Home / Research / Agentic AI Adoption Rate by Industry

Agentic AI Adoption Rate, by Industry: 2026 Report

By Darin ten Bruggencate, Founder, AI Industry Reviews · October 9, 2026

AI agents are software that can plan a task and carry out several steps with limited human direction. Large companies have started to put them to work, and how far each industry has gone shows where the technology is taking hold.

Our analysts studied five surveys from McKinsey, KPMG and Deloitte, fielded between June 2025 and June 2026. Each survey sets its own level of use, so we ranked industries only within a survey. We also tracked McKinsey's industry results from 2025 to 2026. In KPMG's Global AI Pulse sector ranking, technology, media and telco has the highest share at the deploying stage or beyond, 66%.

In the following sections, we explain how each survey defines adoption, compare industries now and two years out, and look at how scaling has changed since 2025. We close with Census Bureau figures on general AI use.

Agentic AI Adoption Rate by Sector in 2026

AI agent engagement by sector, KPMG Global AI Pulse, February to March 2026 (% of respondents)

• Technology, media and telco has the highest share at the deploying stage or beyond, 66%, followed by industrial manufacturing and automotive at 65%.
• Financial services has the lowest share at the deploying stage or beyond, 50%, and the lowest share in the deploying stage alone, 10%. KPMG describes the sector as operating "within tightly regulated conditions, with orchestration more constrained by compliance complexity."
• Technology, media and telco also has the largest share developing multi-agent systems, 21%. Energy and natural resources has the smallest, 11%, and the largest share scaling agents across functions, 23%.
• Life sciences has the largest share in the deploying stage alone, 20%, and the smallest share scaling agents across functions, 8%.
• For business leaders, agent deployment is now common across sectors, so the gap is in how far companies take it. Half or more of each sector is at the deploying stage or beyond, from 50% in financial services to 66% in technology. Regulated sectors are likely to keep moving more slowly toward multi-agent systems.

Agentic AI Use by Industry, Today and in Two Years

Extent of agentic AI use by industry, Deloitte, August to September 2025 (% of respondents)

• Technology, media and telecom has the highest share using agentic AI at least moderately, 28%, against 23% across all respondents. Energy, resources and industrials has the lowest, 20%.
• Minimal use is the most common answer in every industry, from 46% in life sciences and health care to 57% in energy, resources and industrials.
• Life sciences and health care has the largest share not using agentic AI at all, 30%, while its share using it at least moderately, 25%, ranks second.
• Within two years, 82% of technology, media and telecom respondents expect at least moderate use, and 11% expect agentic AI to be fully integrated into core operations. The other four industries expect 71% to 73% at least moderate use.
• Most companies still use agentic AI lightly, but they expect that to change quickly. Only 23% use it at least moderately today, while 71% to 82% expect to within two years. If those plans hold, vendors and service providers face a sharp rise in demand by 2028.

AI Agent Scaling by Industry and Business Function

AI agent scaling by industry, McKinsey, May to June 2026 (% of all respondents in the industry)

• Technology has the highest share in any single function: 31% of technology respondents say their organizations are scaling AI agents in software engineering.
• Media and telecom has the most functions where at least 10% of respondents are scaling agents, 7 of 9. Technology, financial institutions and advanced manufacturing have 6 each.
• Energy and materials and engineering and construction have no function at 10% or more. Their highest shares are 7% and 8%.
• In healthcare payers and providers, IT and software engineering tie for the highest share at 11%, from a sample of 52.
• Companies are scaling agents first in technical work. Across all respondents, IT leads at 12% and software engineering reaches 31% in technology firms. Workers in IT and engineering are likely to see agents change their jobs before other functions do.

Change in AI Agent Scaling From 2025 to 2026

AI agent scaling by industry, McKinsey, 2025 and 2026 (% of all respondents in the industry)

AI agent scaling by business function, all respondents, McKinsey, 2025 and 2026

• Financial institutions and advanced manufacturing each went from no function at 10% or more in 2025 to 6 in 2026. McKinsey's 2025 group was named banking and other financial institutions.
• Technology's share scaling agents in software engineering rose from 24% to 31%, the largest share in either year.
• Insurance's highest share fell from 20% in marketing and sales to 14% in service operations, and its sample fell from 80 to 44.
• Across all respondents, every function rose by 1 to 3 points except risk, legal, and compliance, which stayed at 5%.
• At organizations with revenue of $1 billion or more, 40% were scaling agents in at least one function in 2026, up from 27% in 2025. At smaller organizations the share was 22% in both years.
• Large companies are pulling ahead in scaling agents. At firms with $1 billion or more in revenue, 40% scaled agents in at least one function in 2026, up from 27%, while smaller firms held at 22%. If that gap persists, smaller firms risk falling behind on productivity gains.

AI Agent Deployment in US Banking, Technology and Asset Management

AI agent adoption in three US industries, KPMG, November to December 2025

• Banking's 40% is 14 points above the 26% KPMG reported across industries in the same quarter.
• KPMG's cross-industry deployment share was 11% in Q1 2025 and 42% in Q3 2025 before falling to 26% in Q4 2025. KPMG says leaders have adopted "more sophisticated definitions of what constitutes a true agent."
• Deployment claims are sensitive to how surveys define an agent. KPMG's cross-industry share swung from 11% to 42% and back to 26% in 2025 as leaders tightened their definitions. Buyers and investors should treat single survey readings with care.

General AI Use by Industry, Census Bureau Data

The Census Bureau's Business Trends and Outlook Survey asks US businesses whether they used AI in any business function in the past two weeks. Virtual agents appear only as one example in its definition of AI, so these figures measure general AI use, and the survey covers businesses of every size.

General AI use by sector, US businesses, two weeks of September 7 to 20, 2026 (selected sectors)

• Information, the sector that includes software publishing, telecommunications and data processing, has the highest share at 50.2%. It is the closest Census match to the technology, media and telecom groups that rank first in KPMG's global sector survey and in Deloitte's survey.
• Wider AI use follows the same industry pattern as agents, which suggests the leaders will stay ahead. The information sector reports 50.2% AI use and finance 40.1%, against 23.8% for all US businesses. Industries already ahead on basic AI use are likely to lead agent adoption too.

What Agentic AI Adoption Trends Mean for 2027 and Beyond

Most companies have tried AI agents, but few lean on them for real work yet. The ones moving fastest are large companies and technology firms, and they are putting agents first into IT and software jobs. That means the people who build and run software are the first to see agents take over parts of their day, while workers in banks, energy and drug companies will feel it later.

The next two years will show whether the plans companies describe turn into daily use. Large companies have the money and staff to set agents up across many teams, so the gap between them and smaller firms could widen. Surveys will also keep moving as researchers agree on what counts as an agent, so any single reading can overstate or understate how far things have gone.

So if you are weighing agents for your own company, start with one well-defined IT or software task and measure the result before spreading it across the business.