster into the cloudAurobindo Saha,Capgemini34Everyday AI:next-generation self-service analyticsBridget Shea,Dataiku38The revitalization of data meshBeverley Coy&Neil Parker,Capgemini43Play data like pokerShi Kai,Capgemini48Emerging technologies

26、are transforming the energy and utitilies sectorIsabelle Tachet&Caroline Ball,Capgemini5Data-powered Innovation Review|Wave V 2022 CapgeminiHONEY,I SHRUNK MY CLOUD FROM CLOUD-NATIVE TO EDGE-NATIVE The emergence of IoT devices,autonomous vehicles,wearables and AR/VR devices and the like,plus the read

27、y accessibility of 5G networks,brings a new reality for the modern data platform architecture.And there,bigger is neither better nor faster.A new ecosystem of analytics and computational platforms is born on the edge.N E DYA L K A D E L I S TOYA N OVA Managing Solution Architect,CapgeminiA R N E ROS

28、 S M A N NHead of AI&Data Engineering Germany,Capgemini6Data-powered Innovation Review|Wave V 2022 Capgemini“Once these two cars make their way around the track and Lando Norris and Daniel Ricciardo do it at 200 miles an hour 300 sensors are producing a terabyte-and-a-half worth of information that

29、we have to analyze to try and find the edge.And when I talk about edge,its milliseconds.”How shall we process and react to 1.5 terabytes of data within milliseconds?It is nearly impossible with the usual way of processing,by sending the data into a cloud or on-premises data center,processing it,and

30、sending a result or an action back to the origin.This gets even more complicated when there is no possibility of a proper network connection,or if the surrounding infrastructure or the bandwidth is just too small.What is edge computing?Edge computing is an emerging computing framework that relies on

31、 the distribution of resource-optimized data storage and processing capabilities closer to the point of creation and/or consumption of this data.As opposed to traditional cloud computing frameworks,where data processing is centralized into a remote datacenter,edge computing seeks to relocate activit

32、ies closer to the user or the end device(the edge of the network).Edge computing is a necessary enabler for an increasing variety of critical use cases,ranging from telemedicine to autonomous vehicles and industrial IoT.To sum up the definition:we are able to do more with the data at the place the d

33、ata is being generated and where we need to take the action.That has not only a positive impact on our reaction time,but also allows data-led decisions to be made by autonomous devices.Figure 1:Examples of Edge Computing7Data-powered Innovation Review|Wave V 2022 Capgemini What are the benefits of e

34、dge computing?No latency.Local processing means that data does not need to be sent to a centralized cloud for analyses.AI models deployed on the edge can deliver interpretation of the data on the spot,providing instant decision support.Lower data volume.Analytics capabilities at the source limit the

35、 need for the data to be sent and processed centrally.Only a subset on meaningful point can be communicated and stored centrally.Edge devices can ingest and process data on the spot,determining what data can be discarded and what data is valuable and should be sent to the data centers.Optional conne

36、ctivity.By moving the storage and processing of the data to the edge device,the requirement for stable network connection is significantly reduced.This reduces the bandwidth costs and network traffic,as the devices can store data and results locally,synchronizing with centralized clouds only on dema

37、nd.Strengthened privacy and security.Local processing eliminates the need for private,sensitive data to be sent over the internet to a remote location to perform matching and recognition,where it could potentially be hacked or exposed.As a result,edge AI brings high-performance computing capabilitie

38、s to the edge,where sensors and IoT devices are located.This enables safe real-time analytics and autonomous decision making by creating insights almost instantly through running machine learning(ML)and deep learning(DL)algorithms.The volume of data and the need for higher frequency and faster respo

39、nse to the changing environment raises the topic of edge computing and edge devices and pushes the AI and analytics capabilities out of the cloud and into the edge.Edge computing solutions in industrial settings are predicted to grow rapidly,driven by the need for real-time data analytics and locali

40、zed action.8Data-powered Innovation Review|Wave V 2022 Capgemini Whats in it for us?Industrial organizations view edge computing as key to realizing the full potential of 5G.In 2021,the Capgemini Research Institute found that“close to two thirds(64%)of organizations plan to adopt edge computing serv

41、ices within three years,while a small share(7%)of organizations are already using 5G-based edge computing services.”Figure 2:Edge Computing Infrastructure StackThere are already a multitude of technical solutions and offerings from Amazon and Microsoft,and options from niche suppliers are appearing

42、on the market.Eneco,a producer and supplier of natural gas,electricity,and heating in the Netherlands,collaborated with Microsoft and used Azure IoT components to build a smart grid solution to communicate with the numerous energy meters located at households and industrial units,eliminating the nee

43、d for manual inspection.The company achieved a 10-degree temperature reduction in its heat distribution network and lowered CO emissions.But these innovations happen in other sectors as well,such as consumer business.Apple announced crash detection in its recent smart watch,extending its already exi

44、sting fall detection towards car crashes.It can detect a severe car crash and automatically initiate an emergency phone call.For this,a new three-axis gyroscope and“high g-force”accelerometer that samples up to 256 Gs of force are used in combination with advanced AI models.With the rising adoption

45、of edge AI solution like TinyML,new intelligent solutions are created.An important application of such technology is found in the area of farming,where IoT data and AI models are used to lower the consumption of water and energy while saving operational costs.In another sector,oil and gas,continuous

46、 improvement in miniaturization and AI models is preventing pipeline spills,as these can be autonomously monitored.CLOUD COMPUTING STACKloT&Data Analytics PlatformEDGE COMPUTING STACKloT GatewayBackhaul NetworkCommunicationFeedback LoopFieldNetworkFieldNetworkCamera FeedLow range networksField Netwo

47、rkEmbeddedNetwork DevicesSensorsIn-VehicleSensorCameraPLCCustomApplicationsAPIDevelopmentToolsDataProcessingMachineLearningData LakeApplicationLife CycleManagementHardware LifeCycleManagementSecurityModuleDevice driversData ProcessingEngineDatabaseLife CycleManagementSecurityModule9Data-powered Inno

48、vation Review|Wave V 2022 Capgemini Innovation takeawaysR E A L T I M E W I T H EDG EEdge computing is a scaled down and distributed approach to cloud computing systems bringing the ability to capture insights from real-time data,without impacts from latency and network bandwidth.R E S O U RC E EFFI

49、 C I EN C YEdge resources afford technical advantages over cloud-only processing and greatly improves scalability of systems,while optimizing network efficiency reduces the amount of round-trips to the datacenter.AC T N OW Its now the right time to increase efforts in building edge computing solutio

50、ns,leveraging the latest technologies to increase system resiliency and efficiency.The emergence of IoT devices,autonomous vehicles,wearables and AR/VR devices and the like,and the ready accessibility of 5G networks,brings data with frequency and volume never seen before.The computing capabilities o

51、f the cloud data platforms are not enough to sustain instantaneous storage,processing,and analysis of the results.The volume of data and the need for higher frequency and faster response to the changing environment raises the topic of edge computing and edge devices and pushes the AI and analytics c

52、apabilities out of the cloud and into the edge.Edge computing solutions in industrial settings are predicted to grow rapidly,driven by the need for real-time data analytics and localized action.This unparallel increase in IoT devices and edge AI capabilities brings new challenges for data platform a

53、rchitecture and capabilities.We see a shift in technology and architecture moving away from monolithic to product and domain thinking,and to distributed edge,enabling autonomous decision-making.#E D G ECO M P U T I N G#E D G E A I#A I 4 G O O D#DATA P O W E R E D#I N N OVAT I O N10Data-powered Innov

54、ation Review|Wave V 2022 Capgemini HUMAN AUGMENTATION THE NEXT FRONTIER IN AI Magic happens when humans and machines work collaboratively in a truly symbiotic relationship.They enhance and augment each others capabilities to bring speed,ethical behavior,creativity,disruption,and innovation to how we

55、 do business.It is where AI goes next.Some examples illustrate this key evolution.R A JA S H R E E DA SMaster Architect,Insights&Data,CapgeminiV RU S H A L I J OS H IEnterprise Architect,Insights&Data,Capgemini11Data-powered Innovation Review|Wave V 2022 CapgeminiAI is a transformational technology

56、that helps to imitate the way human minds think and automate various complex,cumbersome,and banal tasks with precision and at a speed.If synergies are combined and AI is learning from human intervention and historical knowledge,it will completely disrupt the ways AI can transform the business,within

57、 constraints of humanity.Cooperation between human and AI,where human cognitive capabilities are enhanced by artificial intelligence technology with sensory or actuating technology,is termed“human augmentation.”Further technologies driving growth in human augmentation are:5G,edge computing,quantum c

58、omputing,IoT precision sensors,augmented reality,and so on.Many industries are already leveraging this human and machine cooperation in sensory augmentation and augmented cognition,or in augmented action like prosthetics or exoskeletons uses in healthcare,gesture tracking,industrial applications,and

59、 3D bioprinting.There are ethical,people,process,and technical considerations though,when implementing this human and AI cooperation.Policy changes and regulatory adherence are needed to build an ecosystem in safe,accountable,ethical,and sustainable ways to embrace AI alongside humans.A massive amou

60、nt of trusted data is required to model data and gain inferences,with the need to extract it from diverse,possibly siloed data sources.Delicate concerns need to be addressed in areas such as data privacy and reliability when sharing human personal data,to avoid negative corporate exposure or even fa

61、talities.Technology solutions need to be selected and designed for the number of contact points and interfaces between humans and AI,for easy handshakes,scalability,and affordability.Market perspectiveMarket studies reveal that the human augmentation market is estimated to be worth more billion by 2

62、026,at a CAGR of 21 percent.Growing demand for wearable human augmentation devices in the healthcare sector,the recent advent of AI-powered wearable devices,and rapid technological advancements such as the availability of fast processors,miniaturization of sensors,use of brain-computer interfaces,an

