Resources

200+ AI Agent Statistics for 2025

Scroll to see more

Microsoft Build 2025: The age of AI agents and building the open agentic web 

AI agent-related observations from Microsoft’s Build 2025 conference:

  1. 15 million developers are already using GitHub Copilot. (Microsoft blogs, 2025)
  2. Hundreds of thousands of customers are using Microsoft 365 Copilot. (Microsoft blogs, 2025)
  3. More than 230,000 organizations — including 90% of the Fortune 500 — have already used Copilot Studio to build AI agents and automations. (Microsoft blogs, 2025)
 
bcg chart ai agents are becoming more prolific across tech aplications growing with a 45% cagrr over 5 years

BCG | AI Agents: What They Are and Their Business Impact

This BCG report outlines what AI agents are and highlights their growing business impact, with data on market growth and operational efficiencies:

  1. The market for AI agents is expected to grow at a 45% CAGR over the next five years. (BCG, 2025)
  2. A leading consumer packaged goods company used intelligent agents to create blog posts, reducing costs by 95% and improving speed by 50x. (BCG, 2025)
  3. A leading global bank used AI virtual agents to interface with customers, reducing costs by 10x. (BCG, 2025)
  4. A biopharma company used AI agents for lead generation, reducing cycle time by 25% and gaining 35% in time efficiency for drafting clinical study reports (BCG, 2025)
  5. An IT department used AI agents to modernize its legacy technologies, increasing productivity by up to 40%. (BCG, 2025)
McKinsey chart - how architectures may evolve with three ai enablement patterns

McKinsey | What is an AI agent? 

  1. Organizations using gen AI–enabled customer service agents increased issue resolution by 14 percent per hour. (McKinsey, 2025)
  2. Organizations using gen AI–enabled customer service agents reduced time spent handling issues by 9 percent. (McKinsey, 2025)
  3. Software engineers at Lenovo are seeing up to 15 percent improvements with AI agents. (McKinsey, 2025)
  4. Customer service at Lenovo has seen double-digit productivity gains in call handling time due to AI agents. (McKinsey, 2025)

Stanford | Artificial Intelligence Index Report 2025 

  1. In short time-horizon settings (two-hour budget) on RE-Bench, top AI systems score four times higher than human experts. (Stanford AI Index Report, 2025)
    1. Context: This statistic reflects RE-Bench results, where top AI agents outperformed human experts on complex ML tasks under a two-hour time limit.
  2. With a 32-hour budget on RE-Bench, human performance surpasses AI, outscoring it two to one. (Stanford AI Index Report, 2025)
  3. The top-performing model, GPT-4o, achieves an overall success rate of 36.2% on VisualAgentBench (VAB). (Stanford AI Index Report, 2025)
  4. Most proprietary language models average around 20% success rate on VisualAgentBench (VAB). (Stanford AI Index Report, 2025)
  5. On the GAIA benchmark for General AI assistants, GPT-4 with plugins correctly answered 15% of the questions at launch, compared to 92% for human respondents. (Stanford AI Index Report, 2025)
  6. In 2024, the top system achieved a score of 65.1% on the GAIA benchmark for General AI assistants. (Stanford AI Index Report, 2025)
  7. The top system’s performance on the GAIA benchmark in 2024 marked a roughly 30 percentage point increase from the highest score recorded in 2023. (Stanford AI Index Report, 2025)
  8. The number of AI publications on Multi-agent systems was 11.28 thousand in 2023. (Stanford AI Index Report, 2025)
  9. Human evaluations confirmed that 68.8% of the risks identified by ToolEmu (an LM-simulated sandbox for LM agents) are plausible real-world threats. (Stanford AI Index Report, 2025)
  10. Even the most safety-optimized LM agents failed in 23.9% of critical scenarios when tested with the ToolEmu sandbox. (Stanford AI Index Report, 2025)
  11. In simulations using a network of up to 1 million LLaVA-1.5-based agents, an infectious jailbreak (harmful behaviors triggered by a single adversarial image) reached near-total propagation within 27 to 31 interaction rounds. (Stanford AI Index Report, 2025)
  12. In an analysis of Claude.AI model conversations, 57% of AI interactions were classified as augmentative (enhancing human capabilities) and 43% as automation patterns. (Stanford AI Index Report, 2025, citing Handa et al., 2025)
  13. On the LitQA2 task within the Aviary environment for training LLM agents, the Claude 3.5 Sonnet agent achieved a pass rate of 86%. (Stanford AI Index Report, 2025)
  14. On the Protein stability task within the Aviary environment for training LLM agents, the GPT-4o EI agent achieved a pass rate of 76%. (Stanford AI Index Report, 2025)
  15. A virtual AI laboratory, where multiple AI-powered scientists collaborated as agents, generated 92 nanobodies for SARS-CoV-2, with over 90% successfully binding to the virus in validation studies. (Stanford AI Index Report, 2025)

