Is your organization currently implementing or planning any AI initiatives (e.g., evaluations, pilots, development, deployments) within the next 12 months?

• Yes (n=538)
  100%
• No (n=0)
  0%

Which of the following best describes your involvement and responsibilities across your organization’s overall approach to AI?

• Strategic Owner (n=126)
  23.4%
• Decision Lead (n=163)
  30.3%
• Technical Sponsor (n=100)
  18.6%
• Contributor / Implementer (n=89)
  16.5%
• Observer / Researcher (n=60)
  11.2%
• None of these describe my role (n=0)
  0%

Please indicate if you have a significant level of involvement in making or influencing decisions with specific elements of the AI Stack.

• Data and Integration: (Data Infrastructure, Databases) (n=396)
  73.6%
• Model Development and Training: (ML Frameworks, Models) (n=295)
  54.8%
• Model Hosting and Orchestration: (Agentic Platforms, Cloud Platforms) (n=269)
  50%
• Application Development and Middleware: (SDKs, libraries, MCP, A2A, Safety, compliance) (n=284)
  52.8%
• AI Observability and Governance: (AI security, Monitoring for drift, bias, hallucinations, cost) (n=329)
  61.2%
• AI Business Applications: (Prepackaged AI apps, Copilots) (n=296)
  55%
• I’m not involved with any of the above (n=0)
  0%

Which of the following best describes the industry which your organization serves?

• Agriculture (farms; aquaculture; forestry) (n=21)
  3.9%
• Banking & Financial Services (banks; insurance; asset managers; payments; FinTech) (n=59)
  11%
• Construction & Real Estate (construction; property management; REITs) (n=41)
  7.6%
• Energy & Utilities (oil & gas; power; renewables; water/waste) (n=35)
  6.5%
• Healthcare Services (hospitals; clinics; long-term care) (n=41)
  7.6%
• Life Sciences & MedTech (pharma; biotech; medical devices) (n=26)
  4.8%
• Manufacturing & Basic Materials (auto; aerospace; machinery; chemicals; food & beverage) (n=53)
  9.9%
• Mining & Minerals (extraction; mining; quarrying; mineral processing) (n=19)
  3.5%
• Professional & Consumer Services (corporate legal/accounting/consulting/marketing; consumer tax prep; counselling; personal care) (n=46)
  8.6%
• Retail & Consumer Goods (brick & mortar; e-commerce; CPG; apparel; grocery) (n=53)
  9.9%
• Software, Cloud & AI (SaaS; platforms; GenAI/AgenticAI; mobile & desktop apps) (n=46)
  8.6%
• Technology & Hardware (chips; servers; devices; OEMs; device firmware) (n=60)
  11.2%
• Telecom, Data Centers & Infrastructure (carriers; ISPs; data centers; hyperscalers; edge providers; WAN/SD-WAN) (n=33)
  6.1%
• Other (Specify) (n=5)
  0.9%

In which country are you personally located?

• Australia (n=17)
  3.2%
• India (n=19)
  3.5%
• Indonesia (n=12)
  2.2%
• Japan (n=14)
  2.6%
• Malaysia (n=11)
  2%
• New Zealand (n=8)
  1.5%
• Singapore (n=18)
  3.3%
• South Korea (n=13)
  2.4%
• Other AsiaPAC (not listed) (n=0)
  0%
• China (n=0)
  0%
• Hong Kong (n=0)
  0%
• Macau (n=0)
  0%
• Taiwan (n=0)
  0%
• France (n=19)
  3.5%
• Germany (n=18)
  3.3%
• Israel (n=6)
  1.1%
• Italy (n=14)
  2.6%
• Nordics (Denmark, Finland, Iceland, Norway, Sweden) (n=11)
  2%
• Poland (n=10)
  1.9%
• Russia (n=8)
  1.5%
• Saudi Arabia (n=10)
  1.9%
• Spain (n=15)
  2.8%
• UAE (n=12)
  2.2%
• Ukraine (n=0)
  0%
• United Kingdom (n=35)
  6.5%
• Other EMEA (not listed) (n=0)
  0%
• Canada (n=40)
  7.4%
• Mexico (n=33)
  6.1%
• United States (n=87)
  16.2%
• Argentina (n=30)
  5.6%
• Brazil (n=39)
  7.2%
• Colombia (n=25)
  4.6%
• Panama (n=14)
  2.6%
• Other South/Central America (not listed) (n=0)
  0%
• Other, Not Listed (n=0)
  0%

Please estimate the total number of employees across your entire company or organization (worldwide).

• 2 to 999 employees (n=91)
  16.9%
• 1,000 to 2,499 employees (n=104)
  19.3%
• 2,500 to 4,999 employees (n=97)
  18%
• 5,000 to 9,999 employees (n=80)
  14.9%
• 10,000 to 19,999 employees (n=76)
  14.1%
• 20,000 to 49,999 employees (n=48)
  8.9%
• 50,000 or more employees (n=42)
  7.8%
• Unsure (n=0)
  0%

Which of the following best describes your department or functional role?

• Customer Service & Support (n=40)
  7.4%
• Executive Leadership & Administration (n=32)
  5.9%
• Finance & Accounting (n=50)
  9.3%
• Human Resources & Employee Success (n=29)
  5.4%
• Information Security (IS) (n=56)
  10.4%
• Information Technology (IT) (n=113)
  21%
• Legal and Compliance (n=21)
  3.9%
• Marketing (n=44)
  8.2%
• Operations (or Delivery/Production) (n=81)
  15.1%
• Research & Development (R&D) or Product Development (n=43)
  8%
• Sales (n=29)
  5.4%
• Other (Not listed) (n=0)
  0%

Which title best describes your 'leadership' responsibilities within your organization?

