《BARC：2024企业架构优化赋能AI创新研究报告（英文版）（35页）.pdf》由会员分享，可在线阅读，更多相关《BARC：2024企业架构优化赋能AI创新研究报告（英文版）（35页）.pdf（35页珍藏版）》请在三个皮匠报告上搜索。

1、 Optimizing Your Architecture for AI Innovation Authors:Shawn Rogers and Merv Adrian Publication:March,2024 Abstract This report examines survey results for 335 Artificial Intelligence(AI)leaders and practitioners regarding their use of AI,including adoption status,technology optimization plans,proj

2、ect challenges,risks,and use cases.The report also shares specific insight into how High Readiness organizations are prioritizing their strategies,selecting trusted suppliers,addressing compliance and regulatory requirements,and the role of cloud platforms in AI innovation.Research sponsored by:2 Op

3、timizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Table of Contents Research Summary.3 Key Takeaways.3 Recommendations.4 Organizational Readiness.5 How Mature is Your Approach?.5 The Intersection of Budget and Strategy.6 Obstacles to Success.7 Strategies to Overcom

4、e the Skills Gap.8 How Enterprises Manage AI Projects.9 Applying Strategic Resources.9 Making the Best Decisions to Drive Success.10 The Role of Global SIs as a Resource.11 Responsible AI.12 Measuring Success and Failure.13 AI Use Cases.14 General AI Use Cases.14 Specialized and Industry-Specific Us

5、e Cases.16 Technology Choices.17 Transforming Your Architecture.17 Augmenting Your Architecture for AI.18 AI Cloud Strategies.20 Data and the Dreaded Model Choices.21 Data is King.21 Biggest Doesnt Always Equal Best with Models.22 Specialty Models in Your Ecosystem.23 Model Management is a Critical

6、Strategy for Success.23 Methodology and Demographics.25 Respondent Panel.25 Authors.28 3 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Research Summary Enterprises across the globe are grappling with an enormously disruptive“new”technology stack as Artificial

7、 Intelligence(AI)implementations are touted(and rightly so)as the next stage in the evolution of information technology.A tsunami of vendor messages,research firm marketing and business pundit proclamations is driving anxious questions about timing,preparedness and possible competitive disadvantage

8、among those firms that have not already ramped up their efforts.Even corporate positioning is affected;boards and executives want to know that they are moving towards adoption and want to tout their progress to fend off competitors who may or may not be moving faster.But clarity is hard to find:Whic

9、h technologies are essential?Where can skilled resources be sourced?What is the status of regulatory requirements?How hard will it be to integrate these new technologies?This research examines the intentions and readiness of potential buyers of these technologies by asking them direct questions abou

10、t what they know or dont know;how they propose to build,govern and measure their plan;who their stakeholders,suppliers and partners are;and finally,how specific to their industry they expect their strategies and technology purchases to be.Worldwide survey respondents totaled 335,with the majority sp

11、lit between Europe and North America balanced by a variety of company sizes,job titles(IT,Executive and line of business)and multiple industries.This is new territory for everyone.Even firms with a history of“AI projects”have been working with highly bespoke technology.The sudden recent ascendance o

12、f several nearly relatively standardized components,such as generative AI built on foundation models,vector databases and retrieval-augmented generation,has caught nearly everyone flat-footed.Existing technology providers have varying degrees of readiness for deploying,integrating and governing thes

13、e technologies,and their service providers,such as systems integrators,are not yet substantially ahead of their prospective clients in experience.Key Takeaways Everything you think you know about the AI market is wrong,perhaps driven by self-interested vendors and research firms with a clickbait app

14、roach to marketing.We were surprised many times,and you will be too.Over 95%of respondents are currently addressing their plans for AI.However,only 20.5%can be characterized as being in a state of“High Readiness.”Respondents have made the most progress in the area of security standards and complianc

15、e,with 56%saying they have either formalized or are reviewing/revising their initiatives.Although conventional wisdom claims IT is becoming less involved in computing decisions as decision-making shifts to business units,for this transformation,IT is involved 95%of the time,ahead even of Executive M

16、anagement.The single largest challenge to the successful implementation of AI innovation is the skills gap,combined with formulating a strategy to overcome it.Unique and innovative technology is not the first criterion companies use to identify who they want to engage with for AI and GenAI projects:

17、familiarity and credible use case experience lead this strategy.4 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Hyperscalers are battling to win the war for AI workloads in the cloud.Microsoft Azure is the most utilized platform overall with respondents and i

18、s winning the race for AI customers.Data of all types is a critical part of a successful AI strategy.Quality,accuracy,access and orchestration have long been challenges for business intelligence and analytics practices.AI has surfaced a renewed requirement for highly performant data systems.Recommen

19、dations 1.Address the skills gap fast and early.Engage your AI workforce with free training and certifications,and embrace knowledge sharing and collaboration at the beginning of your AI journey,not the middle.2.Keep Calm and Carry On.This is an early-stage marketdont let haste and FOMO(fear of miss

20、ing out)drive your strategy.And dont be surprised at the inevitable backlash from dissatisfied customers that will appear in late 2024 and well into 2025.Mistakes will be made.3.Begin with your existing,trusted suppliers.Virtually all of them are seeking to add these capabilities,and thus far few ha

21、ve demonstrated a convincing lead over their competitors.Fit and integration will be key to timely delivery,and your best suppliers will have an advantageassuming you qualify them correctly.4.Compliance and regulatory concerns should be at the forefront of your planning but dont make the critical mi

22、stake of overlooking a complete Responsible AI strategy.Aligning your AI practice with company ethics and KPIs is a foundational strategy for avoiding common mistakes regarding bias,human collaboration,security and accuracy of enterprise AI.5.Augment your existing business intelligence and analytics

23、 stack with AI.Without massive new disruptive capital investments,you can deliver immediate return on investment.Enable your teams with time-saving automation and serve a wider enterprise community with AI-automated insights,narratives and smart dashboards.5 Optimizing Your Architecture for AI Innov

24、ation,March 2024 BARC GmbH,IT MARKET STRATEGY Organizational Readiness How Mature is Your Approach?The overwhelming interest in these projects is evidenced in the answers to a series of questions we asked about organizational readiness and plans.We began with a basic planning question:“At what stage

25、 is your organization with these critical AI planning/process initiatives?”Only 5%are not yet dis-cussing their plansalthough over half are still in the early stages.Figure 1 shows the various stages and where our respondents stood in rolling out each of these critical initiatives.Figure 1:At what s