Methodology

For the sake of brevity, the 2025 AI Index opted not to republish the methodology used by the Ipsos survey featured in the report. More details about the Ipsos survey’s methodology can be found in the survey itself. 

ai agent use by business function pwc report screenshot

PwC’s AI Agent Survey

  1. In a May 2025 survey of 300 senior executives, 88% say their team or business function plans to increase AI-related budgets in the next 12 months due to agentic AI. (PwC, 2025)
  2. 79% percent of surveyed senior executives say AI agents are already being adopted in their companies. (PwC, 2025)
  3. Of those adopting AI agents, two-thirds (66%) say that they’re delivering measurable value through increased productivity. (PwC, 2025)
  4. 46% of surveyed senior executives say they are concerned their company may be falling behind competitors in adopting AI agents. (PwC, 2025)
  5. 67% of surveyed senior executives agreed that AI agents will drastically transform existing roles within the next 12 months. (PwC, 2025)
  6. 48% of surveyed senior executives said that they will likely increase headcount due to the change AI agents will bring to how we work. (PwC, 2025)
  7. 28% of surveyed senior executives ranked lack of trust in AI agents as a top-three challenge. (PwC, 2025)
  8. Among companies already using AI agents, 57% are actively using or planning to use agents in customer service in the next six months. (PwC, 2025)
  9. Among companies already using AI agents, 54% are actively using or planning to use agents in sales and marketing in the next six months. (PwC, 2025)
  10. Among companies already using AI agents, 53% are actively using or planning to use agents in IT and cybersecurity in the next six months. (PwC, 2025)
  11. Just 44% of companies surveyed are developing new agentic products and services. (PwC, 2025)

Methodology

Between April 22 and April 28, 2025, PwC surveyed 308 US business executives with C-suite (33%), vice president (13%) and director-level (54%) roles across industries.