• C-level executive (e.g., CXO, CMO, CIO, CTO) (n=74)
  13.8%
• Senior Management (SVP, EVP, VP, or Business Unit Lead in charge of one or more large departments or divisions) (n=93)
  17.3%
• Mid Leadership (directing or managing multiple teams, managers, or high-level contributors) (n=99)
  18.4%
• Director, Manager, or Team Lead (managing or directing an operational or specialized) (n=136)
  25.3%
• Project Manager (leading ad hoc project or operational teams and/or initiatives) (n=66)
  12.3%
• Practitioner / Specialist (working within a team or as an individual contributor) (n=48)
  8.9%
• Consultant / Advisor (assisting other teams or leadership) (n=22)
  4.1%
• Other (not listed) (n=0)
  0%

Which of the following best describes your primary 'technical' responsibilities within your organization?

• Machine learning (ML) engineer: Building, deploying, and maintaining ML models at the core of the AI stack. (n=47)
  8.7%
• Data scientist: Designing models, algorithms, and experiments that drive the AI solution. (n=44)
  8.2%
• Data architect: Designing enterprise data frameworks, including data ingestion, storage, and processing, which are foundational to any AI stack. (n=35)
  6.5%
• Data engineer: Building and maintaining the data pipelines and infrastructure that feed and support ML models. (n=54)
  10%
• Senior Systems/Platform Engineer: Managing or operating platforms and systems that host and run AI applications and models. (n=47)
  8.7%
• Site Reliability Engineer (SRE): Ensuring the reliability, performance, and scaling of AI models in production. (n=17)
  3.2%
• Director of Cloud/Hybrid Cloud: Making key decisions about the cloud infrastructure, the primary environment for most modern AI stacks. (n=52)
  9.7%
• Enterprise architect: Defining the overall structure and standards for IT systems, ensuring the AI stack aligns with the broader enterprise technology landscape. (n=35)
  6.5%
• Security/SecOps engineer: Securing the sensitive data, models, and deployment pipelines within the AI stack. (n=24)
  4.5%
• Overall AI Stack oversight: Providing the highest-level strategic direction and budget approval for all elements within and supporting the AI stack (including enabling technology and infrastructure). (n=88)
  16.4%
• Oversight of AI-enabling IT: Oversight of the technologies and IT teams enabling AI deployment and adoption. (n=52)
  9.7%
• Oversight of AI-enabling infrastructure: Oversight of the critical infrastructure within IT (compute, networking, data storage, etc.) that underpins the entire AI Stack. (n=43)
  8%
• Other (Not involved in any of the above roles). My technical responsibilities are outside these areas (n=0)
  0%

Which of the following best describes your organization’s overall approach to adopting new technologies?

• Innovator (always first to deploy) (n=87)
  16.2%
• Early Adopter (taking calculated risks) (n=131)
  24.3%
• Mass Market Adopter (look for proven business value) (n=180)
  33.5%
• Late Adopter (deploy to remain competitive) (n=89)
  16.5%
• Laggard (low cost, no risk approach) (n=51)
  9.5%

Specifically regarding AI, what stage of implementation or maturity is your organization at today and where do you estimate it will be in the future?

Currently

• Learning (n=44)
  8.2%
• Planning (n=59)
  11%
• Experimenting (n=142)
  26.4%
• Developing (n=130)
  24.2%
• Initial Deployments (n=110)
  20.4%
• Mass Deployments (n=53)
  9.9%

Within 6 months

• Learning (n=4)
  0.7%
• Planning (n=29)
  5.4%
• Experimenting (n=55)
  10.2%
• Developing (n=125)
  23.2%
• Initial Deployments (n=176)
  32.7%
• Mass Deployments (n=149)
  27.7%

Within 6 - 12 months

• Learning (n=1)
  0.2%
• Planning (n=19)
  3.5%
• Experimenting (n=36)
  6.7%
• Developing (n=71)
  13.2%
• Initial Deployments (n=142)
  26.4%
• Mass Deployments (n=269)
  50%

Within 12 - 24 months

• Learning (n=0)
  0%
• Planning (n=4)
  0.7%
• Experimenting (n=22)
  4.1%
• Developing (n=52)
  9.7%
• Initial Deployments (n=103)
  19.1%
• Mass Deployments (n=357)
  66.4%

Please identify the priority or importance in achieving the following objectives or outcomes through the adoption of AI.

Improved operational efficiency and process automation

• Primary (within 12 months) (n=388)
  72.1%
• Secondary (within 12-24 months) (n=125)
  23.2%
• Tertiary (not a focus) (n=25)
  4.6%

Enhanced decision-making through advanced analytics and insights

• Primary (within 12 months) (n=211)
  39.2%
• Secondary (within 12-24 months) (n=273)
  50.7%
• Tertiary (not a focus) (n=54)
  10%

Accelerated innovation and new product or service development

• Primary (within 12 months) (n=214)
  39.8%
• Secondary (within 12-24 months) (n=263)
  48.9%
• Tertiary (not a focus) (n=61)
  11.3%

Enhanced customer engagement and personalization

• Primary (within 12 months) (n=292)
  54.3%
• Secondary (within 12-24 months) (n=211)
  39.2%
• Tertiary (not a focus) (n=35)
  6.5%

Reduced costs and improved margin performance

• Primary (within 12 months) (n=279)
  51.9%
• Secondary (within 12-24 months) (n=183)
  34%
• Tertiary (not a focus) (n=76)
  14.1%

Improved employee productivity, acquisition, or retention through AI tools or copilots

• Primary (within 12 months) (n=311)
  57.8%
• Secondary (within 12-24 months) (n=146)
  27.1%
• Tertiary (not a focus) (n=81)
  15.1%

Strengthened risk management, governance, and compliance

• Primary (within 12 months) (n=226)
  42%
• Secondary (within 12-24 months) (n=247)
  45.9%
• Tertiary (not a focus) (n=65)
  12.1%

Development of new AI-enabled business models or revenue streams

• Primary (within 12 months) (n=160)
  29.7%
• Secondary (within 12-24 months) (n=226)
  42%
• Tertiary (not a focus) (n=152)
  28.3%