63、d the evolution of nanotechnology are some of the major factors contributing to the market growth.Key players operating in the human augmentation market are Google,Samsung,Ekso Bionics Holdings,Vuzix,Garmin,Fossil Group,B-Temia,Casio,Magic Leap,Rewalk Robotics,Polar Electro,and P&S Mechanics.Human a

64、ugmentation technology has seen a significant influence especially in healthcare in terms of COVID 19 impact.Also,manufacturing,military,and defense industries are heavily investing in adopting human augmentation services.12Data-powered Innovation Review|Wave V 2022 Capgemini SolutionsLets see how h

65、uman augmentation is making a difference in many fields.Human-in-the-loop reinforcement learning in the banking sectorOne of the leading banks has automated and optimized its bank guarantee process via AI-driven automation with human intervention.The success of it has led to enabling many more busin

66、ess functions like home loan,assessment,and document redaction.Reinforcement learning in bankingSolution highlights:The AI/ML components included the classification of the incoming document,extracting requisite information from the agreement advice,and storing it for further usage.Since the system w

67、as new,it generated confidence in the system via a human-in-loop interface that enabled a maker to validate all the information extracted through an interactive screen and make any changes or updates in case there was an error.Any changes or updates made by the maker are captured again and fed back

68、to the system for model retraining to ensure such scenarios are handled adequately in the future.TIFFXMLPDFXMLOCRJSONJSONJSONABCPortal FormEmail requestRequestreportStorageOCR outputDocument in PDFimages formatFiles picked forprocessingAbbyy,DatacapDocumentClassifierXML file form Agreement AdviceReq

69、uestFormClassifierRequestForm IEPurpose&CommodityType ClassifierFormCompletionCheckAgreement Advice IEIE outputJSON to DBLearning Feedback/Reinforcement LearningUpdated JSON andRecord into DBJSON input toMakerScreenWITCommentsMaker ScreenUI-ComponentsAI ML-ComponentsFile Processing Component.xml fil

70、eof RF Indexing Signature Verification Duplicate Check13Data-powered Innovation Review|Wave V 2022 Capgemini Augmented AI for intelligent manufacturingThe combined efforts of intelligent machines and human beings co-bots as they are called are bringing efficiency and value to the manufacturing lifec

71、ycle,meeting the goal of being competitive in market.AI/Human Augmented FactorySolution highlights:AI-enabled humans in analyzing the patterns in data.In a manufacturing scenario,the data patterns in temperature,vibrations,pressure,sound,vision,and electrical parameters like voltage,current,power et

72、c.,play an important role in controlling the operational processes,maintaining assets,producing quality products,and ensuring safety of assets and personnel.The factory robots and equipment bring in many operational efficiencies.AI enabled by IoT data makes robots smart.Such smart machines are conte

73、xt sensitive and can take controlling actions in the fast-paced manufacturing,hazardous process manufacturing,and extreme operating conditions.Techniques include computer vision-based visual inspection,process monitoring and worker safety,machine learning-based yield optimization,predictive maintena

74、nce,smart demand sensing,and inventory control.Human Operators assisted by smart robots to make manufacturing operations lean,automated and agile resulting into improved quality and productivityAI helping to make factory operations green and sustainableAI augumenting demand sensing and improving inv

75、entory controlINTELLIGENT AUTOMATIONREAL-TIME OPERATIONSMANAGEMENTDIGITAL FACTORYENERGY MANAGEMENTPREDICTIVE MAINTENNANCEFLOW SIMULATIONDIGITAL QUALITYENHANCED OPERATOR14Data-powered Innovation Review|Wave V 2022 Capgemini Conversational AI for enhanced customer experienceEvery business requires con

76、versations with its customers.AI can augment this.Conversational AI allows businesses to duplicate human-to-human interaction for human-to-machine conversations.These conversations are made possible via chatbots,voice bots,and virtual assistants,and hence can be referred to as“virtual humans.”Virtua

77、l humans can make conversations personalized and contextualized with humans.They will be able to create and remember experiences from their interactions with humans and will be able to respond to a wider range of situations.Digital human twins can interact and create customer-service responses like

78、resolving queries,completing transactions,and tracking orders in a quite natural way.“The simple truth is that companies can achieve the largest boosts in performance when humans and machines work together as allies,not adversaries,in order to take advantage of each others complementary strengths.”P

79、aul R.Daugherty,Human+Machine:Reimagining Work in the Age of AI15Data-powered Innovation Review|Wave V 2022 Capgemini Augmented and virtual reality in operationsWe are seeing increased use of AR/VR in wearable devices for internal industrial company operations,design,engineering,and field services.T

80、he key benefits are improved productivity and collaboration.Examples:Airline technicians use AR instructions for airplane wiring schematics in their field of view,allowing them to be hands-free.Fieldworkers use AR and VR glasses to see concealed utilities lines under the streets in real-time.In auto

81、motive settings,VR is used to capture human movement during equipment assembly through motion sensors to re-engineer movement to decrease risk of injury and increase productivity.We need to further exploit new ways of embracing AI and human hybrid intelligence to work together to bring in more disru

82、ptive advancements in various industries.AI is considered a“self-learning”algorithm,so let us make it more intelligent by providing good quality data from human intelligence and experience over time to make better decisions and products.There are other technologies in the future with promises of big

83、 potential,such as genetic engineering and brain-computer interfaces.At the same time,the ethical,moral,and legal implications of human augmentation are complex and hard to comprehend.Early and regular engagement with these issues must be thoroughly considered.Innovation takeawaysCOO PER AT I O N I

84、S K E YHumans and AI can work together,as partners,to complement each others capabilities and achieve more,possibly with less.D I S RU P T I O N I S T H E P OT EN T I A LHuman and AI augmentation enables entirely new,unexplored ways of doing business,while keeping it safe,accountable,and ethical.AU

85、G M EN TAT I O N I S E X PA N D I N G AI and other technologies(such as AR and VR)create yet further evolving opportunities to grow human augmentation.#A I 4 G O O D#H U M A N AU G M E N TAT I O N#DATA P O W E R E D#R E S P O N S I B L E A I#E T H I C A L A I16Data-powered Innovation Review|Wave V 2

86、022 Capgemini SERENDIPITY SYSTEMS:BUILDING WORLD-CLASS PERSONALIZATION TEAMS The last best experience we have anywhere sets the bar for all experiences everywhere.Consumers dont want just personalization theyre demanding it.Delivering personalization is no longer bar-raising.Organizations need to mo

87、ve from providing personalization as a feature to delivering serendipitous experiences.The challenge then is serendipity at scale or obsolescence with haste.Without the right teams,organizations are speeding toward obsolescence.CH LO E CH E AUHead of Experience Engineering,Customer First,Insights&Da

88、ta North America,CapgeminiN E E R AV V YA SHead of Customer First,Insights&Data North America,Capgemini17Data-powered Innovation Review|Wave V 2022 CapgeminiGreat basketball teams and great personalization teams have a lot in common.Imagine a shopping experience thats completely generic.Worse than g

89、eneric,it goes out of its way to recommend things you dont want.It recommends actions that are the opposite of what youre looking to do.Its perfectly frustrating.How long will a business based on that sort of experience last?Now imagine a personalization experience that knows you so well its constan

90、tly providing you serendipitously delightful experiences.Youre discovering things you never knew you wanted.But youre never allowed to use it,because the experience never sees the light of day.The MVP never becomes an available product.Both scenarios are terrible.Unfortunately,a variation of the sec

91、ond is more common.Seventy-seven percent of AI and analytics projects struggle to gain adoption.Fewer than 10 percent of analytics and AI projects make an impact financially,because 87 percent of these fail to make it into production.What if we could flip the odds?What if rather than most recommenda

92、tion projects failing,most of them succeeded?Cross-functional,product-centric,teams can do just that.Its how innovators like Amazon and Netflix were able to succeed so quickly and so often in their personalization programs.Its also been critical for the dozens of successful personalization programs

93、weve delivered at Capgemini.18Data-powered Innovation Review|Wave V 2022 Capgemini Recommendation experiencesEverything is a recommendation.That insight came from Netflix:“the Starbucks secret is a smile when you get your latte,ours is that the website adapts to the individuals taste,”said Reed Hast

94、ings,co-founder of Netflix.Recommendations werent features or algorithms.They were the experience;the means to delight,surprise,frustrate,or anger customers.At Amazon,Jeff Bezos original goal was a store for every customer.This wasnt AI for the sake of AI.Both companies made personalization central

95、to their experiences,and personalization enabled Amazon and Netflixs visions for more innovative,delightful,and serendipitous experiences.Recommendation experiences(RX)were critical to customer experiences(CX).Experiences were the product.Building products is hard.Josh Peterson co-founded the P13N(p

96、ersonalization)team at Amazon.He described the early days of Amazon as challenging because the company was siloed.Design,editorial,and software engineering were fragmented.“It was really hard to ever get anything all the way out to the site without begging and borrowing people from silos.The one tim

97、e it was always different was when we did a product launch So if there was a big enough effort like launching music or auctions then you had permission to borrow everyone to put together your team.”In the early days of Amazon,there were many engineering efforts around personalization.Even though the

98、se efforts were led by brilliant engineers,they saw limited success.It wasnt until after the launch of Amazon Auctions that personalization made a real impact.After Amazon Auctions,Peterson and Greg Linden looked to make Bezos vision for a personalized store for every customer a reality.The goal was

99、 a team that could“own its whole space,”to break silos to create a cross-functional team to rapidly experiment and deliver.This was the first team,outside of the design organization,to have designers in their team embedded with web developers and technical project managers.This enabled a higher numb

100、er of launches compared to other teams.The impact of their model was so successful that it became the basis of Amazons famous“Two Pizza Team”approach essentially a team small enough that they could be fed with two pizzas.Small teams that were decentralized,autonomous,and were“owners”of the business