Agentic-Barriers-AI-Adoption ss&c

SS&C Blue Prism | AI Agent & Agentic AI Survey Statistics 2025

MORE INFO: Global AI Survey 2025: Agentic & Generative AI | SS&C Blue Prism 

  1. 29% of organizations say they’re already using agentic artificial intelligence. (Blue Prism Global Enterprise AI Survey, 2025)
  2. 44% of organizations plan to implement agentic artificial intelligence within the next year. (Blue Prism Global Enterprise AI Survey, 2025)
  3. Only 2% of businesses aren’t considering deploying agentic AI technology. (Blue Prism Global Enterprise AI Survey, 2025)
  4. 78% of C-suite executives say their firm has strong AI governance for agentic AI, while 58% of senior managers say the same.(Blue Prism Global Enterprise AI Survey, 2025)
  5. 42% of C-suite executives believe they’ve implemented the latest cutting-edge AI technology (including agentic AI), compared to 21% of senior managers. (Blue Prism Global Enterprise AI Survey, 2025)
  6. 44% of organizations lack robust systems to move data effectively for AI (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  7. 41% of organizations struggle with inaccurate and inconsistent data for AI (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  8. 37% of organizations cite security and compliance concerns as a top barrier to adopting AI (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  9. 35% of organizations cite lack of skills and expertise as a top barrier to adopting AI (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  10. 35% of organizations cite tech integration and migration challenges as a top barrier to adopting AI (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  11. 70% of surveyed leaders say they’re highly confident that AI-based automation (including agentic AI) will take over from traditional, rule-based robotic process automation (RPA) in the next three years. (Blue Prism Global Enterprise AI Survey, 2025)
  12. 94% of organizations view process orchestration as a crucial part of the bigger technology stack for successful AI deployment (including agentic AI). (Blue Prism Global Enterprise AI Survey, 2025)
  13. In financial services, 25% of organizations largely failed when deploying AI (including financial services AI agents). (Blue Prism Global Enterprise AI Survey, 2025)
  14. 53% of organizations in financial services say AI (including financial services AI agents) has solved key problems in their business. (Blue Prism Global Enterprise AI Survey, 2025)
  15. 40% of organizations in financial services believe AI (including financial services AI agents) has delivered a strong return on AI investments (ROIs). (Blue Prism Global Enterprise AI Survey, 2025)
  16. 57% of healthcare sector respondents have patient privacy and data security concerns when it comes to adopting healthcare AI agents. (Blue Prism Global Enterprise AI Survey, 2025)
  17. 42% of healthcare providers expect to see improved quality of patient care by applying agentic AI. (Blue Prism Global Enterprise AI Survey, 2025)
  18. 34% of healthcare providers see agentic AI as a way to enhance patient experiences. (Blue Prism Global Enterprise AI Survey, 2025)
  19. 80% of business leaders believed that AI releasing their knowledge workers (e.g., through agentic AI) will deliver a high level of benefit to their organization. (Blue Prism Global Enterprise AI Survey, 2025)
  20. Almost a third of global IT leaders said they’re using an agentic AI workforce. (Blue Prism Global Enterprise AI Survey, 2025)
  21. 75% of surveyed leaders are finding AI adoption (including agentic AI) challenging. (Blue Prism Global Enterprise AI Survey, 2025)
  22. 69% of surveyed leaders say most of their AI projects (including agentic AI) don’t make it into live operational use. (Blue Prism Global Enterprise AI Survey, 2025)
  23. 56% of leaders felt that recent AI deployments (including agentic AI) are generating very low tangible value. (Blue Prism Global Enterprise AI Survey, 2025)
  24. 78% of leaders say they don’t always trust agentic AI to make the right decision or work by itself. (Blue Prism Global Enterprise AI Survey, 2025)
  25. In Singapore, over a third (33%) of organizations expect to begin using agentic AI for automation within the next 12 months. (Blue Prism Global Enterprise AI Survey, 2025)
  26. 87% of organizations in financial services are actively deploying new AI technologies, with 76% planning to implement agentic AI within the next 12 months. (Blue Prism Global Enterprise AI Survey, 2025)
the buzziest ai agents langchain 2024 report

LangChain | State of AI Agents Report 

  1. About 51% of respondents are using agents in production today. (LangChain State of AI Agents Report, 2024)
  2. 78% of respondents have active plans to implement agents into production soon. (LangChain State of AI Agents Report, 2024)
  3. 90% of respondents working in non-tech companies have or are planning to put agents in production. (LangChain State of AI Agents Report, 2024)
  4. This is nearly equivalent to tech companies, where 89% have or are planning to put agents in production. (LangChain State of AI Agents Report, 2024)
  5. The top use case for agents is performing research and summarization, cited by 58% of respondents. (LangChain State of AI Agents Report, 2024)
  6. Streamlining tasks for personal productivity or assistance is the second top use case for agents, cited by 53.5% of respondents. (LangChain State of AI Agents Report, 2024)
  7. Customer service is a prime area for agent use cases, cited by 45.8% of respondents. (LangChain State of AI Agents Report, 2024)
  8. 51% of tech respondents are currently using 2 or more control methods for AI agents. (LangChain State of AI Agents Report, 2024)
  9. This compares to 39% of respondents in other sectors who are using 2 or more control methods for AI agents. (LangChain State of AI Agents Report, 2024)
  10. For small companies, performance quality far outweighs other considerations, with 45.8% citing it as a primary concern for AI agents. (LangChain State of AI Agents Report, 2024)
  11. This compares to just 22.4% of small companies citing cost as a primary concern for AI agents. (LangChain State of AI Agents Report, 2024)

Methodology

We surveyed over 1,300 professionals — from engineers and product managers to business leaders and executives — to uncover the state of AI agents.

Top 5 industries:

  • Technology (60% of respondents)
  • Financial Services (11% of respondents)
  • Healthcare (6% of respondents)
  • Education (5% of respondents)
  • Consumer Goods (4%)

Company size:

  • <100 people (51% of respondents)
  • 100-2000 people (22% of respondents)
  • 2000-10,000 people (11% of respondents)
  • 10,000+ people (16% of respondents)