Improved competitive advantage

• Primary (within 12 months) (n=223)
  41.4%
• Secondary (within 12-24 months) (n=236)
  43.9%
• Tertiary (not a focus) (n=79)
  14.7%

Established market awareness as a leading AI-enabled business

• Primary (within 12 months) (n=153)
  28.4%
• Secondary (within 12-24 months) (n=283)
  52.6%
• Tertiary (not a focus) (n=102)
  19%

Realizing payback on initial AI investments

• Primary (within 12 months) (n=268)
  49.8%
• Secondary (within 12-24 months) (n=211)
  39.2%
• Tertiary (not a focus) (n=59)
  11%

Achieving an ROI of 15-20 percent or higher

• Primary (within 12 months) (n=110)
  20.4%
• Secondary (within 12-24 months) (n=284)
  52.8%
• Tertiary (not a focus) (n=144)
  26.8%

Improved strategic alignment or innovation potential (even with delayed ROI)

• Primary (within 12 months) (n=140)
  26%
• Secondary (within 12-24 months) (n=292)
  54.3%
• Tertiary (not a focus) (n=106)
  19.7%

Other (please specify)

• Primary (within 12 months) (n=2)
  0.4%
• Secondary (within 12-24 months) (n=2)
  0.4%
• Tertiary (not a focus) (n=1)
  0.2%

Of the objectives you have noted as Primary, which do you consider the most important to achieve?

• Improved operational efficiency and process automation (n=143)
  26.6%
• Enhanced decision-making through advanced analytics and insights (n=52)
  9.7%
• Accelerated innovation and new product or service development (n=69)
  12.8%
• Enhanced customer engagement and personalization (n=47)
  8.7%
• Reduced costs and improved margin performance (n=33)
  6.1%
• Improved employee productivity, acquisition, or retention through AI tools or copilots (n=52)
  9.7%
• Strengthened risk management, governance, and compliance (n=23)
  4.3%
• Development of new AI-enabled business models or revenue streams (n=29)
  5.4%
• Improved competitive advantage (n=32)
  5.9%
• Established market awareness as a leading AI-enabled business (n=10)
  1.9%
• Realizing payback on initial AI investments (n=23)
  4.3%
• Achieving an ROI of 15-20 percent or higher (n=14)
  2.6%
• Improved strategic alignment or innovation potential (even with delayed ROI) (n=11)
  2%
• Other (n=0)
  0%

Which of the following best describes your primary (or planned) approach to adopting AI in the enterprise?

• Buy AI (Embedded): We use AI that is already built into the software and cloud services we purchase (e.g., AI inside our SaaS applications). (n=156)
  29%
• Customize or Adapt AI: We take existing, pre-built AI models (open source or commercial) and customize them using our own company data. (n=237)
  44.1%
• Build AI (Proprietary): We build our own unique AI models from the ground up to maintain complete control over our systems. (n=79)
  14.7%
• Explore AI: We’re actively evaluating AI solutions and have not yet committed to a primary implementation strategy. (n=66)
  12.3%

Which is your organization’s primary approach to selecting/using LLMs?

• Open-source LLMs only (e.g., LLaMA, Falcon) (n=84)
  15.6%
• Proprietary/commercial LLMs only (e.g., OpenAI, Google, Anthropic) (n=142)
  26.4%
• Primarily open-source but some proprietary LLMs (n=114)
  21.2%
• Primarily proprietary but some open-source LLMs (n=105)
  19.5%
• An even balance of open-source and proprietary LLMs (n=72)
  13.4%
• We haven’t made that decision yet (n=21)
  3.9%

What are your plans (timeframes) for using or deploying LLMs to support the following business functions?

Customer service/support (chatbots, virtual assistants, helpdesk automation)

• Currently in use (n=250)
  46.5%
• Within 6 months (n=124)
  23%
• Within 6 - 12 months (n=85)
  15.8%
• Within 12 - 24 months (n=47)
  8.7%
• No Plans (n=32)
  5.9%

Knowledge management/research (market research, internal documentation, insights generation etc.)

• Currently in use (n=203)
  37.7%
• Within 6 months (n=196)
  36.4%
• Within 6 - 12 months (n=83)
  15.4%
• Within 12 - 24 months (n=30)
  5.6%
• No Plans (n=26)
  4.8%

Code generation/software development support (DevOps, automation, code assistants, etc.)

• Currently in use (n=164)
  30.5%
• Within 6 months (n=177)
  32.9%
• Within 6 - 12 months (n=108)
  20.1%
• Within 12 - 24 months (n=55)
  10.2%
• No Plans (n=34)
  6.3%

IT management and operations (IT helpdesk, operations management, data or device security, etc.)

• Currently in use (n=211)
  39.2%
• Within 6 months (n=171)
  31.8%
• Within 6 - 12 months (n=104)
  19.3%
• Within 12 - 24 months (n=29)
  5.4%
• No Plans (n=23)
  4.3%

Financial operations (accounting systems, fraud detection, etc.)

• Currently in use (n=126)
  23.4%
• Within 6 months (n=210)
  39%
• Within 6 - 12 months (n=113)
  21%
• Within 12 - 24 months (n=64)
  11.9%
• No Plans (n=25)
  4.6%

Content creation/marketing automation

• Currently in use (n=229)
  42.6%
• Within 6 months (n=151)
  28.1%
• Within 6 - 12 months (n=91)
  16.9%
• Within 12 - 24 months (n=40)
  7.4%
• No Plans (n=27)
  5%

Other (Specify)

• Currently in use (n=1)
  0.2%
• Within 6 months (n=1)
  0.2%
• Within 6 - 12 months (n=0)
  0%
• Within 12 - 24 months (n=0)
  0%
• No Plans (n=0)
  0%

Please indicate if you agree or disagree with the following statements:

Our organization has a formal process for evaluating, approving, and retiring foundation models