26、tage is your organization with these critical AI planning/process initiatives?(n=335)In each of the dimensions examined here,roughly half of the respondents have arrived at drafts or final formalized planning/process stages.We refer to those who have done so across all seven dimensions20.5%of the pa

27、rticipantsas“High Readiness.”The most mature dimension is Security Standards and Compliance,with over half(56%)saying they have either formalized or are reviewing/revising their initiatives.73%have moved beyond the first two stages(researching and consulting stakeholders)for this issue.29%28%25%24%2

28、2%18%15%21%28%26%27%27%23%29%15%16%19%14%18%22%18%13%9%10%11%11%12%14%18%13%15%19%19%20%18%4435355Identifying AI leadership in organizationSecurity standards and complianceData access/use policiesLegal considerationsAI program standards and policiesProject governance oversightEnterprise architecture

29、 requirementsFormalizedReviewing/revisingDrafting policyConsulting stakeholdersResearchingNot under discussionHigh Readiness 6 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Scores are fairly consistent for other issues,mostly in the mid-40s in percentage term

30、s for draft or formalized plans.In a theme we will see echoed in other questions,the lowest combined readiness score(41%)is for project governance and oversight.Many organizations have not decided just how they will execute and govern projects,or measure their progress.As we discuss our findings,we

31、will highlight the responses from the High Readiness organizations.Their answers will be less speculative and illuminate the early lessons already being learned.North American respondents were far more likely than European ones to be part of the High Readiness cohort87%compared to 10%.The size of th

32、e organization had less variability,clustering around 15%.Only firms with revenue between$10 million and$50 million had fewer than 10%High Readiness participants.Those between$1 billion and$5 billion came in at 18%,but those over$5 billion at 15%.From the industry perspective,information technology,

33、retail/wholesale and manufacturing respondents are significantly more likely to be High Readiness than the other industries represented.The Intersection of Budget and Strategy Although leadership for the initiatives is typically not yet well defined,it is clear where those decisions are influenced.W

34、hen asked,“Which department(s)in your organization are influencing AI projects?”the findings in Figure 2 were clear:Executive Management and IT are very present in this new round of investment.Combining the scores for Setting Strategy,Providing Budget and Both shows that IT is involved 95%of the tim

35、e,with Executive Management close behind with a combined total of 88%.Figure 2:Which department(s)in your organization are influencing AI projects?(n=321)Not surprisingly,Finance ranked highest for providing the budget,but its impact on strategy lags the leaders.While Regulatory and Legal were quite

36、 involved in setting strategy,they are not budget 32%22%35%38%26%31%17%26%25%37%23%17%20%22%12%17%12%12%16%17%33%13%15%12%15%21%20%14%51%49%39%30%38%30%28%36%34%22%29%29%27%22%5%12%13%21%21%22%22%25%26%29%32%32%34%42%Information TechnologyExecutive ManagementCenter of ExcellenceRegulatory and Compli

37、anceOperationsResearch and DevelopmentFinanceMarketingSales/Business DevelopmentLegalCustomer Service/SupportProcurement/Supply ChainManufacturing/ProductionHuman ResourcesSetting strategyProviding budgetBoth strategy and budgetNot presently involved 7 Optimizing Your Architecture for AI Innovation,

38、March 2024 BARC GmbH,IT MARKET STRATEGY sources.HR is the least involved of all,below Manufacturing,Procurement and Customer Service,which all came in very low.Given the number of use cases associated with improving manufacturing processes and customer-facing systems,these results were a bit surpris

39、ing.Obstacles to Success Unsurprisingly,as seen in Figure 3,lack of skills is the most frequently cited obstacle(39%)to new initiatives.High Readiness firms were even more likely to name this41%did so.This was followed by costs or limited budgets(36%),although High Readiness respondents ranked it fi

40、rst at 45%.The latter reflects clear awareness that new investments are likely to be requiredthis is not a routine technology evolution but one that will have far-reaching investment requirements.Moreover,it has become apparent through our conversations with end users and vendors that a new team is

41、likely to be formed alongside existing ones to take on these new projects.The organizational challenges cited earlier are very present herenearly a quarter of respondents cite cross-functional collaboration(24%)and lack of leadership and strategy(20%).However,High Readiness firms seem to have firmed

42、 up their leadership and strategy plans,and only 10%have named this issue.This was the second-largest single difference between the two cohorts.The largest gap between cohorts was the issue of model indemnification and trust,which was named by only 7%of High Readiness respondents but 25%of the other

43、s.Figure 3:What obstacles are slowing/stopping your organization from delivering on your AI strategy?(n=335)39%36%24%24%21%21%20%20%19%16%15%15%10%Lack of AI skills/expertiseHigh costs/budget limitationsCross-functional collaborationIntegration issuesModel indemnification and trustInsufficient data,

44、data accessModel management and complexityLack of leadership and strategyInsufficient data qualityUser/customer resistance to technologyRequired technology/product not available fromvendorsFine tuning models,grounding models and RAGNo obstacles at this time 8 Optimizing Your Architecture for AI Inno

45、vation,March 2024 BARC GmbH,IT MARKET STRATEGY Technology-related issues,such as model management or tuning,simple availability of technology from vendors,or integration,as obstacles tended to fall at or below 24%although 28%of High Readiness firms named integration.And,surprisingly,obstacles relate

46、d to databoth access(21%)and quality(19%)ranked quite low here as well,and High Readiness respondents were even less concerned,citing each only 13%of the time.These technology-related responses may reflect faith in the marketing messages respondents have heard,telling them that data identification,a

47、ccess,quality and governance are easily solved if only they buy the products being offered for those purposes.In fact,many of these technologies not perceived as obstacles are relatively unavailable to AI tools,are in early preview or have significant gaps in many vendor offerings today.Clearly,impl

48、ementers are likely to be hit with surprising technology capability challenges in the next round of deployments.This soon-to-be-seen“buyers remorse”is not unusualin fact,it is a well-understood phase of the technology adoption cycle.But it can be minimized:careful vetting of new technology purchases

49、,development of clear goals for projects and metrics,and governance for their success can all help avoid dead-end investments and activities.Strategies to Overcome the Skills Gap The primary reason why companies are struggling to leverage the opportunities presented by AI and GenAI technology fully

50、is the AI skills gap.How will research respondents address their knowledge gap?First,by upskilling and reskilling their existing workforce,selected by 68%of respondents.This strategy is followed closely by internal knowledge sharing and collaboration at 61%.High Readiness organizations were even mor