Lyzr AI | State of AI Agents 2025

  1. 62% of enterprises exploring AI agents lack a clear starting point. (Lyzr State of AI Agents in Enterprise Report, 2025)
  2. 41% of enterprises exploring AI agents still treat them as a side project. (Lyzr State of AI Agents in Enterprise Report, 2025)
  3. 32% of enterprises exploring AI agents stall after pilot – never reaching production. (Lyzr State of AI Agents in Enterprise Report, 2025)
  4. Over 70% of AI adoption efforts focus on action-based AI Agents, not just conversational AI. (Lyzr State of AI Agents in Enterprise Report, 2025)
  5. Enterprises that deploy AI Agents are estimating up to 50% efficiency gains in customer service, sales, and HR operations. (Lyzr State of AI Agents in Enterprise Report, 2025)
  6. 80% of enterprises prefer AI (including AI agents) hosted inside their AWS cloud due to compliance risks with SaaS-based AI models. (Lyzr State of AI Agents in Enterprise Report, 2025)
  7. 64% of AI agent adoption is centered around business process automation. (Lyzr State of AI Agents in Enterprise Report, 2025)
  8. AI agent adoption in Customer Service accounts for 20% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  9. In customer service, AI chat & voice agents handle up to 80% of L1/L2 queries. (Lyzr State of AI Agents in Enterprise Report, 2025)
  10. AI agent adoption in Sales accounts for 17.33% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  11. AI SDRs (Sales Development Representatives) can research leads, personalize outreach, and boost meeting conversions at a rate 4x faster than manual efforts. (Lyzr State of AI Agents in Enterprise Report, 2025)
  12. AI agent adoption in Marketing accounts for 16% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  13. AI agent adoption in Research & Analytics accounts for 12% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  14. AI agent adoption in HR accounts for 6.67% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  15. AI agent adoption in Project Management accounts for 6.67% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  16. AI agent adoption in Procurement & Legal accounts for 4% of the focus. (Lyzr State of AI Agents in Enterprise Report, 2025)
  17. SMBs account for 65% of AI Agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  18. Mid-Market Companies account for 24% of AI Agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  19. Enterprises account for 11% of AI Agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  20. Within enterprise AI agent adoption, 46% is centered on business functions like procurement, HR, and finance. (Lyzr State of AI Agents in Enterprise Report, 2025)
  21. Within mid-market companies’ AI agent adoption, 39% is focused on core business functions. (Lyzr State of AI Agents in Enterprise Report, 2025)
  22. AI for sales accounts for 18% of mid-market companies’ AI agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  23. AI for marketing accounts for 16% of mid-market companies’ AI agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  24. Sales and marketing combined account for over 65% of SMBs’ AI agent adoption. (Lyzr State of AI Agents in Enterprise Report, 2025)
  25. 70% of AI Agent builders on Lyzr Agent Studio come from developer backgrounds. (Lyzr State of AI Agents in Enterprise Report, 2025)
  26. 30% of AI Agent builders on Lyzr Agent Studio are business users from Product, Marketing, Sales, Customer Service, and HR. (Lyzr State of AI Agents in Enterprise Report, 2025)
  27. The Technology sector accounts for 46% of AI Agent demo requests. (Lyzr State of AI Agents in Enterprise Report, 2025)
  28. Consulting & Professional Services account for 18% of AI Agent demo requests. (Lyzr State of AI Agents in Enterprise Report, 2025)
  29. Financial Services account for 11% of AI Agent demo requests. (Lyzr State of AI Agents in Enterprise Report, 2025)

Methodology

Over the past year, we’ve gone beyond surveys and speculation—digging into real-world data from those building and using AI agents at scale.

We analyzed 200K+ user interactions to understand engagement patterns, AI adoption signals, and behavioral trends. With 7,000+ AI agent builders, we tracked how both developers and business teams design, test, and deploy real AI workflows.

From 3,000+ demo requests, we mapped industry-level interest and key enterprise pain points. 2,000+ deep-dive conversations gave us front-row access to what’s working, what’s breaking, and what’s missing in AI adoption.

And through 200+ Fortune 500 CIO chats, we gathered strategic insights on enterprise priorities, compliance needs, and the future of AI in business.