101、could move faster and launch more experiments.More experiments would enable them to have more successful innovations.19Data-powered Innovation Review|Wave V 2022 Capgemini ExperimentationSuccessful personalization teams foster a culture of experimentation.Creating a culture of experimentation requir

102、es diverse,multi-disciplinary teams.Below we show the various skillsets and domains that are required for modern personalization teams.The circles dont represent people,they represent skills.Great basketball teams and great personalization teams have a lot in common.In basketball,you need defense.Yo

103、u need offense,both close to the rim and from afar.You need diversity in skillsets.You could get lucky and find a unicorn but fielding multiple teams of unicorns is not practical.Creating a team of all-stars sounds good on paper,but there are plenty of examples where those super teams fail to live u

104、p to expectations.A team without a diverse set of skills is unlikely to be very successful,and almost certainly not great.Team roles for personalizationBelow are the types of roles that will typically be needed for Personalization Teams.Not all teams will need all of the roles.One resource may fulfi

105、ll multiple responsibilities of these roles.Domain ExpertiseProduct Management Analytics TranslationGrowth ArchitectureVisualizationData ScienceML EngineeringData EngineeringBack End Expertise Front End Expertise ANALYTICS EXPERTISE+DATA SCIENCE BIG DATA ENGINEERING DOMAIN&BUSINESSEXPERTISE DATA OPS

106、&VISUALIZATION APPLICATIONS DEVELOPMENT AND INTEGRATION 20Data-powered Innovation Review|Wave V 2022 Capgemini Innovation takeawaysD I V ER S I T Y L E A DS TO S PEEDSpeed leads to innovation.Diversity leads to innovation.End-to-end cross-functional teams with dedicated resources are more likely to

107、successfully implement personalization programs and innovate faster than their peersA C U LT U R E O F E X PER I M EN TAT I O N I S C R I T I C A LVelocity,variety,and volume of experiments are leading indicators of innovation.“Our success at Amazon is a function of how many experiments we do per ye

108、ar,per month,per week,per day.”Jeff BezosS PEED I S A COM PE T I T I V E A DVA N TAG E Testing and learning iteratively as well as being able to deploy quickly contribute to faster speed to market.“Companies rarely die from moving too fast,they frequently die from moving too slowly.”Reed Hastings#P

109、E R S O N A L I Z AT I O N#S E R E N D I P I TO U S P E R S O N A L I Z AT I O N#M AC H I N E L E A R N I N G#A R T I F I C I A L I N T E L L I G E N C E#I N N OVAT I O N#R E S P EC T F U L P E R S O N A L I Z AT I O N#DATA A S A P R O D U C T Experimentation requires blending creativity and data.Pr

110、actically,this becomes a blend of statistics,behavioral economics,psychology,marketing,and expertise in experience design.Small teams with most of the skills above are more likely to do end-to-end personalization well.No one person will have all the skills needed,but together theyll bring more exper

111、iments to the table.Early Amazon teams were engineering and data-science heavy.It wasnt until the addition of design,business expertise,and a product-centric approach that they were able to execute end-to-end and achieve Bezos vision.Velocity is a leading indicator.Successful personalization teams t

112、est many ideas.They break experiments into small chunks so no one failure is large enough to disrupt the business.They test and learn quickly.Testing a dozen ideas and refining them will be more efficient than trying to make one idea“perfect.”Our intuition on what is going to work is often wrong.Tes

113、ting many ideas allows the data and results to guide us,rather than intuition.This requires personalization teams to develop many ideas end-to-end quickly.What does the future hold?Cross-functional,product-centric teams are the beginning,not the end.Experimentation requires blending creativity and d

114、ata.Practically,this becomes a blend of statistics,behavioral economics,psychology,marketing,and expertise in experience design.These teams need to track which features drive results to understand what is working and what is not.The goal is to achieve consistent and reliable serendipity from persona

115、lization efforts.The obvious is not serendipitous.Experimentation is needed to discover that which is not obvious and that which drives business outcomes.Without that,we cant scale serendipity.21Data-powered Innovation Review|Wave V 2022 Capgemini THREE STRATEGIES FOR BUILDING AN ADAPTABLE DATA ARCH

116、ITECTUREThe COVID-19 pandemic taught us many things,and one important lesson is the importance of being adaptable.With the uncertainties that organizations have faced in the last two years and with more disruptions on the horizon,being adaptable to rapidly changing market conditions and consumer beh

117、avior will be critical for success in the future.An agile and adaptable data architecture is key to achieving this.I N E S S A G E R B E RDirector of product management Americas,DenodoFE L I X L I AODirector of product management APAC,Denodo22Data-powered Innovation Review|Wave V 2022 CapgeminiThe c

118、ompanies that adapted quickly and innovated based on new data and insights were able to capture emerging market opportunities during COVID.We saw this firsthand with many cloud-native,high-tech organizations that thrived due to their ability to quickly reimagine specific parts of their business usin

119、g new data and insights.When done right,a more agile and adaptable data architecture can increase resilience and accelerate the path to new insights and innovations.Being adaptable means being able to quickly change and evolve to suit different conditions.An adaptable data architecture is a flexible

120、 foundation that can be modified and extended quickly.It is about enabling more rapid delivery of new capabilities and simplifying existing architectural approaches.But data gravity and legacy systems can be significant hurdles when it comes to modernizing any data infrastructure.Here we offer three

121、 strategies that can help an organization on this journey to build a more adaptable and resilient data architecture.1.Decoupling data access from data sourcesThe best way to gain agility and adaptability is to minimize hard constraints and technology dependencies.The shift to objective-based file st

122、orage is the best example of the power of decoupling.Cloud-based object storage solutions such as S3 decouple file access from storage and bring tremendous agility and flexibility,and have formed the foundation of numerous innovations in cloud technology.With an increasingly complex data landscape,d

123、ecoupling data access from data storage can bring the same types of agility and empower the data consumer to do more with data.By connecting data consumers and data sources through a logical layer,we bring additional agility and flexibility.The underlying data repository can be added and modified qu

124、ickly without impacting the end-data consumer.By using data virtualization techniques and presenting a semantic access layer,it is also easier and quicker for the data consumer to connect to and access multiple data sources without needing to know the where and what of the underlying technologies th

125、at are evolving all the time.Data access de-coupled from data source and connected via a data-virtualization layerTraditionalDB&DWSDATA OURCESLOGICAL DATA FABRICDATA CONSUMERSBI ToolsData Science ToolsCloud StoresHadoop&NoSQLFilesAppsSemantic Layer Powered by Data Virtualization StreamingSaaSSaaSR:(

126、).23Data-powered Innovation Review|Wave V 2022 Capgemini 2.Focus on agile data provisioningBeing able to react and adapt to market environments that are changing constantly requires faster and more agile access to data.While lengthy batch ETL and warehousing processes were the norms and acceptable b

127、efore COVID,rapid delivery of real or near real-time data quickly became the gold standard for data-driven decisions during and post peak COVID.By focusing on connecting data instead of just collecting and replicating data,you empower data consumers and applications to quickly tap into all the data

128、within the enterprise,regardless of where they sit.When you focus on an agile approach to data provisioning that is flexible and supports multiple integration styles(batch/real-time),you reduce time to data,time to insight,and,ultimately,time to market.Technologies such as data virtualization today

129、can provide just the connectivity layer that accelerates data provisioning and minimizes lengthy ETL and data pipeline hops.Not only can data virtualization accelerate data provisioning for batch-based analytical requirements,but it can also be used to support real-time,operational requirements toda

130、y.One Australia-based educational organization,for example,leveraged data virtualization to connect to multiple data sources for COVID contact tracing reporting requirements.By leveraging data virtualization,it not only decoupled data delivery and insights across its extensive data ecosystem but als

131、o provided a unified data access layer that surfaced data in real time.The adaptable nature of the data architecture means it could integrate more than 50-plus complex applications in less than two weeks for the contact tracing requirements,something the team there claims would have been impossible

132、without the use of data virtualization.3.Data-as-a-ProductProvisioning data as a product means that data is easily discoverable,understandable,and reusable by everyone.Making data access modular and easily reusable increases adaptability and agility and drives collaboration across the whole organiza

133、tion.An excellent example of this is the increasing adoption of data services based on REST or GraphQL.By focusing on the needs of application developers and making data access modular and API-driven,you enable seamless collaboration between the data teams and applications teams for data-driven inno

134、vation and data-led products.By leveraging product thinking and focusing on the needs of different data consumers,you unleash the value of data to a much broader team.When the concept of Data-as-a-Product is combined with a data marketplace strategy,you further accelerate data adoption and self-serv

135、ice.Anyone in the organization should be able to visit a data marketplace and search for and find the data product they need in the form they want.They should then be able to derive new insights or build a new product with minimal help from a centralized IT or data team.This product-centric,self-ser

136、vice oriented approach to data is very much possible today and can lead to a truly data-powered culture.24Data-powered Innovation Review|Wave V 2022 Capgemini Adapt and innovateWith the data and technology landscape evolving rapidly today,the traditional approach of building long-term data architect

137、ures can be both risky and inefficient.Therefore,the key is creating an adaptable data architecture that allows business leaders to rapidly evaluate emerging business requirements and quickly adopt new data and analytics technologies.The successful companies of the future will not only have to innov

138、ate but innovate faster in highly disruptive and uncertain environments.An adaptable data architecture based on the principles of data decoupling,agile data provisioning,and Data-as-a-Product will therefore be a critical component of any agile and insight-led organization of the future.Innovation ta

139、keawaysB EI N G A DA P TA B L EHaving an adaptable data architecture can both increase resilience and drive innovation.DATA V I R T UA L IZ AT I O N EN AG L E S I N S I G H T SData virtualization techniques can be used to decouple data access from data sources and integrate distributed assets across