51、e likely to cite upskilling(75%)but were less optimistic about internal knowledge(50%)as a strategy.AI as“the disruptor”is often blamed for reorganizations and reductions in force(RIFs)by companies in all industry sectors.To close the AI skills gap,many companies are refitting their teams with new h

52、iring and talent acquisition.High Readiness firms are far more likely to rely on AI-managed services offeringsthey ranked it second at 61%.This was an even greater difference between the two cohorts than the internal knowledge question.Figure 4:Select the top 3 strategies your organization is utiliz

53、ing/will utilize to address the AI skills gap(n=130)68%61%52%48%41%31%Upskilling and reskilling existing workforceInternal knowledge sharing/collaborationHiring and talent acquisitionRelying on AI managed services offeringsStrategic outsourcingCreating Centers of Excellence 9 Optimizing Your Archite

54、cture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY From a global perspective,EU-based firms are nearly 10 percentage points more likely to focus on upskilling and reskilling compared to their North American(NA)counterparts.Additionally,there are significant differences in the utilizatio

55、n of strategic outsourcing as NA-based firms are 6 percentage points more likely to bring in skills from this source.When analyzing strategies by company size,small firms(under$10 million annual revenue)are highly dedicated to upskilling,with 80%of respondents including that strategy in their top 3.

56、This trend is similar to that of global companies having over$5 billion in annual revenue.The trend is lower with companies in between these groups.Previous BARC research of 529 AI professionals in October 2023 shows that after reliance on IT groups,respondents rank Centers of Excellence(BI,analytic

57、s,AI and cloud)at 23%for the group driving AI strategy within their company.How Enterprises Manage AI Projects Applying Strategic Resources Developing a successful strategy for implementing and adopting new technologies like AI and GenAI requires a wide variety of resources.87%of respondents identif

58、ied Internal IT Resources as likely or extremely likely to play a role in supporting or enabling their companys AI innovation strategy,reinforcing the leadership we noted above on strategy and budgeting.IT organizations deliver highly valuable domain expertise that can help fast-track adoption and o

59、vercome challenges with innovative technology.Moreover,they are far more likely to have established governance and process metrics to fast-track new initiatives.Over 83%of respondents ranked reliance on Existing Technology Vendors as the second most popular selection.This is certainly reasonable:the

60、re are few true greenfield scenarios.Every new AI project will draw on,integrate with,and often help operate and improve existing systems.If your existing vendors are not in the lead,they certainly will need to be on the team,and you need to let them know that is expected.Figure 5:Which of the follo

61、wing resources will support/enable your companys AI innovation strategy?(n=335)48%35%34%28%27%23%39%48%45%44%38%35%8%13%13%16%21%22%435%9%9%12%335%7%Internal IT resourcesExisting technology vendorsNew vendors and technologyPartner-ledRegional technology consultantsGlobal technology consulting firm(s

62、)Extremely likelyLikelyNeither likely nor unlikelyUnlikelyExtremely unlikely 10 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY These top responses,coupled with high scores for Partner-Led and Regional Technology Consulting firms,suggest that these existing rel

63、ationships are perceived as offering less risk for companies.With this said,the door remains open for New Vendors and Technology to enable AI within companies,as 79%of respondents indicated they are likely or extremely likely to include them in their strategy.Its important to note that the substanti

64、al number of technology components that will be required are available from specialists,providers of data and analytics products and platforms,and large platform providers,including hyperscalers.For many organizations,a key decision criterion might be current and future scope.Niche vendors will inev

65、itably be acquiredor failas the market evolves.Reliance on them for key components of the strategy may be risky.But the readiness of the big players will not be consistent across all layers of the stack and must be assessed thoroughly.Making the Best Decisions to Drive Success Optimal selection crit

66、eria for users selecting existing or new vendor technology play a critical role in de-risking AI-driven projects while enhancing project success.While Unique and Innovative Technology is part of the scenario at 37%,it is not the first criterion companies are using to identify who they want to engage

67、 with for AI and GenAI projects.Figure 6:When selecting new and existing vendors for AI projects,what drives your selection and reduces risk?(n=332)Survey respondents will begin with players they are already working with and are happy with:they identified“tangible experience”as the leading criterion

68、 for vendor selection in AI projects.42%of respondents included Past Experience with Vendor Performance,resulting in the top driver for selection.Buyers are savvy about whether new technology should be considered in isolation:37%selected Customer Examples and Success Stories.Both selections demonstr

69、ate the importance of proof from the customers perspective.42%37%37%33%31%22%21%20%15%14%Our past experience with vendor performanceCustomer examples/success storiesUnique or innovative technologyVendor professional servicesTechnical alliancesThird-party review sitesPeer referralOut of the box,templ

70、ates and acceleratorsContribution to standards bodiesVendor Net Promoter Score 11 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY The Role of Global SIs as a Resource The list of service providers that will be chosen for these projects is varied.Leading systems

71、 integrators like Accenture,Deloitte Consulting and McKinsey&Company came in at the top of the charts.From a regional perspective,44%of North American respondents identified Accenture as the top choice,followed by Deloitte at 40%,with McKinsey and Company a distant third at 28%.The European responde

72、nts selected“Other”as the top choice at 38%.Niche regional solution integrators and consultants varied the write-in answersa very different trend compared to the North American respondents.Accenture was named second most by European respondents at 27%.Figure 7:Which of the following is your organiza

73、tion likely to work with on AI strategy and projects?(n=229)Large platform providers who have their own consulting offeringsmost notably,leading hyperscalers AWS,Google,IBM,Microsoft and Oraclewere not listed among the choices for this question,in part to test whether respondents in the space provid

74、ed would write them in.They did notonly a handful of references to these vendors as providers for strategy and projects were made.This is another surprise since those vendors are all aggressively marketing their offerings in this market.Future planned research will include more focus on this issue.3

75、8%33%24%20%19%19%18%17%16%16%11%8%6%14%AccentureDeloitte ConsultingMcKinsey&CompanyBoston Consulting GroupCognizant Technology SolutionsPricewaterhouseCoopersKPMGErnst and YoungCapgeminiInfosys ConsultingAtosCoforgeTata ConsultingOther 12 Optimizing Your Architecture for AI Innovation,March 2024 BAR

76、C GmbH,IT MARKET STRATEGY Responsible AI Informed companies preparing to leverage AI-driven innovation are monitoring the quickly evolving landscape of global AI regulations.While that terrain can prove complicated with the EU AI Act,the US NIST Guidelines and others,its important to include Respons

77、ible AI as an equally critical part of the overall success strategy.Responsible AI acts as a guiding set of principles for how your company will manage data,train models,and deploy and leverage AI.Its important that these principles align with your corporate ethics and culture.These policies need to