• Strongly Agree (n=155)
  28.8%
• Agree (n=220)
  40.9%
• Neutral/Unsure (n=80)
  14.9%
• Disagree (n=52)
  9.7%
• Strongly Disagree (n=31)
  5.8%

We anticipate having multiple foundation models concurrently deployed

• Strongly Agree (n=203)
  37.7%
• Agree (n=153)
  28.4%
• Neutral/Unsure (n=94)
  17.5%
• Disagree (n=53)
  9.9%
• Strongly Disagree (n=35)
  6.5%

We anticipate adding additional or replacing existing models every year

• Strongly Agree (n=199)
  37%
• Agree (n=163)
  30.3%
• Neutral/Unsure (n=109)
  20.3%
• Disagree (n=41)
  7.6%
• Strongly Disagree (n=26)
  4.8%

The ability to add or replace foundation models will provide our business with agility and improve our competitive posture

• Strongly Agree (n=161)
  29.9%
• Agree (n=184)
  34.2%
• Neutral/Unsure (n=118)
  21.9%
• Disagree (n=54)
  10%
• Strongly Disagree (n=21)
  3.9%

AI is strategically important to our organizations overall success

• Strongly Agree (n=246)
  45.7%
• Agree (n=161)
  29.9%
• Neutral/Unsure (n=59)
  11%
• Disagree (n=43)
  8%
• Strongly Disagree (n=29)
  5.4%

Deploying AI and leveraging LLMs will be critical to maintaining our competitive advantage over the coming 1 - 3 years

• Strongly Agree (n=160)
  29.7%
• Agree (n=211)
  39.2%
• Neutral/Unsure (n=87)
  16.2%
• Disagree (n=52)
  9.7%
• Strongly Disagree (n=28)
  5.2%

Agentic AI (autonomous workflows/agents) will play a significant role in our AI strategy over the coming 12 months

• Strongly Agree (n=211)
  39.2%
• Agree (n=194)
  36.1%
• Neutral/Unsure (n=59)
  11%
• Disagree (n=51)
  9.5%
• Strongly Disagree (n=23)
  4.3%

Which of the following best describes your organization’s approach to providing enterprise context to AI systems?

• We primarily use basic document retrieval or RAG (n=124)
  23%
• We use governed retrieval connected to enterprise data sources (n=177)
  32.9%
• We use semantic layers, knowledge graphs, catalogs, or metadata to improve context quality (n=86)
  16%
• We use memory, user/project context, or workflow context across AI applications (n=70)
  13%
• We are evaluating context architecture but have not implemented it (n=62)
  11.5%
• We do not have a formal approach to AI context yet (n=17)
  3.2%
• Unsure (n=2)
  0.4%

Which of the following best describes your organization’s current stage of agentic AI adoption?

• Researching or evaluating agentic AI (n=139)
  25.8%
• Running pilots or proofs of concept (n=172)
  32%
• Deploying agents for limited internal workflows (n=113)
  21%
• Deploying agents in production across multiple workflows (n=57)
  10.6%
• Deploying autonomous or semi-autonomous agents with governance and monitoring (n=33)
  6.1%
• No current plans (n=20)
  3.7%
• Unsure (n=4)
  0.7%

What best describes the current state of your core data architecture (warehouse, lake, or lakehouse)?

• Legacy on-premises data warehouse (n=106)
  19.7%
• Hybrid setup (n=211)
  39.2%
• Cloud-based data lake or lakehouse (n=141)
  26.2%
• Fully modernized AI-ready architecture (n=80)
  14.9%
• Other (please specify) (n=0)
  0%

What is the status of modernizing your organization's data architecture?

• Complete (n=102)
  19%
• In process now (n=194)
  36.1%
• Planning to begin within 6 months (n=115)
  21.4%
• Planning to begin within 6 - 12 months (n=75)
  13.9%
• Waiting at least 1 year or more (n=47)
  8.7%
• No plans to modernize our data architecture (n=5)
  0.9%

How important are the following as drivers for modernizing your data architecture?

Need for better scalability, performance, and real-time access

• Very Important (n=339)
  63.6%
• Somewhat Important (n=171)
  32.1%
• Not Important (n=23)
  4.3%

Base: n=533

Integration and governance of siloed or fragmented data

• Very Important (n=320)
  60%
• Somewhat Important (n=182)
  34.1%
• Not Important (n=31)
  5.8%

Base: n=533

Reduction of infrastructure costs and technical debt

• Very Important (n=243)
  45.6%
• Somewhat Important (n=253)
  47.5%
• Not Important (n=37)
  6.9%

Base: n=533

Compliance, security, and data sovereignty requirements

• Very Important (n=361)
  67.7%
• Somewhat Important (n=143)
  26.8%
• Not Important (n=29)
  5.4%

Base: n=533

Other (please specify)

• Very Important (n=1)
  0.2%
• Somewhat Important (n=1)
  0.2%
• Not Important (n=0)
  0%

Base: n=533

Which of the drivers you’ve selected is the most important or primary driver?

• Need for better scalability, performance, and real-time access (n=162)
  30.4%
• Integration and governance of siloed or fragmented data (n=157)
  29.5%
• Reduction of infrastructure costs and technical debt (n=97)
  18.2%
• Compliance, security, and data sovereignty requirements (n=116)
  21.8%
• Other (n=1)
  0.2%

Base: n=533

To what extent is your data platform designed or ready to support AI/ML workloads?

• Completely (100%) (n=122)
  22.7%
• Mostly (75% there) (n=139)
  25.8%
• Somewhat (50% there) (n=122)
  22.7%
• Barely (less than 50% there) (n=144)
  26.8%
• Not at All (no design or readiness at this time) (n=11)
  2%

Which of the following are (or were) primary barriers to updating your data architecture for AI/ML?