140、 the enterpriseCO N N EC T I N G OV ER CO L L EC T I N GBy focusing on connecting data instead of just collecting and replicating data,you empower data consumers and applications to quickly tap into all the data within the enterprise regardless of where they sit.T H I N K PRO DU C TData-as-a-Product

141、 combined with a data marketplace strategy can accelerate self-service and data democratization.#A DA P TA B L E#DATA D ECO U P L I N G#DATAV I R T UA L I Z AT I O N#DATA A S A P R O D U C T#AG I L E DATA P R OV I S I O N I N G“A more agile and adaptable data architecture can increase resilience and

142、 accelerate the path to new insights and innovations.”25Data-powered Innovation Review|Wave V 2022 Capgemini YOUR 2023 DATA STRATEGY IN FOUR RESOLUTIONSAs the year winds down,this is a good time to assess personal resolutions you have for the new year and,as a data leader,its also an opportunity to

143、take a fresh look at your data and AI strategy.Following a volatile year in the market,you can get ahead of your 2023 plans and see where your organization can improve processes,bring on new tools,and set goals that make sense for your team.S A B I N A S H A I K HVP,Global System Integrators,Databri

144、cks26Data-powered Innovation Review|Wave V 2022 CapgeminiFor the second year in a row,Databricks recently partnered with MIT Technology Review Insights to survey 600 CIOs,CTOs,and CDOs from large enterprises.The key result:CxOs and boards recognize that their organizations ability to generate action

145、able insights from data,often in real-time,is of the highest strategic importance.All respondents agreed that companies must view AI adoption as mission-critical in order to succeed.But without effective data strategies,businesses miss massive opportunities to better understand their customers,offer

146、 high-value products,and streamline operations.With such a significant link between effective AI strategies and strong data,not using the right AI tools or neglecting to leverage AI in the most effective ways can foil even the best-laid data plans.Here are four resolutions to make your data strategy

147、 pay off this year.1.Reassess your data architectureMost executives(72 percent)say that data,both fragmented and with poor quality,is likely to be the biggest issue when aspiring to achieve AI goals.The only way to better prepare for these challenges is to invest in a flexible data and computing arc

148、hitecture,like a lakehouse,that embraces open standards and can scale to meet the changing needs of the business.By creating a lakehouse,a company gives every employee the ability to access and employ data and artificial intelligence to make better business decisions.Many organizations that implemen

149、t a lakehouse as their key data strategy are seeing lightning-speed data insights with horizontally scalable data-engineering pipelines.Walgreens specifically shared that a lakehouse enabled smarter algorithms and generated new types of reporting that help people understand the supply chain and stor

150、e labor and productivity,patient vaccine scheduling,and prescription pickup processes.2.Build your tech stack in the multi-cloudMany data and technology leaders believe its not enough to think about the cloud in the singular sense instead,they think about building a multi-cloud environment.As the ad

151、option of cloud-based technology grows,many look for solutions that can move across major clouds(such as those from AWS,Azure,and Google Cloud).In our survey,78 percent of executives agreed that a multi-cloud approach ensures the most flexible foundation possible for AI development.It offers organiz

152、ations easy integrations when bringing on new solutions or businesses that use other cloud providers,flexibility to run workloads anywhere,and the assurance that they can comply with regulations down the road.Organizations that adopt a multi-cloud approach can also create new revenue opportunities a

153、nd enhance customer experiences.3.Invest in low/no codeLow-and no-code approaches are opening new pathways to innovation and lowering the barrier to entry for people who want to get quick insights from their data.Given how competitive it is to find the right tech talent in todays hiring market,low-a

154、nd no-code tools are key to relieve some of the pressure on data teams,empowering less technical teams to build models even with just a basic understanding of machine learning.No-code platforms make it possible to leverage AI without hiring expensive developers and data scientists,which means smalle

155、r businesses can more easily harness its power.Columbia Sportwear embraced this resolution and has seen more business units using the platform in a self-service manner that was not possible before.This has sped up the time to insights for all groups.27Data-powered Innovation Review|Wave V 2022 Capge

156、mini 4.Embrace open-source AI and open standardsOpen-source data lakehouses are quickly becoming the standard for how the most innovative companies handle their data and AI.It prevents teams from building tricky solutions in-house from scratch,which eats up resources.In our survey,50 percent of resp

157、ondents said that open-source standards and open-data formats were at the top of their dream tech stack list.Open source usually comes at little to no cost and,more importantly,its tried and true its a community effort,and solutions have been adopted and vetted by many,which equates to fewer headach

158、es for your IT team down the road.On top of this,the commitment to open data standards fuels the open-source community,which helps create a large talent pool of data experts who are better equipped to move use cases to production.Its easy to get overwhelmed with your new years resolutions-especially

159、 the hard ones.Organizations must continue to stay inspired when they think about their data strategy.Our survey showed that a smart data strategy will ultimately provide better data value.Stay the course to be prepared for a happy new year!Innovation takeawaysR E A S S E S S YO U R DATA A RC H I T

160、EC T U R EInvest in a flexible data and computing architecture,like a lakehouse.B U I L D YO U R T EC H S TAC K I N T H E M U LT I-C LO U DA multi-cloud approach ensures the most flexible possible foundation for AI development.I N V E S T I N LOW/N O CO D E Leverage AI without hiring expensive devel

161、opers and data scientists.E M B R AC E O PEN-S O U RC E A I A N D O PEN S TA N DA R DS Take advantage of an open-source community and talent pool.#DATA R E S O L U T I O N S#L A K E H O U S E#DATA#O P E N S O U R C EWithout effective data strategies,businesses miss massive opportunities to better un

162、derstand their customers,offer high-value products,and streamline operations28Data-powered Innovation Review|Wave V 2022 Capgemini FLY FASTER TOWARDS THE CLOUD Organizations seeking to modernize their AI and data estate on the cloud struggle with many challenges.A solution is necessary to understand

163、 the friction in the modernization journey,so businesses achieve time to value.Modern accelerators can smooth the experience,helping organizations realize the value of cloud modernization faster and at a lower cost.AU RO B I N DO S A H AGlobal Data Engineering&Cloud Architect Leader,Insights&Data,Ca

164、pgemini29Data-powered Innovation Review|Wave V 2022 CapgeminiWhether an organization is looking to unleash the power of its data estate or liberate the full potential of its business through data and AI,faster realization of the value of modernization is necessary.This is because market demands are

165、highly dynamic,and may change even before realizing the benefits of IT investment.Modernization is also a cultural change for an organization to become data powered.But the benefits are clear in terms of overall performance and financial parameters,such as revenue generation and profitability.Data m

166、astery and the cloudCompanies which are data masters deliver an advantage of between 30 to 90 percent in various metrics across customer engagement,top-line benefits,operational efficiency,and cost savings.There is 70 percent higher revenue per employee,and they are 22 percent more profitable than a

167、verage organizations.Today,about 39 percent of such organizations are turning data-powered insights into a sustained competitive advantage.Data masters manage data as a strategic asset that generates additional revenues by providing new intelligent services and products,resulting in better customer

168、engagement with personalized marketing.Insights are embedded into current processes,products,and services to make faster and better decisions to provide high value.Many organizations are still struggling to modernize and migrate their existing data platform,particularly towards the cloud let alone a

169、ddress pressing business needswith the right agility through data and AI platforms.The technology of today often is unable to meet the agility and innovation required by the business and the varying customer demands of today and tomorrow.There is a need for a framework of tools and processes that ca

170、n:Empower the end users to make better,faster,more collaborative decisions Enable business and technology collaboration through data democratization,while maintaining the highest level of security,compliance,ethicality,and trust Prime the organization for data activation and data ecosystems with a f

171、lexible,modular approach.Reasons to transform the organizationDisparateworlds of bigdata&operationalanalyticsLack ofbusinessautonomyExcessivetime tomarketPoor dataquality,privacy&securityDataGovernanceInhibitingInnovationIncreasingTCO and ITdebtData silos&lack offederatedmasterdataBusinessimperative

172、 to transformthe data fabricof theorganization30Data-powered Innovation Review|Wave V 2022 Capgemini Accelerating the journey with IDEAEnter Capgeminis IDEA(Industrialized Data&AI Engineering Acceleration):a new capability that helps organizations turn their data sprawl into a true strategic asset.I

173、t doesnt just help organizations do things faster,but also differently,by modernizing.This combines production-ready,out-of-the-box applications and modules that are also fully customizable granting organizations speed,flexibility,and accelerations as they transform their data estate towards the clo

174、ud.The IDEA frameworks help organizations to discover their current data estate,find the data inventory,and analyze complexity in code at blazing speed,so the effort of months is reduced to hours.The Data Ingestion Factory provides a flexible,scalable architecture that federates data from many sourc

175、es to quickly deliver business insights.Cross-platform code conversions,and the ability to create parallelized data pipelines with low-code or no-code skills,enable easier adaption of the modern solution.The use of OpenAI APIs such as GPT3 helps with Natural Language Processing to ease self-service.