78、 be shared transparently across your organization and with customers and partners.Setting goals and policies for data and model bias,accuracy,human-AI collaboration,job impact,security and privacy are all topics responsible companies need to explore and align with.Figure 8:What components of Respons

79、ible AI is your organization prioritizing?(n=335)This data clearly illustrates that companies are prioritizing the Security of AI Systems and Models and Data Privacy as the top Responsible AI strategies.High Readiness respondents had slightly higher scores for both of these.And,as noted above,this i

80、s the issue respondents identified as having the most mature planning efforts in this survey.Previous BARC research on this topic,conducted in October 2023 with 539 participants,recorded that over 6%of respondents did not have a strategy for Responsible AI five months later.We see some incremental p

81、rogress on the topic.40%36%32%29%21%21%20%19%19%16%12%10%9%8%3%Security of AI systems and modelsHuman-AI collaborationData privacyRegulatory complianceRisk management and indemnificationData biasTransparency and auditabilityAccountability processes and policiesFinancial budgeting and governanceAccur

82、acy,hallucination managementEquity and inclusivityEnvironmental sustainability and impactJob impact and lossModel biasNo responsible AI strategy at this time 13 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY We expect risk management and indemnification to bec

83、ome more important over the coming months as customers adopt large language models trained on public data models.In 21%of our responses,this was selected as a priority.High Readiness respondents were far less likely to cite Human-AI Collaboration:at 22%,it was not in the top 4 responses,with a gap o

84、f 17 percentage points compared to non-High Readiness respondents.It may be that they see the challenges of defining and tackling these issues more clearly.Measuring Success and Failure Its impossible to measure success for any project without clear metrics,and few are easily quantified.Organization

85、s will be challenged to define how they will measure accuracy and alignment,although they are admirable goals.Only 1%of organizations say they have no plans to measure success,which is encouraging,though it may be either naive or disingenuous.However,it is concerning to note that the speed of decisi

86、ons,regulatory requirements and timely completion were at the bottom,all cited by fewer than 19%of respondents.All are relatively measurable and perhaps assumed to be well in hand.In the analysts experience,they rarely are.High Readiness firms ranked them higherwith completion on time coming in fift

87、h overall with 26%,exceeding the overall rate by 9 percentage points,and the non-High Readiness response by 11,the largest difference between the two groups.As elsewhere,High Readiness respondents were less concerned about datain this case,protection and auditability(19%)than the non-High Readiness

88、cohort(28%).This was the second largest difference.Figure 9:Select the top 3 ways your organization will measure success with its AI-driven projects(n=335)33%32%31%27%26%26%24%21%21%19%18%17%1%Accuracy and trust of insightsAlignment to business function(s)or goalsCustomer satisfactionUser adoption o

89、f AI application or solutionData protection and auditabilityInternal quality standardsOverall project costsSpeed and performance of the systems in useAlignment to corporate level KPIsSpeed of decisions or insightRegulatory requirementsCompleting the project on timeSuccess not(planned)to be measured

90、14 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY AI Use Cases General AI Use Cases When it comes to leveraging AI and GenAI,many users ask,Where should we use AI first?Companies feel pressure to rush because of a fear of missing out(FOMO),so they often begin

91、with general AI use cases that appear likely to help them expedite the project and experience an immediate return on investment(ROI)without taking on a lot of risks.This doesnt mean that these use cases lack value,but they do provide a faster,less complex on-ramp to integrating AI into a companys cu

92、lture.The most popular general use cases are AI chatbots and intelligent assistants,with 28%of respondents deploying them and another 24%in proof-of-concept(POC)testing.Figure 10:Which of the following general AI use cases has your organization deployed,planned or is researching?(n=334)28%25%23%22%2

93、2%21%20%20%19%18%18%15%14%24%22%28%23%23%27%25%23%23%25%24%23%20%20%24%25%21%26%22%27%21%23%26%22%17%18%18%13%13%19%14%16%21%21%13%19%13%15%17%10%16%12%15%15%14%7%16%23%12%22%30%31%ChatBot/intelligent assistantData security and privacyDocument analysisCustomer service and supportCustomer experience

94、enhancementContent creationAI-infused business intelligence and analyticsCoding and/or API DevRecommendation enginesData management tasks and integrationImage processing and recognitionSentiment analysisMass personalizationDeployed to productionProof of concept(POC)in progressPlanned target next 12

95、monthsResearching target next 2 yearsNot applicable/not in plans 15 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Use cases for AI-infused business intelligence and analytics,as well as AI-driven data management tasks and integration,are becoming critical for

96、 companies early on.When combining deployed,POC and planned projects in the next 12 months,the data shows that both are top-tier initiatives.Data management use cases fit this group with 69%,and business intelligence and analytics at 72%.AI has been used to improve user experience via recommendation

97、 engines,automated interactions and personalized data.Its surprising that this use case came in lower,as early adopters have found positive ROI by infusing AI in it.23%of respondents surveyed plan to adopt this use case in the next 12 months,and the same number are actively testing it.Across all gen

98、eral use cases,mass personalization is the least popular,with 31%of respondents indicating that it is not part of their AI projects,and only 14%indicating that they have deployed that type of use case.16 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Specializ

99、ed and Industry-Specific Use Cases Specialized and industry-specific opportunities can be attractive because they may offer rapid competitive advantage and differentiation.They may also be technically“smaller”because they work with highly specific,already documented fact basesa frequently cited reas

100、on for investing in retrieval-augmented generation(RAG,discussed below in“Model Management is a Critical Strategy for Success”).The most frequently named use cases or industry-focused plans shown below in Figure 11 occur in Sales Forecasting/RFP Generation,with this use case deployed or underway wit

101、h 80%of the respondents.Fraud Detection at 77%,Predictive Maintenance at 76%and Robotic Process Automation at 75%deployed or underway round out the four.Disease diagnosis was far down at the bottoma huge surprise considering the early popularity of predictive healthcare stories in the market.Figure

102、11:Which of the following specialized/industry-specific AI use cases has your organization deployed,planned,or is researching?(n=334)17%24%20%20%15%15%14%17%12%13%9%322%20%24%19%23%22%22%21%20%20%18%7%26%19%19%18%20%23%19%20%22%18%13%314%13%14%18%16%14%16%13%14%15%13%420%23%24%25%26%26%28%29%32%33%4