• Data quality/governance (n=191)
  35.5%
• Scalability/performance (n=280)
  52%
• Budget limitations or cost uncertainty (n=249)
  46.3%
• Operationalization/MLOps readiness (n=164)
  30.5%
• Talent or skills gaps (n=243)
  45.2%
• Security/privacy/regulatory (n=193)
  35.9%
• Resource allocation conflicts (n=110)
  20.4%
• Other (Specify) (n=0)
  0%

Which of the barriers you selected is/was the top most significant or challenging to overcome?

• Data quality/governance (n=80)
  14.9%
• Scalability/performance (n=123)
  22.9%
• Budget limitations or cost uncertainty (n=115)
  21.4%
• Operationalization/MLOps readiness (n=75)
  13.9%
• Talent or skills gaps (n=81)
  15.1%
• Security/privacy/regulatory (n=49)
  9.1%
• Resource allocation conflicts (n=15)
  2.8%
• Other (n=0)
  0%

How important are the following in managing and ensuring data quality, governance and lineage for AI models?

Automated data validation and cleansing pipelines

• Very Important (n=332)
  61.7%
• Somewhat Important (n=172)
  32%
• Not Important (n=34)
  6.3%

Centralized data catalog and governance tools

• Very Important (n=313)
  58.2%
• Somewhat Important (n=193)
  35.9%
• Not Important (n=32)
  5.9%

Feature stores or curated datasets

• Very Important (n=252)
  46.8%
• Somewhat Important (n=246)
  45.7%
• Not Important (n=40)
  7.4%

Monitoring and auditing for model input/output

• Very Important (n=327)
  60.8%
• Somewhat Important (n=168)
  31.2%
• Not Important (n=43)
  8%

Other (please specify)

• Very Important (n=1)
  0.2%
• Somewhat Important (n=1)
  0.2%
• Not Important (n=0)
  0%

How well integrated is your data platform with AI/ML operationalization (MLOps/LLMOps) workflows?

• Not integrated - Data platform and AI/ML workflows operate in silos (n=94)
  17.5%
• Partially integrated (less than 50%) (n=178)
  33.1%
• Mostly integrated (50% or more) (n=163)
  30.3%
• Fully integrated - End-to-end automation connecting data platform with MLOps/LLMOps, including monitoring and model retraining (n=103)
  19.1%

Which cloud or on-premises infrastructure providers are you using today to run AI workloads and how might that change ahead?

Today

• Amazon Web Services - AWS (n=316)
  58.7%
• Microsoft Azure (n=275)
  51.1%
• Google Cloud - GCP (n=208)
  38.7%
• Oracle Cloud Infrastructure (n=92)
  17.1%
• IBM Cloud (n=91)
  16.9%
• Alibaba (n=70)
  13%
• Tencent (n=39)
  7.2%
• Neo Cloud provider, such as (CoreWeave, Vultr or OVHCloud) (n=12)
  2.2%
• Dell (n=161)
  29.9%
• HPE (n=68)
  12.6%
• Lenovo (n=52)
  9.7%
• Cisco (n=66)
  12.3%
• SuperMicro (n=54)
  10%
• Other (n=0)
  0%

12-24 months from now

• Amazon Web Services - AWS (n=250)
  46.5%
• Microsoft Azure (n=258)
  48%
• Google Cloud - GCP (n=178)
  33.1%
• Oracle Cloud Infrastructure (n=122)
  22.7%
• IBM Cloud (n=76)
  14.1%
• Alibaba (n=97)
  18%
• Tencent (n=48)
  8.9%
• Neo Cloud provider, such as (CoreWeave, Vultr or OVHCloud) (n=42)
  7.8%
• Dell (n=161)
  29.9%
• HPE (n=65)
  12.1%
• Lenovo (n=101)
  18.8%
• Cisco (n=68)
  12.6%
• SuperMicro (n=64)
  11.9%
• Other (n=0)
  0%

Do you have a dedicated AI governance committee or cross-functional team responsible for AI strategy and oversight?

• Yes (n=230)
  42.8%
• Not yet, but we plan to (within 6 months) (n=151)
  28.1%
• Not yet, but we plan to (within 6 - 12 months) (n=87)
  16.2%
• No, and there are no plans yet (n=46)
  8.6%
• Don’t know, Unsure (n=24)
  4.5%

Which organization is currently (or will be) the primary leader or driver of your AI committee or team?

• Customer Service & Support (n=16)
  3.4%
• Executive Leadership & Administration (n=88)
  18.8%
• Finance & Accounting (n=29)
  6.2%
• Human Resources & Employee Success (n=15)
  3.2%
• Information Security (IS) (n=47)
  10%
• Information Technology (IT) (n=131)
  28%
• Legal and Compliance (n=35)
  7.5%
• Marketing (n=10)
  2.1%
• Operations (or Delivery/Production) (n=23)
  4.9%
• Research & Development (R&D) or Product Development (n=66)
  14.1%
• Sales (n=7)
  1.5%
• Other (Not listed) (n=1)
  0.2%

Base: n=468

How much is your organization concerned about the following aspects of AI and LLMs?

User Data Privacy

• Significantly Concerned (n=244)
  45.4%
• Somewhat Concerned (n=212)
  39.4%
• Not Concerned At All (n=82)
  15.2%

Corporate Data Privacy

• Significantly Concerned (n=270)
  50.2%
• Somewhat Concerned (n=188)
  34.9%
• Not Concerned At All (n=80)
  14.9%

Security, Hacks

• Significantly Concerned (n=291)
  54.1%
• Somewhat Concerned (n=164)
  30.5%
• Not Concerned At All (n=83)
  15.4%

Hallucinations / Inaccurate Outputs

• Significantly Concerned (n=195)
  36.2%
• Somewhat Concerned (n=241)
  44.8%
• Not Concerned At All (n=102)
  19%

Bias in LLM Data

• Significantly Concerned (n=197)
  36.6%
• Somewhat Concerned (n=241)
  44.8%
• Not Concerned At All (n=100)
  18.6%