176、The ability to do data reconciliation on terabyte-scale data,track the data lineage,audit the data quality,and visualize them in real-time requires libraries such as vaex.io.It can do statistical operations on an N-dimensional grid of up to 1 billion objects per second.The semantics and knowledge gr

177、aph on collected metadata helps find interesting relationships about why a certain workload operation takes more time and money.The machine learning algorithms are federated and pushed down on edge devices so that privacy is preserved and distributed computes in local devices can work faster.The adv

178、ent of 5G makes this cloud-to-device communication faster.31Data-powered Innovation Review|Wave V 2022 Capgemini How do the accelerators work?The accelerators are built on some simple premises:Observe the pattern of repeated manual operations Extract the metadata associated with operations Automate

179、similar operations in the future.Accelerator Powered SolutionUI LayerEvolved Technical architectureto meet business agilitySourcesPlatform BuildSecurity,Data Trust and GovernanceIntegration LayerMetadataDatabaseOPERATIONAL METADATA32Data-powered Innovation Review|Wave V 2022 Capgemini This three-ste

180、p approach helped a UK retail company set up a cloud platform and deploy warehouses in 25 minutes through infra code automation and DevSecOps;this was originally estimated as a two-month effort.A North American healthcare device manufacturer used accelerators to quickly demonstrate the value achieve

181、d from data federation from ERPs,Hadoop,data warehouses,and other applications in two months.A pattern-based data ingestion factory can validate the data schema,bring better data observability,and manage complex pipeline orchestration.Accelerators work like a catalyst,without changing the chemistry

182、of the underlying architecture.With“light-touch”on operational metadata,it helps to accelerate the journey.This benefits from the faster adaption of modern technologies at a lower cost.This agility helps to beat the competition and achieve data mastery for facing current and future business competit

183、ion.Innovation takeawaysN EED TO B E N I M B L EModern business requires modern solutions to understand customers and requires addressing business change.Easy-to-use solutions need to scale to address this need.The accelerators require tailoring to fit the business ask and show more agility to custo

184、mize for specific needs.ACC EL ER ATO R S C A N WO R K A S A C ATA LY S TThis fast-tracks the modernization journey by minimizing repeated manual tasks and bringing business and IT closer by providing intuitive self-service tools.W I D E-S PA N S O LU T I O N AT LOW COS T Accelerators enable experim

185、entations and fail-fast methods at a lower cost,making the outcome more resilient and tested.#DATA P O W E R E D#DATA M A S T E R S#I N S I G H T S&DATA#C A P G E M I N I#ACC E L E R ATO R S#A Z U R E#AW S#G C P#M E TA DATA#ACC E L E R AT E#DATA FAC TO R Y#M AC H I N E L E A R N I N G#DATAT R U S T

186、Accelerators work like a catalyst,without changing the chemistry of the underlying architecture.33Data-powered Innovation Review|Wave V 2022 Capgemini EVERYDAY AI:NEXT-GENERATION SELF-SERVICE ANALYTICS Organizations have been using the term“self-service analytics”for nearly a decade now,but for many

187、 companies its not a source of value generation.But businesses cannot afford to waste time or money spinning up or supporting a self-service analytics program that people arent actually using or thats not generating tangible value.Self-service analytics in the age of AI needs to be about truly enabl

188、ing people to ask infinite questions of their data and then empowering them to find or build trusted answers on their own.B R I DG E T S H E AChief Customer Officer at Dataiku 34Data-powered Innovation Review|Wave V 2022 CapgeminiImagine the smoothest self-service analytics experience possible at yo

189、ur organization today.It probably goes something like this:Business user looks at someone elses data product(such as a dashboard)built to answer some specific set of questions.Business user asks a question that is within the realm of possibility to answer with this data product(e.g.,“How did sales p

190、erform last quarter?”or“How did my marketing campaign perform?”).Business user gets a trusted answer without having to ask anyone in the middle,such as IT or a data team,for help.The future of self-service analytics is about empowering peopleBut what happens if this business user wants to ask new qu

191、estions that are outside the realm of what that dashboard was built for?For example,the marketing manager sees her campaign did not perform well and wants to understand who she should target for her next campaign,potentially even with a score predicting who will be most likely to open her emails.Or

192、a supply-chain manager has identified a pattern of shortages,but doesnt have the tools to dig in and get more visibility to address the problem.Most likely,every new question or business challenge is a new ask to a team to build a data product(dashboard)that provides answers from which they can self

193、-service.Its easy to say that the future of self-service analytics is about moving from descriptive to predictive(and even prescriptive)analytics.But its more than that.Its about empowering people,especially business users,to ask questions of their data and find or build trusted answers on their own

194、,whether that means building a dashboard or a machine-learning model for themselves.This is where the term“citizen data scientists”comes into play and why,in the future,the concepts of self-service analytics and citizen data scientist will become somewhat intertwined.Ultimately,putting the full powe

195、r of data in the hands of the people involved in the day-to-day business(we call this Everyday AI)is what will move self-service analytics from providing answers to providing impact and,with it,value.Yet,with great power comes great responsibility,so the key in the coming years will be for leaders t

196、o provide the framework that allows for this fundamental transformation.35Data-powered Innovation Review|Wave V 2022 Capgemini Building self-service analytics for impactA world in which data is accessible and anyone can build data or AI projects and solutions to answer business questions might sound

197、 scary.To be honest,without the right tools,technology,and processes,it is scary and can devolve quickly into data chaos.Seeing that risk,its critical in this new world of self-service analytics that the initiative:Doesnt exist in a vacuum.When businesspeople have the data,their questions for IT com

198、e up a level and can be more impactful.For example,how can I automate what Ive built so I dont have to update it every week with new data?Is built on trust.Leaders need to trust employees ability to use data in a self-service context.Business users working on self-service analytics need to trust the

199、 data that theyre working with.Managers and executives alike need to trust the insights delivered from self-service analytics projects.If just one of these layers is missing,it doesnt work.Has the proper governance built in,complete with appropriate guardrails.This can be as simple as proper permiss

200、ions management at the dataset or the project level,but it goes all the way up to the macro level.How are data and models being used?Who is monitoring this to lower the overall risk to the organization?For example,Dataiku customer GE Aviation implemented its own version of a self-service system that

201、 allows it to use real-time data at scale to make better and faster decisions throughout the organization.Engineering uses data from these tools to redesign parts and build jet engines more efficiently,the supply chain team uses it to get better data insights into its shop floors and streamline supp

202、ly-chain processes,finance uses it to understand key metrics,and more.At its core,its self-service program equips everyone(with proper access rights)with the ability to discover and use data,prepare that data,and create a data product,including developing predictive models within Dataiku.At the same

203、 time,it also ensures projects pass a set of checks,balances,and governance measures.36Data-powered Innovation Review|Wave V 2022 Capgemini Next-generation self-service analytics technologyThere are people in the business who have the ambition to go on their own data journey and will do it if the po

204、ints of friction are reduced and they are enabled to do so.This is the essence of the next generation of self-service analytics and,as previously discussed,of citizen data science.The idea behind the next generation of self-service analytics isnt that individuals can do and build whatever they want

205、with data(which would lead to data chaos).Its about empowering people,and choosing the right technology is an important milestone.The right technology should connect doers with data by bringing people of diverse skill sets together to work with data in a common ground.Ultimately,it should be second

206、nature for anyone in the business to produce new insights and to work with data in a way that is easily reusable.Individuals should benefit from the expertise of the many as timely new data products are created and maintained across the whole enterprise.Thats where the value lies.Innovation takeaway

207、sE M P OW ER T H E PEO PL EThe future of self-service analytics in the age of AI is intertwined with the idea of citizen data science both are about truly empowering people.PROV I D E T RU S T A N D I N D EPEN D EN C EBusinesspeople must be able to ask questions of their data and find or build trust

208、ed answers on their own.A L L A BO U T T H E TOO L I N G The right tools and technology are critical to enabling people while also maintaining the appropriate level of governance and control.#S E L F S E R V I C E#C I T I Z E N DATA S C I E N C ESelf-service program equips everyone with the ability

209、to discover and use data,prepare that data,and create a data product,including developing predictive models.37Data-powered Innovation Review|Wave V 2022 Capgemini THE REVITALIZATION OF DATA MESHThe great ambitions of the data mesh approach cannot be fully achieved with cloud-based or legacy-based th

210、inking.The use of emerging technical concepts around decentralization and distribution extending way beyond the enterprise can provide the answers.B E V E R L E Y COY Senior Solutions Architecture,Insights and Data,CapgeminiN E I L PA R K E RManaging Solutions Architecture,Insights and Data,Capgemin

211、i38Data-powered Innovation Review|Wave V 2022 CapgeminiData mesh is a new way of thinking.It is a concept proposed by Zhamak Dehghani in her introductory article,published in 2019,in which an organizations data-related capabilities are no longer organized into technically aligned functions.Rather,we

212、 see the data,process,people,governance,security,and other enterprise capabilities migrated away from monolithic,centralized functions and incorporated into the business domains.Within a business context of exponentially growing data volumes,multiple customer touchpoints,just-in-time supply chains,a

213、nd dynamic consumer relationships,business stakeholders need agility.They can no longer afford the time to navigate layer upon layer of technical teams to get answers to business questions.Rather,business users need to be empowered to create the data products they need autonomously,and without the c

214、onstraints of centralization.This is the promise of data mesh.Happily,the coalescence of cloud computing,serverless architecture,no-code platforms,DevSecOps,automated testing,semantic design,and other abstractions have reduced the technical expertise required to build a data product.Insight can be d

215、erived from data without coding or deep technical understanding of the underlying storage,compute,or network resources.Data mesh provides the overarching methodology for restructuring organizational capabilities across people,processes,data,and governance,such that these technical abstractions can b

216、e leveraged.Within the business domains,expertise is thereby connected with the ability to deliver data products.However,the interconnectedness of people,processes,data,and technology means that if any of these elements are constrained by residual centralization,all will be compromised and a decentr

217、alized data mesh will not be fully achievable.Dispelling mythsIn the book Sapiens:A Brief History of Humankind,Yuval Noah Harari observed that human progress was initially constrained to tribal units.To enable cooperation at scale,humanity created certain myths,such as law,governments,nations,and cu

218、rrency.Within an enterprise,these myths manifest as complex functional layers,such as IT,legal,and finance,which sit between data producers and data consumers.To understand how these centralized functions constrain data-mesh adoption,its easiest to ask the question:“Imagine we have implemented data

219、mesh.Is there anything which,if it ceased to exist or be valid,would prevent a business user from creating or consuming a data product?”The answer to this question,based on current approaches to implementing a data mesh,would likely include items such as network access,licenses,security credentials,