103、7%82%Sales forecasting/RFP generationFraud detectionPredictive maintenanceRobotic process automationSupply chain managementPrice optimizationTalent and HR managementInventory managementFinance,trading and marketsEnergy consumption optimizationAutonomous vehiclesDisease diagnosis/predictive healthcar

104、eDeployed to productionProof of concept(POC)in progressPlanned target next 12 monthsResearching target next 2 yearsNot applicable/not in plans 17 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Technology Choices Transforming Your Architecture Innovation often

105、leads to disruption,especially when new technology is being incorporated.This can result in significant strategy shifts to ensure enterprise architectures can accommodate new use cases and workloads.The important question is whether your infrastructure is prepared for AI and GenAI,and what technolog

106、y will be needed to meet new challenges.When asked,“How will AI innovation and projects transform your companys technology architecture?”our respondents rejected the notion that previous architecture investments should be overhauled or scrapped to make way for AI.The data thus indicates that this is

107、 not necessary,and most respondents(61%)are simply integrating AI-specific technology into their existing architectures to fill gaps and add functionality.Figure 12:How will AI innovation and projects transform your companys technology architecture?(n=335)Not at all surprising in the research is tha

108、t only 4%of respondents state that their company architecture is already optimized for AI,further supporting the idea that while a sense of urgency exists around taking part in AI,you are likely in the 96%group that still has work to do to prepare for AI/GenAI innovation.The survey data reveals that

109、 companies are focused on making changes to their data management,analytics and data science architectures,with consolidation around AI being a priority.This is not a surprise.Every year,data management software spending is the largest single category of IT infrastructure spending,and business intel

110、ligence/data science is not far behind.This has been the case for decades.61%44%41%32%4%Add AI-specific technology to our existingarchitectureOverhaul existing data managementarchitectureOverhaul existing analytics and data sciencearchitectureMigrate more of our architectureto hyperscaler cloud comp

111、aniesOur architecture is already optimized for AI 18 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Hyperscaler cloud companies have made a compelling value proposition out of helping clients take advantage of their one-stop shops for easier AI implementation.

112、Bringing AI projects on board will significantly accelerate cloud transformations for one-third of surveyed companies:32%of respondents plan to migrate more of their architecture and workloads to cloud solutions.However,we predict that cloud platform costs will play a role over the long term as cust

113、omers face difficulty managing monthly cloud transactions.Those costs have not gone unnoticed,and platform providers are already taking concrete steps to demonstrate their commitment to help.Augmenting Your Architecture for AI A bewildering set of new technologies is vying for buyer attention.The co

114、nfusion is evident in the similarity of responses on this question,and even more so in the size of the“not in our plans”or otherwise called“I dont know”responses for eachtypically in the 15%range.However,both vector databases(24%)and knowledge graphs(18%)came in higher.Foundation models and optimize

115、d hardware have moved the fastest,with combined testing and production rates at 58%,but the gap to other technologies is small.Vector databases,AutoML environments and centralized feature stores are the laggards overall,but here again,the gap is small.Figure 13:What is the status of the following AI

116、 technology in your existing environment?(n=298)24%24%23%18%18%17%17%14%14%12%29%26%27%27%24%30%28%31%34%30%26%25%25%29%30%29%27%28%29%33%20%24%25%26%28%24%29%28%24%25%Vector databasesCentralized feature storesAutoML environmentsKnowledge graphsOptimized compute hardwareModelOps,management and monit

117、oringAdditional cloud servicesCompliance and regulatory managementData fabric,data meshFoundational models,model selectionNot in our plansResearching/RFPTesting/POCsIn production 19 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Perhaps the most interesting da

118、ta comes from comparing the responses of the High Readiness cohort to the rest of the survey sample.It comes as no surprise that respondents who are identified as highly ready would be ahead of the other panels.The following chart,Figure 14,reflects the experience and priorities of the two groups an

119、d can be used as guidance for anyone who has yet to qualify as a High Readiness user,as we defined in the first section of this research(How Mature is Your Approach?).From a priority standpoint,optimized compute hardware has risen to the top for the High Readiness group,with 83%already in production

120、 or testing/POC.From the analysts viewpoint,this is likely due to their focus on cost optimization and budgeting for AI projects.Choosing foundational models early in their journey has also been a priority,with 77%already in production or testing/POC,representing a gap of 29 percentage points betwee

121、n their experience and that of all other respondents.Knowledge graphs round out the top three with High Readiness respondents at 77%in production or testing/POC,representing a 28 percentage point gap in their experience versus all others.The same level of attention has not been paid to the inclusion

122、 of vector databases.Our analyst view is that often,this technology is integrated with other components of the stack,especially DBMS,and while it delivers significant value in an AI technology stack,its not always upfront in the solutions being used but powering them instead.Figure 14:AI technologie

123、s with status“Testing/POCs”or“In Production”by Readiness(n=296)83%77%77%70%76%74%78%72%75%65%51%48%49%43%50%48%53%47%51%42%Optimized compute hardwareFoundational models,modelselectionKnowledge graphsAutoML environmentsCompliance and regulatorymanagementAdditional cloud servicesModelOps,management an

124、dmonitoringData fabric,data meshCentralized feature storesVector databasesHigh ReadinessAll other respondents 20 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY AI Cloud Strategies Companies utilize cloud platforms for a variety of mission-critical workloads.Th

125、is research investigated AI use and compared and contrasted how companies are using the cloud today for AI,adding cloud to support AI workloads,and whether they are already using the cloud for data analytics.The most utilized cloud platforms among respondents are Microsoft Azure at 76%,Amazon Web Se

126、rvices(AWS)at 69%and Google Cloud Platform(GCP)at 61%.Leadership by the“Big Three,”as they are often called when referring to Hyperscalers,is not a surprise in these results.General platform utilization is consistent with High Readiness customers,without significant differences.The least utilized fo

127、r any workloads are Alibaba Cloud at 30%,Baidu Cloud at 24%and Tencent Cloud at 23%.These findings match nearly all global cloud utilization research and our global audience sample trends.There is a similar trend when examining how respondents rank their use for data and analytics workloads.38%of re

128、spondents use Azure for data and analytics,30%utilize AWS and 22%are using GCP.There are some differences when examining how High Readiness users adopt these platforms for data and analytics,though their rank is the same.For Azure,its 41%,AWS is 34%,and GCP is used by 26%of these respondents.Figure

129、15:Which of the following cloud platforms does your company utilize for general data,analytics and AI use cases?(n=352)The most interesting data comes from analyzing how these companies are competing for AI use workloads.Hyperscalers,in general,are leading the race to build new and competing models