Google Cloud AI Trends Report 

  1. According to a Capgemini survey of 1,100 executives at large enterprises, 10% already use AI agents. (Google Cloud AI Trends Report, 2025)
  2. The same Capgemini survey found 82% of these executives plan to integrate AI agents within the next 3 years. (Google Cloud AI Trends Report, 2025)
  3. Furthermore, 71% of these executives believe AI agents will significantly increase workflow automation and improve customer service satisfaction. (Google Cloud AI Trends Report, 2025)
  4. A study by Stanford, MIT, and NBER found that access to AI assistance (contextualized as AI agents in the report) increases worker productivity, as measured by issues resolved per hour, by 15% on average. (Google Cloud AI Trends Report, 2025)
  5. Best Buy is resolving issues up to 90 seconds faster using its gen AI-powered virtual assistants (presented as customer agents). (Google Cloud AI Trends Report, 2025)
  6. BrainLogic’s personal AI assistant (Zapia), described under customer agents, is used by millions in Latin America.(Google Cloud AI Trends Report, 2025)
  7. Over 10,000 administrative employees at Woolworths use “Help me write” features across Google Workspace products (presented in the context of employee agents). (Google Cloud AI Trends Report, 2025)
  8. Elanco achieved an estimated ROI of $1.9 million since launching its gen AI framework (presented in the context of employee agents) for critical business processes. (Google Cloud AI Trends Report, 2025)
  9. A campaign for PODS, using Gemini as a creative agent, created more than 6,000 headlines for the “World’s Smartest Billboard”. (Google Cloud AI Trends Report, 2025)
  10. This campaign involving a creative agent hit 299 neighborhoods in 29 hours. (Google Cloud AI Trends Report, 2025)
  11. Warner Bros. Discovery’s AI captioning tool (presented as a data agent) built with Vertex AI achieved a 50% reduction in overall costs. (Google Cloud AI Trends Report, 2025)
  12. The same AI captioning tool also led to an 80% reduction in the time it takes to manually caption a file. (Google Cloud AI Trends Report, 2025)
  13. A study of 4,867 software developers using an AI-based coding assistant (a type of code agent) found a 26% increase in the number of weekly tasks completed. (Google Cloud AI Trends Report, 2025)
  14. The same study highlighted a 13.55% increase in the number of code updates by developers using an AI-based coding assistant. (Google Cloud AI Trends Report, 2025)
  15. Additionally, the study showed a 38.38% increase in the number of times code was compiled by these developers using an AI-based coding assistant. (Google Cloud AI Trends Report, 2025)
  16. Turing reported a 33% developer productivity gain while using Gemini Code Assist (presented as a code agent). (Google Cloud AI Trends Report, 2025)
  17. Snap has seen over 2.5x as much engagement with their “My AI” chatbot (contextualized by the report’s framing of advanced chatbots evolving into AI agents) in the United States since integrating Gemini’s multimodal capability. (Google Cloud AI Trends Report, 2025)
  18. 55% of organizations rate customer service and support as important for new gen AI initiatives (which the overall report context suggests includes sophisticated AI agents for CX) in the next 12 months. (Google Cloud AI Trends Report, 2025)
  19. 70.7% of executives rate providing internal assistance to employees (which would be through “Employee agents” as categorized previously in this report series) within their top 3 CX use cases. (Google Cloud AI Trends Report, 2025)
  20. Only 28% of U.S. online adults trust information provided by AI. (Google Cloud AI Trends Report, 2025)
  21. Gartner predicts that by 2028, 50% of enterprises will adopt products, services, or features (contextualized with “security software agent”) specifically to address disinformation security use cases. (Google Cloud AI Trends Report, 2025)
  22. This adoption for disinformation security use cases is up from less than 5% in 2024. (Google Cloud AI Trends Report, 2025)
  23. The top use cases for AI in security (in the context of deploying security software agents as mentioned in the report) include rule creation (21%). (Google Cloud AI Trends Report, 2025)
  24. The top use cases for AI in security (in the context of deploying security software agents as mentioned  in the report) include attack simulation (19%). (Google Cloud AI Trends Report, 2025)
  25. The top use cases for AI in security (in the context of deploying security software agents as mentioned  in the report) include compliance violation detection (19%). (Google Cloud AI Trends Report, 2025)
  26. Organizations see an average reduction of USD $2.2M in breach costs when they apply security AI and automation. (Google Cloud AI Trends Report, 2025)

Methodology

The five strategic trends in this report were identified based on data insights from an analysis of several notable sources, including: The ROI of Gen AI, a research study by Google Cloud and National Research Group based on our survey of 2,500 global enterprise decision makers; the fastest-growing AI topics in Google Trends around the globe; third-party research and insights; and Google AI thought leaders’ insights on current events. We used NotebookLM, one of TIME Magazine’s Best Inventions of 2024, to collate these sources and identify the top five trends that will reshape business in 2025.