220、agreements with technology providers and use of their services,installed software,access to existing banking,clearing and payments,and others.39Data-powered Innovation Review|Wave V 2022 Capgemini A new paradigmOur vision of a further evolution of data mesh is one which enables an organization to pr

221、ogress from partial decentralization to a fully decentralized and distributed approach.We therefore propose a future data mesh where the following can all be true:Any data product can use unlimited storage and any compute(no resource constraints)Any data product can be invoked using the services of

222、any vendor,in an open marketplace,and this can be changed for each invocation without friction or disruption(no vendor constraints)Any data product can negotiate autonomously on each invocation,using programmatically defined,auto-executing contracts to access the most cost-effective storage,compute,

223、and other required services(no legal or relationship constraints)Any data product can be used by any trusted person within an industry or sector(no organizational constraints)Any data product can be created and accessed using only a browser and internet connection(no network or access constraints)An

224、y data product can pay for the services it consumes without recourse to a centralized financial institution(no monetary constraints)Any data product can be used by any trusted person without prior granting of privileges(no security constraints)Every data product is self-describing and contains both

225、data and metadata(no knowledge constraints)Any data product can negotiate with any other data product to determine who is master of a given data asset and owner of the requisite standards.Such negotiations will ensure there is always a master available,and all standards and metadata are correctly di

226、stributed across the data products(no governance constraints).Two worlds collideWhile some of the points above are aspirational and are included to demonstrate thought-leadership in data mesh,many are already practical and achievable.This is due to an opportune correlation between our thinking and a

227、nother emerging technical,social,and commercial paradigm:Web 3.0.Web 3.0 StatisticsThe attributes of Web 3.0 work beautifully as enablers for our future vision for data mesh.Web 3.0 is designed around the ideals of:Distributed cooperation through peer-to-peer protocols,contracts,and transactions Fri

228、ctionless participation using only a browser and internet connection Public or private interaction without a trusted(typically centralized)third party Open-source software,community maintained and technology/vendor agnostic Retained ownership,where the data product owner retains ownership of the con

229、tent Inherent privacy,where data and transactions remain private and unconstrained or exploited by intermediary service providers Community governed without invasive centralized control.195,000Searches per month forWeb 3.0 on Googleat the end of 2021(Google Analytics)$4.5bnThe investment fundingfor

230、438 start-ups offeringWeb3.0 services endtechnologies in 2021(Dove Metrics)350,000Active developers workingon blockchaindevelopment platforms in 2021(Dove Metrics)40Data-powered Innovation Review|Wave V 2022 Capgemini Web 3.0 services,based on the principles described above,are already emerging,incl

231、uding Distributed Ledger Technology(DLT),cryptocurrency,digital tokenization(such as NFTs),and decentralized finance(DeFi).Applying these concepts to a Web 3.0-driven data mesh can be shown schematically as follows:Fullty Decentralised Data MeshDATA PRODUCTDOMAIN USERS EXTENDING BEYOND THE ORGANISAT

232、IONORGANISATIONALBOUNDARY3.However,each organisationsinterpretation of the sum of theircontractual relationship is localised:these are the standards etc.representedby company specific metadataadministered by the Data Product.2.The same data product can be sharedacross multiple organisations that are

233、 in a contractual relationship.1.The data products are independent of organisational boundary.41Data-powered Innovation Review|Wave V 2022 Capgemini The creation and deployment of new data products follows the same philosophy:using open-source,vendor-agnostic frameworks,and leveraging no-code techni

234、ques,to deploy data products into a decentralized,distributed Web 3.0 architecture.So,while much of this is still theoretical,and testing of the concept still has a long way to go,if you are asked how to implement data mesh the answer may well be at least in the medium to longer term to start with W

235、eb 3.0 thinking.“With the concept of data mesh,I can see the data engineers role becoming more of a high-level SME(subject matter expert)being an advisor to the individual domains on the metadata,data catalogue,daily processing,and future requirements rather than the bottleneck.I look forward to the

236、 day where I can be looked upon as a trusted advisor instead of a bottleneck!”Beverley Coy,CapgeminiInnovation takeawaysB U S I N E S S E M P OW ER E M EN TData mesh is an important concept for empowering business users to create their own data productsMY T H I C A L CO N S R A I N T SThe aspiration

237、s of data mesh cannot be fully met with current enterprise resource organization models as legacy centralized artefacts will still persist.A N E W PA R A D I G M The emerging Web 3.0 paradigms offer a solution to enable a fully realized,distributed,and decentralized data mesh.#DATA M E S H#W E B 3#D

238、 EC E N T R A L I Z E D#A R C H I T EC T U R E#DATA P R O D U C T S#D I S T R I B U T E D L E D G E R#D LT42Data-powered Innovation Review|Wave V 2022 Capgemini PLAY DATA LIKE POKEREvery enterprise aims to find valuable business scenarios powered by data.However,there are many communication barriers

239、 between business people and data experts.Most of the currently applied approaches feel too traditional,not doing justice to the dynamics of data being closer to the business than ever before.The Lean Data methodology is a set of data-powered transformation activities and tools that integrate lean m

240、anagement,design thinking,and strategic planning methods.It has been successfully verified by many companies in the Chinese market.It includes the Lean Data Workshop,which use a unique card game.S H I K A ICTIO Insights&Data APAC,Capgemini43Data-powered Innovation Review|Wave V 2022 CapgeminiThe Lea

241、n Data Workshop is a lightweight,active engagement session,having businesspeople and data experts“play”with data like it is a poker game.It makes full use of the concept of game thinking and combinatorial innovation,enabling businesspeople and data experts to fully embrace data-powered principles an

242、d use cases.It creates an intimate,open,and free atmosphere for jointly developing and exploring a list of data-powered value scenarios.It then formulates a roadmap for data-powered transformation projects and implementation,to help enterprises use smaller input-output data strategies.A workshop in

243、four partsThe workshop is divided into four parts:preparation,guidance,co-creation,and planning,as shown in Figure 1.Input for the process is the business vision of its data-powered(digital)transformation.The output consists of blueprints,a list of data-powered projects and initiatives,and supportin

244、g activities and assets.Workshop phases2.GUIDE3.CO-CREATION4.PLANNINGInterview Research Design Vision aligment Concept training Technology empowerment Target decomposition Data blueprint Scenne innovation Roadmap Projects list Action plan44Data-powered Innovation Review|Wave V 2022 Capgemini Through

245、out the workshop process,a set of Lean Data cards is used as props to allow businesspeople and data experts to jointly master data-powered concepts and use cases,while playing a poker-like game.Highly valuable business scenarios are the key output,ensuring no blind spots are forgotten or left behind

246、.The entire card deck includes 10 card types,as shown in Figure 2:Deck of cardsThe workshop attendees play these cards following the facilitators guidance,as follows:The Business Vision and Business Target cards are used to get agreement with all the attendees on the overall direction.The Lean Value

247、 Tree(LVT)is used as a conceptual tool to facilitate,share,and drive a companys strategy and vision across all levels and teams of an organization.This provides an excellent kickoff of the workshop.Then,various ways are used to populate the enterprise Data Asset cards,arranging and combining them wi

248、th the Data&AI Technology cards and Toolkit cards to form a Business Value scenario position.This process has multiple supporting rules,making it very interesting and compelling to participate in.Next,Business Scenarios are created through Value Cards,leading to Project cards,required Resource cards

249、,and Action cards for steps to be carried out.In this way,the deck of Lean Data cards translates the complexities and difficulties of data-powered(digital)transformation into a compelling,engaging game that business personnel can easily understand and play mastering the secrets of data for business

250、in a relaxed,open,and creative atmosphere.LDC-C11Business Vision Card The agreed business vision is the foundation of digital transformation successLean Data Methodology Copyright Shi KaiLDC-C12BusinessTargetCardBreak down the business vision into measurable clear business goalsLean Data Methodology

251、 Copyright Shi KaiLDC-C2BusinessValueCardAligning business value is prerequisites for exploring valuable scenariosLean Data Methodology Copyright Shi KaiLDC-C3DataAssetCardData is a digital form of businessLean Data Methodology Copyright Shi KaiLDC-C4Data&AITechnologyCardLeverage data and ai technol

252、ogies ways to empower business InnovationLean Data Methodology Copyright Shi KaiLDC-C6ToolkitCardLeverage a variety of transformation tools and methodologiesLean Data Methodology Copyright Shi KaiLDC-C5BusinessScenarioCardBusiness scenarios are digitally transformed handle and driveLean Data Methodo

253、logy Copyright Shi KaiLDC-C7ResourceCardDigital transformation project requires support and investiment of many resourcesLean Data Methodology Copyright Shi KaiLDC-C8ProjectCardAll scenarios require projects to implementLean Data Methodology Copyright Shi KaiLDC-C9InitiativeCardLets get startedLean

254、Data Methodology Copyright Shi Kai45Data-powered Innovation Review|Wave V 2022 Capgemini BenefitsAs simple as playing a card game may seem,the Lean Data Workshop helps enterprises obtain four major data-powered values:1.Align businesspeople,data experts,and other stakeholders on a commonly understoo

255、d,shared business vision2.Jointly define and explore the Data Asset and Data Product blueprints of the enterprise3.Break down the typical barriers between businesspeople and data/technology experts while exploring the full range of potential value scenarios,eliminating blind spots through permutatio

256、ns and combinations4.Systematically prioritize value scenarios,based on a growing common understanding of the scenarios as they evolve.The Lean Data Workshop pioneered the combination of cards representing business,technology,and data to generate business scenarios and make the hidden data value exp

257、licit.The use of data has played a very beneficial role in promoting.imagine that everyone draws out a user data card,and then draws a clustering technology card,and then the business personnel see the clustered sample card scene and then say We want to make use cases.What an exciting scene!”Wang Gu