130、to power AI.They have the resources,depth of skill sets and money to make this investment.From the user perspective,it is an enticing message to shift AI workloads and often other data and analytics to these platforms,all offering varying degrees of an integrated approach to simplifying AI.37%34%30%

131、18%18%12%11%9%8%38%30%22%24%16%16%9%8%7%32%24%25%15%20%19%15%9%9%24%31%39%50%53%58%70%76%77%Microsoft AzureAmazon Web Services(AWS)Google Cloud Platform(GCP)Salesforce CloudIBM Cloud(watson)Oracle CloudAlibaba CloudBaidu CloudTencent CloudUse today for AIUse for data and analyticsAdding for AI use c

132、asesDont Use 21 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY 37%of respondents are already using Azure for AI today,and AWS is close behind with 34%.There is a significant gap between Azure and third place with GCP coming in at 30%.High Readiness respondents

133、 are closely aligned with these results except for GCP,where we see higher usage at 37.5%.The final measurement of success focuses on where companies are adding new AI use cases.Azure remains in the lead,with 32%of respondents growing their AI footprint on this platform.AWS and GCP are falling behin

134、d,with a significant gap of up to 8 percentage points.This gap is not helped by High Readiness respondents as their planned growth is mainly focused on Azure.Taking the overall total addressable market(TAM)for cloud AI into consideration,estimated to be well over 62 billion dollars,this gap should b

135、e concerning for the market laggards.Oracle,IBM and Salesforce are in the middle of the pack,and our conversations with end users and vendors indicate they remain focused on enabling their existing clients with greater value through AI across their platforms.Data and the Dreaded Model Choices Data i

136、s King Structured data(61%)leads the data type rankings.It is followed closely by real-time data(56%,with time series at 43%and sensor data at 29%)although streaming data comes in near the bottom with 27%.The surprise here is that geospatial data was cited less frequently than any other response.Any

137、one who has seen vendor demos in use case storiesthey have been hard to avoid at conferenceswill recall how frequently this data is at the heart of solutions in many categories.It is also worth noting that despite years of hype,commercial product development and marketing,and widespread sales effort

138、s to address it,unstructured(and/or semi-structuredthey are often conflated)data is only cited by a third of respondents.The story for many organizations will begin with their structured dataand that will actually be likely to accelerate progress because of the metadata,quality management functions

139、and governance capabilities provided by the database management software used for it.Figure 16:What data types are critical to your AI innovation?(n=335)61%56%43%36%34%33%29%27%24%1%Structured dataReal-time dataTime-series dataImage,video,sound dataSemi-structured dataUnstructured dataSensor dataStr

140、eaming dataGeospatial dataOther 22 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Datas role in these projects is central for obvious reasons,not least of which is how it informs,enriches and validates models.But models themselves are data,and the processes as

141、sociated with them mirror those in broad use for data today.Organizations can leverage this knowledge and skill set to their advantage as they initiate a new layer of architecture.Biggest Doesnt Always Equal Best with Models We are witnessing a global arms race initiated by OpenAI with ChatGPT in th

142、e LLM and foundational model sector.The number of models available has grown exponentiallynew models premier daily on Hugging Face and other centralized distribution points.The Hugging Face community hosts over 400,000 models to browse.In a short number of months,it has become clear that there will

143、not be“one model to rule them all.”Our respondents identified Google Gemini and BARD(renamed to Gemini in February 2024)at a combined 57%as the top choice for LLMs,followed by GPT 3.5 plus 4.0 at 53%.Coding models continue to find a specialized home in many environments,serving development and data

144、science teams at 19%.High Readiness respondents submitted similar responses but showed a significantly higher level of interest in GPT 3.5 and 4.0,with 70%choosing this model,17 percentage points higher than the rest of the respondents.Early adopters are very likely to have this pioneering model in

145、their environment.Only 3%of High Readiness users were still researching what models they plan to use,in comparison to the 12%of standard respondents.Figure 17:Which of the following large language models(LLMs)and foundational models are part of your companys AI strategy?(n=309)53%29%28%19%12%11%10%1

146、0%9%7%6%5%4%GPT 3.5,4.0GeminiBARDCodexNone so farLLama2ClaudeGraniteMistralCoherePaLM2FerretOther 23 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Early adopters are already evolving toward multi-model environments to align the work with the model capabilitie

147、s,and it is also clear that there will be a variety of model sizes in most sophisticated environments.Companies are aligning narrow models to specific work and medium and larger-performing models to cover wider-scoped projects.Specialty Models in Your Ecosystem 21%of respondents indicated that domai

148、n-specific models are not a priority at this time.However,that number drops when analyzing High Readiness respondents.Only 16%of these users are not planning to leverage domain-specific models.Finance models at 40%and code models to assist developers at 39%have generated significant interest with st

149、andard respondents and are even more popular with the High Readiness group,where finance models are at 55%and code models are at 49%.Again,the surprise was that medical(13%)and biomedical(10%)came in lowest for the standard group and nearly the same for the High Readiness cohort.Figure 18:Which indu

150、stry or domain-specific models are included in your AI model strategy?(n=335)Model Management is a Critical Strategy for Success Effectively managing AI models allows companies to better manage quality and accuracy,leading to stronger decision-making and customer experiences.In a world of significan

151、t regulation and compliance pressures,managing and understanding the model and the information within a model is a requirement for growing and innovating with AI.Monitoring and maintenance of models is a foundational capability.For overall model management,survey respondents were equally split on ex

152、ecuting that work in-house(33%),on third-party platforms(32%)or a hybrid of both(31%).Respondents defined as High Readiness are managing monitoring and maintenance primarily from a hybrid mix at a rate of 42%,15 percentage points more than standard respondents.This is similar to the results for mode

153、l benchmarking,where 30%are executing this in-house,32%are relying on third-party platforms and 30%cite hybrid strategies.Meanwhile,High Readiness respondents are 8 percentage points more likely to leverage a hybrid strategy.40%39%25%21%21%13%13%13%10%4%FinanceCodeScienceLaw/legalDomain-specific mod

154、els are not a priorityInsuranceClimateMedicalBiomedicalOther 24 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Prompt engineering,the skill of creating and refining input prompts to achieve optimal model outputs,is a growing competency.Highly proficient prompt

155、s can drive quality,accuracy and overall output relevance.For companies that include prompt engineering in their plans,the majority(35%)are doing this work in-house.High Readiness respondents leverage an in-house strategy at a rate of 38%.RAG or retrieval-augmented generation is the technique of ret