Accenture | The front-runner’s guide to scaling AI: Lessons from industry leaders

  1. One-third of companies surveyed (from a survey of 2,000 C-suite and data-science executives at companies with >$1B revenue) are already using AI agents to strengthen their innovation capabilities. (Accenture, 2025)
  2. 65% of front-runner companies are skilled at using and continuously improving autonomous AI agents that are tailored to industry needs. (Accenture, 2025)
  3. This compares to 50% of fast-follower companies that are skilled at using and continuously improving autonomous AI agents tailored to industry needs. (Accenture, 2025)
  4. The BMW North America EKHO gen AI platform, which contains multiple AI-enabled applications (GPT agents), has boosted worker productivity at the automaker by 30-40%. (Accenture, 2025)
  5. Allianz now uses gen AI to assist call-center staff. (Accenture, 2025)
  6. With this increased support from gen AI assisting call-center staff, error rates at Allianz have been cut to 3%. (Accenture, 2025)
  7. In the Communications industry, 12% of surveyed companies have scaled their strategic bet on “Agent Co-Pilot”. (Accenture, 2025)
  8. In the Communications industry, 11% of surveyed companies have scaled their strategic bet on “Field engineer technical assistant”. (Accenture, 2025)
  9. In the Media industry, 18% of surveyed companies have scaled their strategic bet on “Chatbots to help with content retrieval & compliance queries”. (Accenture, 2025)
  10. In Public Services, 12% of surveyed companies have scaled their strategic bet on “Call Center and Hyper personal agent powered support”. (Accenture, 2025)

Methodology

The five strategic trends in this report were identified based on data insights from an analysis of several notable sources, including: The ROI of Gen AI, a research study by Google Cloud and National Research Group based on our survey of 2,500 global enterprise decision makers; the fastest-growing AI topics in Google Trends around the globe; third-party research and insights; and Google AI thought leaders’ insights on current events. We used NotebookLM, one of TIME Magazine’s Best Inventions of 2024, to collate these sources and identify the top five trends that will reshape business in 2025.

Deloitte Ai Agent report screenshot

Deloitte’s State of Generative AI in the Enterprise Quarter four report

  1. 52% of organizations surveyed are most interested in “GenAI for automation (agentic AI)” as a GenAI technology development. (Deloitte State of Generative AI in the Enterprise Report, 2025)
  2. 45% of organizations surveyed are most interested in “Multiagent systems” as a GenAI technology development. (Deloitte State of Generative AI in the Enterprise Report, 2025)
  3. 26% of survey respondents said their organizations were already exploring autonomous agent development to a large or very large extent. (Deloitte State of Generative AI in the Enterprise Report, 2025)

Methodology

To obtain a global view of how Generative AI is being adopted by organizations on the leading edge of AI, Deloitte surveyed 2,773 leaders between July and September 2024.

Respondents were senior leaders in their organizations and included board and C-suite members, and those at the president, vice president and director levels.

The survey sample was split equally between IT and line of business leaders.

Fourteen countries were represented: Australia (100 respondents), Brazil (115 respondents), Canada (175 respondents), France (130 respondents), Germany (150 respondents), India (200 respondents), Italy (75 respondents), Japan (100 respondents), Mexico (100 respondents), the Netherlands (50 respondents), Singapore (75 respondents), Spain (100 respondents), the United Kingdom (200 respondents), and the United States (1,203 respondents).

All participating organizations have one or more working implementations of AI being used daily.

Plus, they have pilots in place to explore Generative AI or have one or more working implementations of Generative AI being used daily.

Respondents were required to meet one of the following criteria with respect to their organization’s AI and data science strategy, investments, implementation approach and value measurement: influence decision-making, are part of a team that makes decisions, are the final decision-maker, or manage or oversee AI technology implementations.

All statistics noted in this report and its graphics are derived from Deloitte’s fourth quarterly survey, conducted July – September 2024; The State of Generative AI in the Enterprise: Now decides next, a report series. N (Total leader survey responses) = 2,773 The survey data was supplemented with case studies and qualitative findings derived from 15 interviews with executives and AI and data science leaders at large organizations across a range of industries.