258、angsheng,chairman of the China Branch of the International Data Management Association(DAMAChina)46Data-powered Innovation Review|Wave V 2022 Capgemini All of this comes together even when initially in the midst of uncertainty and lack of understanding and direction to establish the right data-power

259、ed activities and start transformation as soon as possible.And its triggered by businesspeople and data experts,playing data like it is poker.The democratization of data is a key trend for the near future,as data more and more becomes central to powering the enterprise vision and strategy.It is nece

260、ssary to endow every businessperson with the ability to understand and activate the value of data,analytics,and AI.For that,an end-to-end,lightweight,agile,and lean approach to becoming a data-powered enterprise is instrumental.Lean Data is such a new(yet already heavily field-tested)methodology.It

261、contains many different activities,tools,and perspectives more to be addressed in follow-up articles,posts,and a book.However,the Lean Data Workshop particularly stands out as it brings together all key stakeholders in the data-powered enterprise with something as low-entry,playful,and joyful as a c

262、ard game.It is truly the ace of spades for anybody seeking to get the most out of data.Innovation takeawaysP OW ER TO T H E PEO PL EData and AI should not be mastered only by a few highly skilled data experts but should be fully utilized in the daily work of all businesspeople,realizing the democrat

263、ization of data.DO I N G I T D I FFER EN T LYData strategy and consulting activities should be carried out in a more engaging way,allowing businesspeople to express real and comprehensive needs in an interactive and immersive experience.L E T S PL AY C A R DS The Lean Data workshop approach lets bus

264、inesspeople and data experts jointly discover data-powered business value scenarios while playing a poker-like card game providing a fresh,innovative way of leveraging data.#DATA P O W E R E D#L E A N DATA M E T H O D O LO G Y#DATA D R I V E N47Data-powered Innovation Review|Wave V 2022 Capgemini EM

265、ERGING TECHNOLOGIES ARE TRANSFORMING THE ENERGY AND UTILITIES SECTORThe energy and utilities sector faces significant environmental and geopolitical challenges.To counter these,organizations are looking to take advantage of a flood of emerging technologies from artificial intelligence to the Interne

266、t of Things(IoT)to enhance efficiency,reduce waste,and facilitate regulatory compliance,while also inspiring new products and services and improving customer satisfaction.I S A B E L L E TACH E TLead Data Architect For Energy,Utilities,Data strategy,Sustainable AI,CapgeminiC A RO L I N E B A L LInno

267、vation Ecosystem Lead for Southern&Central Europe,Capgemini48Data-powered Innovation Review|Wave V 2022 CapgeminiThe challenges are not only significant,but also related.As the source of 73 percent of the worlds carbon emissions,the energy and utilities sector has a central role to play as economies

268、 transition to net zero.Meantime,global conflicts in todays politically destabilized world have highlighted the fragile nature of energy security in many countries driving up prices and making it more difficult to meet emissions-reduction targets.Similar challenges face the members of this sector th

269、at are not directly related to energy.For example,water utilities also face pressure to reduce their impact on the environment while ensuring a reliable supply of clean water for their customers.In this volatile environment,companies in this sector are embracing technology to address current challen

270、ges while also making their operations more resilient in the medium and long term.Here are some of the ways in which the energy and utilities sector is leveraging data-powered innovation.Modernizing the data estateAn organizations transformation starts by ensuring its working with a best-in-class da

271、ta estate.Many players in this sector have data ecosystems that have developed over decades.The result is a patchwork of systems,a mix of structured and unstructured data,and information locked in silos.Merging these into a single,easily accessible platform helps deliver additional value across the

272、enterprise while making it easier to enforce proper governance.As an example,Capgemini helped a water utility in eastern England modernize its data estate with a framework to consume,manage,and orchestrate information from smart meters and customer databases.The benefits are numerous.The utility can

273、 now apply enterprise analytics to smart metering and other projects.Analytics provides useful insights to automate the creation of work orders to more rapidly fix leaks or replace failed devices.The new architecture allows the utility to onboard new datasets more quickly with reduced costs,enabling

274、 the organization to more rapidly turn insights into action.The new architecture improves how the utility produces regulatory reports,and reduces the risk of non-compliance with the General Data Protection Regulation.And the utility has introduced an application that allows residential customers to

275、view their own water consumption and compare it to similar homes in their area,which helps customers reduce their water use.The power of SCADA and IoTBecause of the nature of the energy and utilities sector,companies operating in it manage a lot of physical assets.As consumers,all of us interact wit

276、h these on a daily basis.Think gas pumps.Think electricity and water meters.Further upstream,these companies monitor and control a broad range of industrial equipment.Integration of Information Technology(IT)and Operational Technology(OT)systems enable these assets to provide rich streams of telemet

277、ry data for next-generation supervisory control and data acquisition(SCADA)and industrial IoT.This in turn enables enterprises to make informed choices about when and how to maintain,repair,and upgrade assets.Capgemini worked with another UK-based water utility to connect asset data with customer ge

278、ospatial data.Now when theres a service disruption,this system can automatically alert customers in the affected area through a mass-messaging solution.This reduces the number of service calls the utility must process and by providing regular updates on progress and scheduled works in a timely manne

279、r the system has improved customer satisfaction.“Companies in this sector are embracing AI,machine learning,analytics,SCADA,IoT,and other technologies to address significant current challenges and make their operations more resilient in the medium to long term.”49Data-powered Innovation Review|Wave

280、V 2022 Capgemini Digital twinsIncreasingly,organizations are exploring opportunities to leverage digital twins virtual models of a physical asset that can be used to understand and optimize the assets performance or improve processes.Sensors in the asset collect real-time data.Machine learning and A

281、I can then be applied to this data to recognize patterns,predict faults,and call attention to deviations from normal operation.Capgemini helped an American petrochemical company apply AI,machine learning,and advanced analytics to its core chemical and polymer production processes.A new process impro

282、ved efficiencies and optimized product quality.Meantime,a digital twin used real-time production data including temperature,pressure,and feed rate to accurately and reliably predict quality specifications for more than 20 products.The solution helped this customer achieve process improvements of up

283、to 96 percent and generated more than$50 million in costs savings and revenue opportunities.Digital twins have numerous benefits for utilities operators,too.For example,they can be used to enable engineers to remotely monitor the performance of wind turbines and provide better information to field-s

284、ervice personnel when an issue occurs.Find out moreTechnoVision 2022 the latest edition of Capgeminis annual look at key technology trends describes how pioneering businesses leaders are leveraging technology to innovate,adapt,and achieve corporate and societal goals in a sustainable manner.The tren

285、ds are organized into multiple themes covering user experience,collaboration,data,processes,infrastructure,applications,and balance by design.For more on the technology trends affecting this sector,read the TechnoVision 2022 sector report on energy and utilities.Innovation takeawaysG E T I N S H A P

286、EEmerging applications demand top-quality data and lots of it.Modernizing the data estate and ensuring proper governance is the first step to unlocking the full potential of any organizations data.U S E YO U R A S S E T A DVA N TAG EAny enterprise with physical assets can leverage data from them to

287、enhance a broad range of business objectives.G E T T W I N N EDDigital twins are poised to leverage data to revolutionize everything from production processes to field service.#DATA E S TAT E#S M A R TG R I D#D I G I TA LT W I N S50Data-powered Innovation Review|Wave V 2022 Capgemini SUSTAINABILITY

288、SERVED52Putting AI in sustainability Tijana Nikolic&Robert Engels,Capgemini57Finding the code for a cureAnne Laure Thibauld,CapgeminiMike Miller,AWS61Creating a circular economythrough AI Faizan Pratyasha,Soumitra Upadhyay&Pratyasha Shishodia,frog66Data sharing is caring.Or is it a waste of investme

289、nt?Daniela Rittmeier,Capgemini71Three data-foundation strategies for Scope 3Roosa Sntti,Capgemini76AI roundtable on sustainabilityRajeswaran Viswanathan,AishwaryaJagtap,and Karan Kajrolkar,Capgemini80The paradigm shift of enterprise sustainabilityUmamaheswari Kathirvel,Yashowardhan Sowale,Capgemini8

290、4Feeding the world better with Project ENHANCEMaartje De Laat&Marijn Markus,Capgemini88Sust-AI-UN-ability:AI meets UN Sustainable Development Goals Niharika Kalvagunta,Capgemini51Data-powered Innovation Review|Wave V 2022 CapgeminiPUTTING AI IN SUSTAINABILITY Could AI offer a route to net zero,while

291、 realizing we are already late?And can current initiatives and developments play a role in it?What can we do to ensure AI can be used responsibly in reaching net-zero goals,accounting for its impact on the climate and unlocking its full potential in the climate action ecosystem?T I JA N A N I KO L I

292、 AI Specialist,Quality,Ethics,Sustainability,SogetiRO B E R T E N G E L SCTO Insights&Data NCE,Capgemini52Data-powered Innovation Review|Wave V 2022 CapgeminiThe impact of climate change is very real,and organizations in both the public and private sectors should play a leading role in mitigating cl

293、imate change.How?By embedding sustainable actions and social responsibility into their business models,and pledging to reach net zero in the coming years.Only when businesses are made accountable for reduction of the greenhouse gas(GHG)emissions they produce will we see change.The GHG protocol defin

294、es three scopes for this:scope 1 being direct emissions from organizational assets,scope 2 being purchased energy,and scope 3 being indirect company emissions across the supply chain.Making companies accountable has little effect without enabling their sustainability actions with pragmatic solutions

295、.Therefore,businesses investigate technologies such as artificial intelligence(AI)to help them reach their sustainability goals.Google achieved its net-zero goal already in 2018 and now uses AI to predict wind patterns to improve windmill yields,creating direct impact on scope 1 and 2 emission reduc