156、rieving information from often proprietary enterprise sources to enhance the“knowledge”of large language models(LLMs)and decrease“hallucinations.”Coupling this highly valuable and nuanced corporate information with the model allows it to respond in a more relevant and diverse way.Survey respondents(

157、35%)are primarily relying on third-party solutions to accomplish this task.14%of respondents have yet to deploy this sophisticated approach to model management,and that is likely driven by the use cases and adoption rates of these respondent companies.High Readiness respondents are only slightly ahe

158、ad in RAG adoption but are more likely to execute it in-house or in a hybrid configuration.Fine-tuning allows companies to adapt pre-trained AI models to their specific needs and contexts.This process is important because it enables businesses to tailor the models responses,style and focus to align

159、with their brand,objectives and customer expectations.By fine-tuning models on proprietary data or specific use cases,companies can significantly improve the performance,relevance and effectiveness of AI applications,leading to better outcomes and higher customer satisfaction.Most survey respondents

160、,including High Readiness respondents,do this continuously or sporadically as required.The largest segment(34%)rely on third-party platforms for this type of work.There is a significant cost involved that needs to be considered in ongoing AI strategies when utilizing RAG and fine-tuning.Compute reso

161、urces,data preparation,storage,integration and API calls are all part of this scenario.Utilizing third-party platforms,i.e.,cloud environments,can result in surprising bills.The return on investment(ROI)of models that produce fewer hallucinations and highly relevant responses is clear.This needs to

162、be balanced with the costs of achieving it.Figure 19:How will your organization manage LLMs?(n=289)35%33%30%30%29%26%24%28%32%34%32%32%31%35%26%31%29%30%34%30%27%11%47%9%512%14%Prompt engineeringMonitoring and maintenanceModel fine-tuning(periodic,continuous)Model benchmarkingModel trainingOutput dr

163、iftRetrieval augmented generationIn-houseOn a third-party platformHybridNot in plans 25 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Methodology and Demographics This research was designed to explore the impact of AI and GenAI on enterprise companies.Its goa

164、l is to understand better how prepared these firms are to take advantage of this disruptive technology,with a specific focus on how users are optimizing their technology architectures to leverage AI.The study employs a quantitative approach utilizing survey responses from a qualified panel of respon

165、dents.The design of the survey instrument allows for a detailed exploration of participant subject matter expertise.It focuses on AI maturity,challenges to success,funding,strategy,technology gaps and strategies,and other critical components of leveraging this type of technology.The respondent popul

166、ation is global with significant European and North American audiences,and a total sample size of 335 completed surveys.Random sampling was used to ensure representation across different job roles,company sizes and industries.Participants were included based on their knowledge of their companys AI s

167、trategy and projects.Those with no project experience were excluded.Respondent Panel Survey results were drawn from business and IT functional professionals whose job titles included CXO,VP/EVP/SVP,Senior Director,Director,Manager,Engineer and Analyst.Respondents represent a wide variety of industri

168、es,including information technology,manufacturing,retail,financial services,healthcare and others.Respondents represent companies of all sizes based on company worldwide employee size and annual revenue.Figure 20:In which region are you located?(n=335)73%24%1%1%1%North AmericaEuropeSouth AmericaAsia

169、 and PacificROW 26 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Figure 21:Select the primary sector served by your organization(n=335)Figure 22:How many employees does your organization have worldwide?(n=335)23%18%15%12%9%6%5%4%3%3%2%ITManufacturingServicesR

170、etail/WholesaleFinancial ServicesHealthcareOtherPublic sector and EducationTelecommunicationsTransportUtilities23%10%11%13%15%13%5%10%Less than 250250-499500-9991,000-2,4992,500-4,9995,000-9,99910,000-19,99920,000 or more 27 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARK

171、ET STRATEGY Figure 23:What is the annual revenue range that best fits your organization?(n=317)Figure 24:What best describes your job level?(n=335)The survey was administered online,and data collection occurred over a three-week period starting in February and ending in March 2024.The survey was als

172、o quality-tested in the field,with a test phase to refine clarity and ensure response quality.The survey instrument was designed by Shawn Rogers and Merv Adrian and deployed by BARC GmbH.20%12%9%17%13%18%12%Less than$10 million$10-$49.9 million$50-$99.9 million$100-$499.9 million$500-$999.9 million$

173、1-$5 billionOver$5 billion31%24%21%7%5%5%3%3%0%ManagerExecutive:CxOVP/DirectorEngineerAnalystArchitectOtherConsultantVendor 28 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Authors Shawn Rogers CEO BARC,US,and BARC Fellow For over 28 years,Shawn Rogers has be

174、en an internationally respected industry analyst,thought leader,speaker,author and instructor on data,business intelligence,analytics,AI/ML and cloud technologies.His former executive strategy roles with Dell,Statistica,Quest Software and TIBCO Software give him a unique perspective on the enterpris

175、e software industry.As a CEO of BARC US and BARC Fellow,he is responsible for analyst coverage of these technologies in the North American market.Merv Adrian Founder and Principal,IT Market Strategy Merv Adrian,founder and principal at IT Market Strategy,is a trusted advisor to leading software firm

176、s and investors,a frequent conference speaker,and a prolific researcher,editor and writer on IT issues.He is a former Research Vice President of Gartner,Senior Vice President of Forrester Research and Research Manager at Giga Information Group,specializing in data management.Mr.Adrian began his care

177、er at the Federal Reserve Bank and was an independent programmer and consultant before moving into the software industry as a technical journal editor and strategic marketer with Information Builders and Sybase before becoming an analyst in 1997.29 Optimizing Your Architecture for AI Innovation,Marc

178、h 2024 BARC GmbH,IT MARKET STRATEGY About BARC Research BARC studies are based on internal market research,software tests and analyst comments,giving you the security to make the right decisions.Our independent research brings market developments into clear focus,puts software and vendors through th

179、eir paces and gives users a place to express their opinions.Events Decision-makers and IT industry leaders come together at BARC events.BARC seminars in small groups,online webinars and conferences with more than 1,000 participants annually all offer inspira-tion and interactivity.Through exchange w

180、ith peers and an overview of current trends and market de-velopments,you will receive new impetus to drive your business forward.Consulting In confidential expert workshops,coaching and in-house consultations,we transform the needs of your company into future-proof decisions.We provide you with succ

181、essful,holistic concepts that ena-ble you to use the right information correctly.Our project support covers all stages of the successful use of software.BARC(Business Application Research Center)is one of Europes leading ana-lyst firms for business software,focusing on the areas of data,business int