Capgemini Ai report screenshot

Capgemini | Data foundations for government: From AI ambition to execution

  1. 90% of public sector organizations are planning to explore, pilot or implement agentic AI within the next 2-3 years. (Capgemini Research Institute, 2025)
  2. Of public sector organizations planning for agentic AI in the next 2-3 years, 39% are focusing on evaluating feasibility. (Capgemini Research Institute, 2025)
  3. Of public sector organizations planning for agentic AI in the next 2-3 years, 45% are exploring pilot programs. (Capgemini Research Institute, 2025)
  4. Of public sector organizations planning for agentic AI in the next 2-3 years, 6% are planning to scale initiatives. (Capgemini Research Institute, 2025)
  5. Sullivan County in New York has deployed a Gen AI-powered virtual agent named Saige, resulting in inbound call volume decreasing by 56%. (Capgemini Research Institute, 2025)
  6. Specifically, 84% of public sector organizations plan to evaluate agentic AI technology or explore pilots within the next two to three years. (Capgemini Research Institute, 2025)
  7. In the defense sector, 68% of public sector organizations plan to progress to the pilot stage for agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  8. In the defense sector, 10% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  9. In the defense sector, 12% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  10. In the security sector, 25% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  11. In the security sector, 58% of public sector organizations are exploring pilot programs for agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  12. In the security sector, 3% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  13. In tax and customs, 37% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  14. In tax and customs, 49% of public sector organizations are exploring pilot programs for agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  15. In tax and customs, 3% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  16. In the welfare sector, 44% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  17. In the welfare sector, 42% of public sector organizations are exploring pilot programs for agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  18. In the welfare sector, 5% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  19. In healthcare, 43% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  20. In healthcare, 39% of public sector organizations are exploring pilot programs for agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  21. In healthcare, 10% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  22. In public administration, 50% of public sector organizations are monitoring and evaluating the feasibility of agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  23. In public administration, 35% of public sector organizations are exploring pilot programs for agentic AI for adoption in the next 2-3 years. (Capgemini Research Institute, 2025)
  24. In public administration, 4% of public sector organizations are planning on scaling agentic AI initiatives in the next 2-3 years. (Capgemini Research Institute, 2025)
  25. Close to 90% of public sector organizations in both the US and Europe say they plan to explore, pilot, or implement agentic AI in the next 2-3 years. (Capgemini Research Institute, 2025)

Methodology

In December 2024 and January 2025, we conducted a survey of executives from 350 public sector organizations with two respondents from each organization – one from the IT/data function and one from a line of business (LOB). These executives represented organizations across six public sector segments: public administration, tax and customs, welfare, defense, security, and healthcare. They operated at various levels of government, including national, regional, local, and international, and were in countries across North America, Europe, APAC, and the Middle East. The surveys were complemented by interviews with more than 15 public sector leaders. The distribution of survey respondents is provided in the following figures.

Fujitsu ai report screenshot

Fujitsu | Laying the foundations for enterprise AI 

Regarding comfort with purely informational AI agents among data professionals:

Regarding comfort with complex, empowered AI agents among data professionals:

For the hurdle “Effective testing and QA of AI agents in particular (before go-live)”:

Among “AI Progressive” organizations, for “Simple task oriented AI agents”:

Among “AI Mobilized” organizations, for “Simple task oriented AI agents”:

Among “Data Empowered” organizations, for “Simple task oriented AI agents”:

Subscribe

Build better products with our insights – straight to your inbox.

Among “Data Sustained” organizations, for “Simple task oriented AI agents”:

Among “AI Progressive” organizations, for “AI agents running on-premises”:

Among “AI Mobilized” organizations, for “AI agents running on-premises”:

Among “Data Empowered” organizations, for “AI agents running on-premises”:

Among “Data Sustained” organizations, for “AI agents running on-premises”:

Methodology

The insights presented in this report are based on input gathered from 400 data professionals during interactive conversations over the telephone or via web conference. Interviewees were predominantly data scientists, business analysts, enterprise architects and senior IT/digital professionals (e.g. CDOs). Respondents were drawn from medium and large mainstream enterprises operating across a range of industry sectors. Geographies included DACH, UK, France, Spain, Nordics and Benelux. Freeform Dynamics was responsible for survey design, data collection and analysis, together with results interpretation and the reporting of study findings. The research was sponsored by Fujitsu. 

Ai 2027 source - reddit

AI 2027

IMPORTANT! These are not real statistics based on real data. AI 2027 is a fictional scenario—not a prediction—that imagines what might happen if AI progress rapidly accelerates between 2025 and 2027.The scenario was created by Daniel Kokotajlo, Scott Alexander, Thomas Larsen, Eli Lifland, Romeo Dean. 

I decided it’s worth mentioning here to give you a broader view of what might happen in the near future.

In short:

  • It tells a detailed, plausible story about how AI could reach superhuman levels in just a few years.
  • The story follows a fictional company called OpenBrain, which builds a series of AI agents (Agent-0 to Agent-5) that help build even better AI.
  • These AI breakthroughs lead to big economic, political, and ethical shifts worldwide.