296、tion.Scope 1 and 2 seem to be manageable for footprint reporting,as most of the related data and operations are known and often controlled by organizations themselves.It is scope 3 reduction where it gets complicated.Not all the data is known,owned,or available for use.And scope 3 impact is often at

297、 least five times greater than the other two,so there is no path to net-zero reduction without incorporating it.This is no easy feat,due to a lack of transparency and standards on data collection in the supply chain and beyond.So,how can AI support this?Model-driven prediction and simulation can hel

298、p reduce all scopes by finding optimal choices in design,production,logistics,and life-cycle management.The lack of transparent data and implemented standards need to be dealt with.Digital twins bring together much of these needs in concise solutionsThe duality of“green”AIAlthough vastly positive,th

299、e impact of AI on the climate is a two-sided coin as development of AI its interaction with carbon-intensive applications and lock-in effects has potential negative impacts on the climate.The duality of using AI for net-zero goals is reflected in the UNs sustainable development goals(SDGs).This set

300、of 17 overarching goals has been described as“the worlds best plan to build a better world for people and our planet”.While supporting many of the SDG targets,the use of AI technologies also hinders some of them,highlighting that the fast development of AI needs to be supported by appropriate polici

301、es and regulation.53Data-powered Innovation Review|Wave V 2022 Capgemini The footprint of AI is shaped by the responsible use of AI,capacity building,international collaboration,and impact monitoring at every step.Impactful use of AI on net-zero goals depends on governments providing support in data

302、 and infrastructure,research and innovation funding,as well as supporting new AI-based solutions.AI in sustainable actionShaping AIs impactDATA&DIGITALINFRASTRUCTURERESEARCH&INNOVATION FUNDINGDEPLOYMENT&SYSTEMS INTEGRATIONREDUCING AISNEGATIVE IMPACTSON THE CLIMATERESPONSIBLE AICAPACITY BUILDINGINTER

303、NATIONAL COLLABORATIONIMPACTASSESSMENTData,simulation environments,testbeds,libraries,computational hardwareInterdisciplinary&cross-sectoral workguided by climate impactPolicy design&evaluation,market design,business modelsApplication andcompute-related impactsAI as a tool for climate actionImplemen

304、tation,evaluation and governance capabilities54Data-powered Innovation Review|Wave V 2022 Capgemini Enable sustAInabilityFocusing on using AI as a tool for climate action,data and the digital ecosystem are at its core.From a technical perspective,AI depends on data streams,applications,and processes

305、.Pitfalls can be expected with data availability and quality.Representative data is difficult to come by and is often blocked through regulations or otherwise inaccessible to data-processing pipelines.At the same time,computing power and hardware are required for large-scale AI deployment.Companies

306、and research institutions should be encouraged to create data portals for data sharing,establish data taskforces in climate-critical sectors,work on elaborate data standards,and provide affordable cloud-compute resources for AI development.NASA and ESA built a collaborative platform with computing c

307、apabilities,algorithms,and data for research and understanding of above-ground terrestrial carbon dynamics in the Multi-Mission Algorithm and Analysis Platform(MAAP),enabling efficient knowledge sharing and green computation.55Data-powered Innovation Review|Wave V 2022 Capgemini Reduction of AIs neg

308、ative impactsMoving on to the reduction of AIs negative impacts on the environment,we need to consider the AI practitioners responsibilities.Several aspects of AI require decisions which can make or break footprints.The choice of AI models and data pipelines is important for final emissions,as are t

309、he use of algorithm-optimized hardware(e.g.,specific,energy efficient,task-specific processors),their location(e.g.,local,cloud),the cost of data transport,and the type of energy used for computation.Given the computational power needed for training,deploying,and executing AI models,this becomes cru

310、cial in scalable and responsible use of AI.AI has a clear potential in driving positive climate action and helping companies reach their net-zero goals.But it is important to ensure this potential is not hindered by the negative impacts of AI on the climate.A responsible use of AI in lowering GHG em

311、issions is supported through data sharing,collaboration,and the responsible development of AI models.The EU has established the European Strategy for Data,enabling the creation of data spaces that ensure data is available and standardized for use across industries.Combined with GO FAIR,this encourag

312、es accessible and usable data for AI models and data for net zero.Innovation takeawaysCO N S I D ER A L L S CO PE SUtilize AI to its fullest extent and lower all three GHG emission scopes.R ECOG N IZE A N D COM B ATMeasure,monitor,and bring awareness to negative impact of AI on the planet,making sur

313、e to take steps to lower it.S U PP O R T A N D ED U C AT EEducate on the positive impact of AI on the climate and support research efforts in this topic.CO L L A BO R AT ERecognize partners and collaborate in the climate action ecosystem.#DATA P O W E R E D#DATA M A S T E R S#S U S TA I N A B I L I

314、T Y#G R E E N A I#A I 4 G O O D#S U S TA I N A B L E56Data-powered Innovation Review|Wave V 2022 Capgemini FINDING THE CODE FOR A CURECAPGEMINIS FLAGSHIP HACKATHON ADVANCES THE FIGHT TO ERADICATE RIVER BLINDNESSCapgeminis Global Data Science Challenge 2022 demonstrated the power of bringing together

315、 smart people and advanced technology,as more than 400 teams worked to help solve a health crisis that affects 20 million people worldwide.A N N E L AU R E T H I B AU DData AI&Analytics Group Offer Leader,CapgeminiM I K E M I L L E RDirector of AI Devices,Amazon Web Services57Data-powered Innovation

316、 Review|Wave V 2022 CapgeminiAI and machine learning have great potential to help address the many environmental and social challenges facing the planet.But to get the most out of these tools,its essential that developers get hands-on training in their use.Capgeminis Global Data Science Challenge(GD

317、SC)is proving to be an excellent way to enable this.Launched in 2016,the GDSC has grown to become Capgeminis flagship hackathon.For each event,Capgemini Insights&Data team identify a challenge and a participating NGO to the sustainability of the planet and invite teams to develop a solution using ar

318、tificial intelligence,machine learning,and other state-of-the-art data-science tools.The 2022 GDSC:Code for a CureFor the 2022 challenge,GDSC organizers joined the fight against onchocerciasis,more commonly known as“river blindness.”More than 20 million people worldwide live with this parasitic dise

319、ase.Most of those people experience severe skin irritations but in more than one million cases,the disease has also caused permanent vision loss.The World Health Organization is leading a global effort to eradicate onchocerciasis,and several drug candidates are currently undergoing clinical trials.B

320、ut evaluating drug efficacy is a slow,manual process.The 2022 challenge,titled Code for a Cure,invited teams to leverage the power of AI to accelerate the efficacy trials and help bring these important drugs to market more quickly.(The Data-Powered Innovation Review Wave 3 described river blindness

321、and the goals of this challenge in more detail.)The response was spectacular.More than 900 people from 31 countries took up the challenge,forming into more than 400 teams.These teams presented 8,457 possible solutions.AI-powered processes whittled down these submissions to five finalists and scored

322、them using criteria such as code quality,documentation and approach,and sustainable resource use.These finalists then presented their solutions to a panel of six judges representing University Hospital Bonn,Capgemini,and AWS.The judges assessed each finalist based on the effectiveness of the solutio

323、n,how well the solution could be applied to other use cases,the teams contribution to the overall learning experience during the“More than 900 participants hacked their way to a better understanding of AI and machine learning.Thats a huge win for these important technologies.”58Data-powered Innovati

324、on Review|Wave V 2022 Capgemini hackathon,and how well the team presented its results.We identified three big winners.The first winner isThe 2022 winning team is untitled.ipynb(named after the default file created when storing a Jupyter Notebook,a popular tool used in data science).The teams four me

325、mbers Abhijeet Gorai,Deepak Kumar Pandey,Utkarsh Prakash,and Prince Raj,all from Capgemini in India developed a solution that leverages the unique strengths of two architectures:Faster R-CNN a form of region-based convolutional neural network that can quickly and accurately predict the locations of

326、different objects Cascade R-CNN a multi-stage extension of a region-based convolutional neural network that uses the output of one stage to train the next.With increasing selectivity,this architecture can address problems such as close false positives in an image.untitled.ipynbs winning entry detect

327、s worm sections in prepared images taken from nodules extracted from the patient,then crops these sections and classifies them using various criteria,including dead/alive,male/female,and adult worm/child worm.The results will help researchers determine the state of the worms and hence the effectiven

328、ess of the drug being tested.Among its benefits,the teams model can be easily changed to train more than one class of objects,which enables the object detection model to be used to predict the class of the worm section.This means the approach does not require a separate classification model.59Data-p

329、owered Innovation Review|Wave V 2022 Capgemini The second winner isWhile only one team could place first,everyone who took part also won by gaining experience with advanced technologies.Participants invested more than 10,000 hours in learning about AI,including important tools such as Amazon SageMak

330、er.This platform is used to both prepare the data used to train machine-learning models and develop algorithms that enable the model to decide whats interesting.SageMaker lets developers and data scientists collaborate in a single space and because its cloud-based,it does not require a significant i

331、nvestment in computers or other on-premises equipment.This years GDSC also introduced participants to the idea of minimizing the cost of proposed solutions by measuring each entrys cloud compute time as part of the judging process which is an important,practical consideration whenever building machi

332、ne-learning models.The third winner isThis hackathon is a great contributor to a brighter future and not just for those affected by river blindness.Knowing how to use advanced technologies will be essential if were to address the significant challenges facing our societies and our environment.At the

333、 2022 GDSC,more than 900 participants hacked their way to a better understanding of AI.Thats a huge win.Innovation takeawaysR E M E M B ER T H E PR AC T I C A LDont let them become barriers,but as people explore the power of emerging technologies its important to consider practical issues such as“Can this be applied to other use cases?”and“How much compute time does this require?”T H E P OW ER O F