182、elli-gence(BI)and analytics,enterprise content management(ECM),customer re-lationship management(CRM)and enterprise resource planning(ERP).Our passion is to help organizations become digital companies of tomorrow.We do this by using technology to rethink the world,trusting data-based deci-sions and

183、optimizing and digitalizing processes.Its about finding the right tools and using them in a way that gives your company the best possible ad-vantage.This unique blend of knowledge,exchange of information and inde-pendence distinguishes our services in the areas of research,events and con-sulting.30

184、Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY About Aerospike Aerospike is the real-time,multi-model database built for infinite scale,speed,and savings.Blazing fast and reliable,Aerospike performs on gigabytes to petabytes of data with sub-millisecond latenc

185、y,so enterprises can make better decisions faster with new and constantly changing data.Built for efficiency and sustainability from the start,Aerospike enables organizations to operate on a fraction of the infrastructure required from legacy databases,helping them reduce their server footprint by 8

186、0 percent to drive lower all-around costs and significantly reduce carbon emissions.Aerospike customers are ready for whats next with the lowest latency and the highest throughput data platform and no re-platforming required even as their data grows.Aerospike serves organizations as a catalyst for p

187、rogress,believing that there are no limits to how businesses can innovate with their data.Cloud and AI-forward,Aerospike empowers businesses like Adobe,Airtel,Criteo,DBS Bank,Experian,PayPal,Snap,and Sony Interactive Entertainment to optimize for the cloud and expand their use of transformative AI a

188、pplications.Aerospike is headquartered in Mountain View,California,with offices in London,Bangalore,and Tel A.Contact info Contact info Aerospike 2440 W.El Camino Real,Suite 100 Mountain View,CA 94040 (+1)408-462-AERO(2376) 31 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MA

189、RKET STRATEGY About Boomi Productivity.Progress.Innovation.Growth.In Boomis world,nothing gets in their way.So data can flow,business can grow,and people can thrive.Inventing iPaaS was the rock-solid foundation for Boomi to build the global SaaS company thats proud to have 800+partners in its networ

190、k.The likes of Accenture,Capgemini,Deloitte,and SAP all choose Boomi,its people,and its award-winning platform to connect everyone to everything,anywhere and everywhere.And with Boomi AI,nothing can stop apps,data,processes,and people doing everything they need to do.End-to-end is just the beginning

191、.Its time for a world where complexity cant get in the way anymore.Boomi powers the future of business with intelligent integration and automation.As a category-leading,global software as a service(SaaS)company,Boomi celebrates more than 20,000 global customers and a worldwide network of 800 partner

192、s.Organizations turn to Boomis award-winning platform to connect their applications,data,and people to accelerate digital transformation.For more information,visit .Contact info Contact info Boomi,LP 1 W Elm St,Suite 200 Conshohocken,PA 19428 +1-800-732-3602 32 Optimizing Your Architecture for AI In

193、novation,March 2024 BARC GmbH,IT MARKET STRATEGY About Domino Domino Data Lab empowers the largest AI-driven enter-prises to build and operate AI at scale.Dominos Enterprise AI platform unifies the flexibility AI teams want with the vis-ibility and control the enterprise requires.Domino enables a re

194、peatable and agile ML lifecycle for faster,responsible AI impact with lower costs.With Domino,global enterprises can develop better medi-cines,grow more productive crops,develop more competi-tive products,and more.Founded in 2013,Domino is backed by Sequoia Capital,Coatue Management,NVIDIA,Snowflake

195、,and other leading investors.Learn more at www.domino.ai.Contact info Contact info www.domino.ai 33 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY About InterSystems InterSystems is a leading provider of next-generation solu-tions for enterprise digital transf

196、ormations that provide a foundation for artificial intelligence in the healthcare,fi-nance,manufacturing,and supply chain sectors.Its cloud-first data platforms solve interoperability,speed,and scala-bility problems for large organizations around the globe.InterSystems offers a range of software,inc

197、luding its flag-ship data platform software,InterSystems IRIS,and a set of industry specific applications that speed and simplify the process of making data“AI ready”.InterSystems is commit-ted to excellence through its award-winning,247 support for customers and partners in more than 80 countries.P

198、rivately held and headquartered in Cambridge,Massachusetts,InterSystems was established in 1978 and now has 39 offices in 28 countries worldwide.For more information,please visit InterS.Contact info Contact info Global Headquarters One Memorial Drive,Cambridge,MA 02142,USA Main Office:+1.617.621.060

199、0 InterS 34 Optimizing Your Architecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY About Qlik Qlik stands at the forefront of data integration and analytics,propelling businesses into a new era of informed decision-making and strategic agility.With a heritage rooted in innovation,Qli

200、k has grown to serve over 40,000 customers worldwide,transcending industries and geographies.Our comprehensive portfolio,underpinned by advanced AI and machine learning technologies,emphasizes not just the acquisition of data,but its transformation into actionable insights,driving significant busine

201、ss outcomes.Our technology is distinguished by its robust data integration and governance capabilities,ensuring data is not only accessible but trustworthy,relevant,and timely.This foundation enables our intuitive analytics platform to uncover hidden patterns,providing businesses with the clarity an

202、d foresight needed to navigate complex challenges and capitalize on emerging opportunities.Qliks approach is decidedly customer-centric,prioritizing flexibility and platform agnosticism to avoid vendor lock-in,thereby empowering businesses to adapt and evolve their AI strategies with changing market

203、 dynamics.This ethos extends to our commitment to education and empowerment,equipping our clients with the tools and insights necessary for a data-driven transformation.In an era where data is the most valuable asset,Qlik aims to turn complex data landscapes into a strategic advantage,making busines

204、ses more competitive and resilient.Our commitment to innovation,customer success,and ethical data use positions Qlik as a strategic partner in the journey toward AI-enabled transformation.Contact info Contact info QlikTech Inc.211 S.Gulph Road,Suite 500 King of Prussia,PA 19406 35 Optimizing Your Ar

205、chitecture for AI Innovation,March 2024 BARC GmbH,IT MARKET STRATEGY Data Decisions.Built on BARC Germany BARC GmbH Berliner Platz 7 D-97080 Wrzburg+49 931 880651-0 Austria BARC GmbH Hirschstettner Strae 19/I/IS314 A-1220 Wien+43 660 6366870 Switzerland BARC Schweiz GmbH Tfernstr.22a CH-5405 Baden-Dttwil+41 56 470 94 34 United States BARC US 13463 Falls Drive Broomfield,CO.80020 USA+01 720-381-4988 United States 39654 Gladiolus Ln Palm Desert,CA 92211