Reliability / Accuracy of Corporate Data

• Significantly Concerned (n=257)
  47.8%
• Somewhat Concerned (n=196)
  36.4%
• Not Concerned At All (n=85)
  15.8%

Cost to Deploy or Maintain

• Significantly Concerned (n=207)
  38.5%
• Somewhat Concerned (n=243)
  45.2%
• Not Concerned At All (n=88)
  16.4%

Misuse by Employees

• Significantly Concerned (n=227)
  42.2%
• Somewhat Concerned (n=202)
  37.5%
• Not Concerned At All (n=109)
  20.3%

Lack of Employee Training / Engagement

• Significantly Concerned (n=198)
  36.8%
• Somewhat Concerned (n=257)
  47.8%
• Not Concerned At All (n=83)
  15.4%

Other Concerns (please specify)

• Significantly Concerned (n=1)
  0.2%
• Somewhat Concerned (n=1)
  0.2%
• Not Concerned At All (n=0)
  0%

Which of the concerns you’ve selected is the most important or primary concern that needs to be addressed?

• User Data Privacy (n=135)
  25.1%
• Corporate Data Privacy (n=87)
  16.2%
• Security, Hacks (n=133)
  24.7%
• Hallucinations / Inaccurate Outputs (n=33)
  6.1%
• Bias in LLM Data (n=29)
  5.4%
• Reliability / Accuracy of Corporate Data (n=58)
  10.8%
• Cost to Deploy or Maintain (n=32)
  5.9%
• Misuse by Employees (n=24)
  4.5%
• Lack of Employee Training / Engagement (n=7)
  1.3%
• Other (n=0)
  0%

Which of the following best describes your organization’s current position or status on Shadow AI?

• Strictly prohibited and enforced (only sanctioned AI tools are allowed or in use) (n=107)
  19.9%
• Strictly prohibited with enforcement issues (we know there is shadow AI we haven’t been able to identity) (n=195)
  36.2%
• Guided allowance / support (permitted under internal guidelines and oversight) (n=128)
  23.8%
• Flexible / Experimental (tolerated to encourage innovation; minimal restrictions within limits) (n=99)
  18.4%
• Unsure (n=9)
  1.7%

How important are the following approaches in helping manage risks associated with Shadow AI?

Implementing strict monitoring and access controls

• Very Important (n=352)
  65.4%
• Somewhat Important (n=160)
  29.7%
• Not Important (n=26)
  4.8%

Providing training and guidelines for safe AI use

• Very Important (n=367)
  68.2%
• Somewhat Important (n=132)
  24.5%
• Not Important (n=39)
  7.2%

Conducting audits or reviews of AI tool usage

• Very Important (n=284)
  52.8%
• Somewhat Important (n=206)
  38.3%
• Not Important (n=48)
  8.9%

Accepting some risk as part of an innovation strategy

• Very Important (n=218)
  40.5%
• Somewhat Important (n=237)
  44.1%
• Not Important (n=83)
  15.4%

What is the primary or most important approach you’ve selected in managing risks of shadow AI?

• Implementing strict monitoring and access controls (n=130)
  24.2%
• Providing training and guidelines for safe AI use (n=258)
  48%
• Conducting audits or reviews of AI tool usage (n=103)
  19.1%
• Accepting some risk as part of an innovation strategy (n=47)
  8.7%

What conditions and controls must be in place before your organization allows agentic AI systems to take autonomous actions?

• Human approval before high-impact actions (n=346)
  64.3%
• Identity, access control, and permissions management (n=306)
  56.9%
• Audit trails and explainability of agent actions (n=208)
  38.7%
• Tool/API access restrictions (n=124)
  23%
• Data loss prevention and sensitive data controls (n=239)
  44.4%
• Runtime monitoring and rollback (n=145)
  27%
• Policy-based governance (n=63)
  11.7%
• Testing/simulation before deployment (n=183)
  34%
• Unsure (n=0)
  0%

How much does your organization trust AI-generated answers for business-critical decisions today?

• High trust (AI answers can be used with minimal human review) (n=62)
  11.5%
• Moderate trust (AI answers are useful but require human validation) (n=287)
  53.3%
• Low trust (AI answers are used only for low-risk tasks) (n=133)
  24.7%
• Very low trust (AI answers are not trusted for business decisions) (n=42)
  7.8%
• Unsure (n=14)
  2.6%

Which capabilities would most improve your organization’s trust in AI-generated answers?

• Source citations and traceability (n=213)
  39.6%
• Confidence scoring (n=131)
  24.3%
• Human review workflows (n=194)
  36.1%
• Grounding in approved enterprise data (n=236)
  43.9%
• Explainability of reasoning or steps taken (n=201)
  37.4%
• Automated hallucination detection (n=182)
  33.8%
• Evaluation/testing against known answers (n=150)
  27.9%
• Model/output monitoring over time (n=145)
  27%
• Clear ownership and accountability for AI outputs (n=99)
  18.4%
• Certification/approval by 3rd party or government agency (n=63)
  11.7%

How mature is your organization’s AI governance model today?

• Ad-hoc (No formal AI governance) (n=88)
  16.4%
• Emerging (Policies exist, but are applied inconsistently across teams or projects) (n=186)
  34.6%
• Defined (Governance processes exist and are applied to AI initiatives, but are not yet standardized across the organization) (n=153)
  28.4%
• Managed (Governance is consistently integrated into development, deployment, and monitoring workflows) (n=73)
  13.6%
• Optimized (Governance is automated through policy enforcement, access controls, monitoring, and audit systems) (n=32)
  5.9%
• Unsure (n=6)
  1.1%

Which AI security risks is your organization prioritizing most?