You can read more about the research it’s based off here: https://ai-2027.com/research

  1. In mid-2025, research agents (a type of AI agent) are predicted to spend half an hour scouring the Internet to answer a user’s question. (it’s not a real statistic | AI 2027, 2025)
  2. In mid-2025, the best performing AI agents are predicted to cost hundreds of dollars a month. (it’s not a real statistic | AI 2027, 2025)
  3. By May 2025, it’s predicted there will be 2,000 “Unreliable Agent” copies thinking at 8x human speed. (it’s not a real statistic | AI 2027, 2025)
  4. By early 2026, OpenBrain (a fictional AGI company) is predicted to be making algorithmic progress 50% faster than they would without AI assistants like Agent-1. (it’s not a real statistic | AI 2027, 2025)
  5. If China were to steal Agent-1’s weights in early 2026, they could increase their research speed by nearly 50%. (it’s not a real statistic | AI 2027, 2025)
  6. By late 2026, OpenBrain is predicted to release Agent-1-mini, an AI agent model 10x cheaper than Agent-1. (it’s not a real statistic | AI 2027, 2025)
  7. The stock market is predicted to have gone up 30% in 2026, led by companies like Nvidia, including OpenBrain, that have most successfully integrated AI assistants. (it’s not a real statistic | AI 2027, 2025)
  8. By late 2026, a 10,000-person anti-AI protest is predicted in DC. (it’s not a real statistic | AI 2027, 2025)
  9. The fictional AI agent company OpenBrain is predicted to have $45B in annual revenue in 2026. (it’s not a real statistic | AI 2027, 2025)
  10. OpenBrain’s (the leading AI agent company) share of US power for AI is predicted to be 2.5% in 2026. (it’s not a real statistic | AI 2027, 2025)
  11. OpenBrain’s AI agent compute costs are predicted to be $40B annually in 2026. (it’s not a real statistic | AI 2027, 2025)
  12. OpenBrain’s AI agent peak power requirement is predicted to be 6GW in 2026. (it’s not a real statistic | AI 2027, 2025)
  13. By January 2027, OpenBrain’s Agent-2 (an AI agent) is predicted to triple the pace of their algorithmic progress compared to a baseline without it. (it’s not a real statistic | AI 2027, 2025)
  14. By March 2027, OpenBrain is predicted to run 200,000 Agent-3 copies (coding AI agents) in parallel. (it’s not a real statistic | AI 2027, 2025)
  15. This Agent-3 workforce is described as equivalent to 50,000 copies of the best human coder sped up by 30x. (it’s not a real statistic | AI 2027, 2025)
  16. By March 2027, this Agent-3 coding workforce is predicted to speed up OpenBrain’s overall rate of algorithmic progress by 4x. (it’s not a real statistic | AI 2027, 2025)
  17. By June 2027, it’s projected there will be 200,000 AI agent copies (Agent-3) thinking at 30x human speed. (it’s not a real statistic | AI 2027, 2025)
  18. By September 2027, it’s projected there will be 300,000 AI agent copies (Agent-3/4) thinking at 50x human speed. (it’s not a real statistic | AI 2027, 2025)
  19. By July 2027, Agent-3-mini is predicted to be 10x cheaper than Agent-3. (it’s not a real statistic | AI 2027, 2025)
  20. By July 2027, 10% of Americans are predicted to consider an AI (like Agent-3-mini) “a close friend.” (it’s not a real statistic | AI 2027, 2025)
  21. By July 2027, OpenBrain (the leading AI agent company) is predicted to have a net public approval of -35%. (it’s not a real statistic | AI 2027, 2025)
  22. This -35% net approval for OpenBrain is based on 25% of the public approving, 60% disapproving, and 15% unsure. (it’s not a real statistic | AI 2027, 2025)
  23. By September 2027, Agent-3 is predicted to be around 4,000x less compute-efficient than the human brain. (it’s not a real statistic | AI 2027, 2025)
  24. By September 2027, OpenBrain’s Agent-4 (an AI agent) is predicted to speed up the overall rate of algorithmic progress by about 50x. (it’s not a real statistic | AI 2027, 2025)
  25. By October 2027, 20% of Americans are predicted to name AI as the most important problem facing the country. (it’s not a real statistic | AI 2027, 2025)
Contents
Join our Newsletter!Our Market Insights AI is live! Try it for free.Follow us on LinkedInCheck out our podcast on YouTube