• Sensitive data exposure through prompts or outputs (n=283)
  52.6%
• Unauthorized access to enterprise data through AI tools (n=255)
  47.4%
• Prompt injection or malicious inputs (n=220)
  40.9%
• Model, agent, or API misuse (n=137)
  25.5%
• Data leakage through retrieval or connectors (n=189)
  35.1%
• Shadow AI usage (n=159)
  29.6%
• Supply chain risk from third-party models/tools (n=119)
  22.1%
• Compliance or regulatory exposure (n=164)
  30.5%
• Insider misuse (n=88)
  16.4%
• Unsure (n=0)
  0%

What percentage of your data is currently in the cloud (vs on-prem) and what do you expect for the future?

Today (within 6 months)

• 0% (n=21)
  3.9%
• 10% (n=27)
  5%
• 20% (n=45)
  8.4%
• 30% (n=55)
  10.2%
• 40% (n=60)
  11.2%
• 50% (n=82)
  15.2%
• 60% (n=83)
  15.4%
• 70% (n=66)
  12.3%
• 80% (n=48)
  8.9%
• 90% (n=30)
  5.6%
• 100% (n=21)
  3.9%

Within 6 - 12 months

• 0% (n=8)
  1.5%
• 10% (n=21)
  3.9%
• 20% (n=26)
  4.8%
• 30% (n=41)
  7.6%
• 40% (n=57)
  10.6%
• 50% (n=59)
  11%
• 60% (n=84)
  15.6%
• 70% (n=83)
  15.4%
• 80% (n=77)
  14.3%
• 90% (n=49)
  9.1%
• 100% (n=33)
  6.1%

Within 12 - 24 months

• 0% (n=2)
  0.4%
• 10% (n=13)
  2.4%
• 20% (n=23)
  4.3%
• 30% (n=29)
  5.4%
• 40% (n=44)
  8.2%
• 50% (n=57)
  10.6%
• 60% (n=72)
  13.4%
• 70% (n=87)
  16.2%
• 80% (n=97)
  18%
• 90% (n=63)
  11.7%
• 100% (n=51)
  9.5%

Please rank the following criteria from most (top) to least important when selecting infrastructure/model vendors:

Performance

• Rank 1 (n=161)
  29.9%
• Rank 2 (n=135)
  25.1%
• Rank 3 (n=103)
  19.1%
• Rank 4 (n=78)
  14.5%
• Rank 5 (n=36)
  6.7%
• Rank 6 (n=25)
  4.6%

Cost

• Rank 1 (n=111)
  20.6%
• Rank 2 (n=106)
  19.7%
• Rank 3 (n=91)
  16.9%
• Rank 4 (n=70)
  13%
• Rank 5 (n=76)
  14.1%
• Rank 6 (n=84)
  15.6%

Ecosystem compatibility

• Rank 1 (n=50)
  9.3%
• Rank 2 (n=64)
  11.9%
• Rank 3 (n=101)
  18.8%
• Rank 4 (n=108)
  20.1%
• Rank 5 (n=112)
  20.8%
• Rank 6 (n=103)
  19.1%

Security

• Rank 1 (n=174)
  32.3%
• Rank 2 (n=141)
  26.2%
• Rank 3 (n=86)
  16%
• Rank 4 (n=68)
  12.6%
• Rank 5 (n=49)
  9.1%
• Rank 6 (n=20)
  3.7%

Governance

• Rank 1 (n=28)
  5.2%
• Rank 2 (n=60)
  11.2%
• Rank 3 (n=93)
  17.3%
• Rank 4 (n=116)
  21.6%
• Rank 5 (n=139)
  25.8%
• Rank 6 (n=102)
  19%

Support/training

• Rank 1 (n=14)
  2.6%
• Rank 2 (n=32)
  5.9%
• Rank 3 (n=64)
  11.9%
• Rank 4 (n=98)
  18.2%
• Rank 5 (n=126)
  23.4%
• Rank 6 (n=204)
  37.9%

Does your organization have a clear process for evaluating, testing, and deploying new AI technologies, or is it case by case?

• Structured process in place (n=184)
  34.2%
• Partially structured (n=215)
  40%
• Case-by-case (n=139)
  25.8%

What steps are you prioritizing to ensure the security of sensitive data when interacting with or fine-tuning LLMs?

• Data anonymization or de-identification (sanitization) before use with LLMs (n=301)
  55.9%
• On-premises or private cloud deployment of LLMs (n=286)
  53.2%
• Strict access controls, encryption for data, pen testing, auditing (n=345)
  64.1%
• Using synthetic or generated datasets for fine-tuning (n=212)
  39.4%
• Network, API security (n=222)
  41.3%
• Data cleansing / minimization before training or data accessible in real time by LLMs (n=248)
  46.1%
• Other (please specify) (n=0)
  0%

Using a scale of 1 (lowest) to 10 (highest), please rate the maturity of your MLOps, AIOps or LLMOps practices today and estimate where you believe they will be in 12 - 24 months:

Today

• 1 (Lowest) (n=6)
  1.1%
• 2.0 (n=20)
  3.7%
• 3.0 (n=31)
  5.8%
• 4.0 (n=53)
  9.9%
• 5.0 (n=80)
  14.9%
• 6.0 (n=78)
  14.5%
• 7.0 (n=114)
  21.2%
• 8.0 (n=78)
  14.5%
• 9.0 (n=34)
  6.3%
• 10 (Highest) (n=44)
  8.2%

12 - 24 months from now

• 1 (Lowest) (n=1)
  0.2%
• 2.0 (n=3)
  0.6%
• 3.0 (n=17)
  3.2%
• 4.0 (n=42)
  7.8%
• 5.0 (n=70)
  13%
• 6.0 (n=67)
  12.5%
• 7.0 (n=78)
  14.5%
• 8.0 (n=117)
  21.7%
• 9.0 (n=81)
  15.1%
• 10 (Highest) (n=62)
  11.5%

What percent of all distinct AI/ML projects (pilots, proofs-of-concept, and deployments) launched at your organization in the past 12 months are at the following stages:

• Successful. Fully deployed to production and meeting original business objectives or ROI
  26.5%
• Partially Successful. Fully deployed to production but not yet meeting original business objectives or ROI
  21.7%
• Stalled / In Process. Reached pilot/PoC phase but has not yet been fully deployed to production
  22.2%
• Failed / Abandoned. Cancelled entirely or discontinued before completing pilot phase
  13.2%
• Pending / Planning. Still in the initial ideation, planning, or research stage
  16.4%

How significant are the following challenges in adopting/scaling your AI Stack?

Data integration & curation (data pipelines and preparation)

• Very Significant (n=312)
  58%
• Somewhat Significant (n=176)
  32.7%
• Not Significant (n=50)
  9.3%

Infrastructure (compute, storage, and platform capacity)

• Very Significant (n=302)
  56.1%
• Somewhat Significant (n=200)
  37.2%
• Not Significant (n=36)
  6.7%

Security (protecting models, data, and systems)

• Very Significant (n=341)
  63.4%
• Somewhat Significant (n=166)
  30.9%
• Not Significant (n=31)
  5.8%

Tool complexity (managing diverse AI tools and platforms)

• Very Significant (n=194)
  36.1%
• Somewhat Significant (n=283)
  52.6%
• Not Significant (n=61)
  11.3%

Executive buy-in (leadership support for AI initiatives)

• Very Significant (n=245)
  45.5%
• Somewhat Significant (n=190)
  35.3%
• Not Significant (n=103)
  19.1%

Talent or skills gaps (shortage of AI-related skills)

• Very Significant (n=253)
  47%
• Somewhat Significant (n=210)
  39%
• Not Significant (n=75)
  13.9%

Organizational resistance to change (cultural or process barriers)

• Very Significant (n=224)
  41.6%
• Somewhat Significant (n=230)
  42.8%
• Not Significant (n=84)
  15.6%

Regulatory compliance (meeting laws and industry standards)

• Very Significant (n=291)
  54.1%
• Somewhat Significant (n=186)
  34.6%
• Not Significant (n=61)
  11.3%

AI governance (enforcing internal AI policies, standards, and controls)

• Very Significant (n=235)
  43.7%
• Somewhat Significant (n=249)
  46.3%
• Not Significant (n=54)
  10%

Cost (financial constraints for AI adoption)

• Very Significant (n=290)
  53.9%
• Somewhat Significant (n=211)
  39.2%
• Not Significant (n=37)
  6.9%

Data quality (accuracy and consistency of business data)

• Very Significant (n=377)
  70.1%
• Somewhat Significant (n=130)
  24.2%
• Not Significant (n=31)
  5.8%

Trust & safety (model bias, harmful outputs, and user protection)

• Very Significant (n=260)
  48.3%
• Somewhat Significant (n=224)
  41.6%
• Not Significant (n=54)
  10%

Which of the above challenges is the most significant or difficult to overcome?

• Data integration & curation (data pipelines and preparation) (n=50)
  9.3%
• Infrastructure (compute, storage, and platform capacity) (n=52)
  9.7%
• Security (protecting models, data, and systems) (n=124)
  23%
• Tool complexity (managing diverse AI tools and platforms) (n=52)
  9.7%
• Executive buy-in (leadership support for AI initiatives) (n=30)
  5.6%
• Talent or skills gaps (shortage of AI-related skills) (n=42)
  7.8%
• Organizational resistance to change (cultural or process barriers) (n=23)
  4.3%
• Regulatory compliance (meeting laws and industry standards) (n=34)
  6.3%
• AI governance (enforcing internal AI policies, standards, and controls) (n=53)
  9.9%
• Cost (financial constraints for AI adoption) (n=37)
  6.9%
• Data quality (accuracy and consistency of business data) (n=28)
  5.2%
• Trust & safety (model bias, harmful outputs, and user protection) (n=13)
  2.4%
• Other (n=0)
  0%

On average, how long does it take to move an AI project from development to production?

• <1 month (n=46)
  8.6%
• 1-3 months (n=105)
  19.5%
• 4-6 months (n=206)
  38.3%
• 7-12 months (n=104)
  19.3%
• 12+ months (n=51)
  9.5%
• Unsure, we haven’t made it there (n=26)
  4.8%

How much AI-driven transformation do you expect in the following areas of your organization over the coming 3-5 years?

Customer experience and engagement

• Significant Transformation (n=300)
  55.8%
• Moderate Transformation (n=177)
  32.9%
• Limited or No Transformation (n=61)
  11.3%

Operations and process automation

• Significant Transformation (n=323)
  60%
• Moderate Transformation (n=158)
  29.4%
• Limited or No Transformation (n=57)
  10.6%

Product development and innovation

• Significant Transformation (n=252)
  46.8%
• Moderate Transformation (n=208)
  38.7%
• Limited or No Transformation (n=78)
  14.5%

Manufacturing and engineering

• Significant Transformation (n=296)
  55%
• Moderate Transformation (n=134)
  24.9%
• Limited or No Transformation (n=108)
  20.1%

Back office (HR, Finance & Legal)

• Significant Transformation (n=274)
  50.9%
• Moderate Transformation (n=183)
  34%
• Limited or No Transformation (n=81)
  15.1%

Coding and development

• Significant Transformation (n=311)
  57.8%
• Moderate Transformation (n=165)
  30.7%
• Limited or No Transformation (n=62)
  11.5%

Sales and marketing

• Significant Transformation (n=274)
  50.9%
• Moderate Transformation (n=201)
  37.4%
• Limited or No Transformation (n=63)
  11.7%

How important are vendor partnerships and ecosystem strength (e.g., cloud providers, specialized AI vendors, integrators) to the success of your AI strategy?

• Very important - Critical to AI strategy and adoption (n=297)
  55.2%
• Somewhat important - Helpful but not essential (n=163)
  30.3%
• Neutral / Moderate - Plays a minor role in decision-making (n=61)
  11.3%
• Not important - Vendor ecosystem does not influence our AI strategy (n=17)
  3.2%