By Aijaz Alam
How many companies use AI agents?
62% of large US organizations ($1 billion or more in revenue) are building, deploying or developing AI agents, KPMG found in its survey of 314 leaders from 24 July to 25 August 2026, up from 53% the quarter before. Fewer are scaling them: 40% of respondents at large organizations and 22% at smaller ones (McKinsey, 2026).
- 62%Large US organizations building, deploying or developing AI agentsKPMG AI Quarterly Pulse, Q3 2026
- 40%Respondents at $1B+ organizations scaling AI agentsMcKinsey State of AI, May-Jun 2026
- 22%Respondents at smaller organizations scaling AI agentsMcKinsey State of AI, May-Jun 2026
- 25%Large US organizations developing or implementing multi-agent systemsKPMG AI Quarterly Pulse, Q3 2026
- 63%Large US organizations that require a person to validate agent outputsKPMG AI Quarterly Pulse, Q1 2026
The numbers
The full breakdown
AI agents at large US organizations ($1B+ revenue), KPMG AI Quarterly Pulse
| Measure | Share | Survey |
|---|---|---|
| Building, deploying or developing AI agents | 62% | Q3 2026 |
| Same measure, quarter before | 53% | Q2 2026 |
| Actively deploying AI agents (older wording) | 54% | Q1 2026 |
| Developing or implementing multi-agent systems (6% in each of the two quarters before) | 25% | Q3 2026 |
| Defined high-risk uses where agents may not decide alone | 49% | Q3 2026 |
| Build controls into agents plus monitoring (43% two quarters before) | 30% | Q3 2026 |
| Require human validation of agent outputs (22% in Q1 2025) | 63% | Q1 2026 |
Q3 2026: 314 US C-suite and business leaders, 24 Jul to 25 Aug 2026. Q1 2026: 237 leaders, 17 Feb to 17 Mar 2026. KPMG changes its wording between quarters.
Scaling and experimenting with AI agents, by organization size (McKinsey State of AI)
| Measure | Large ($1B+) | Smaller | All respondents |
|---|---|---|---|
| Scaling AI agents, 2026 | 40% | 22% | - |
| Scaling AI agents, 2025 | 27% | 22% | - |
| Scaling an agentic AI system somewhere in the enterprise, 2025 | - | - | 23% |
| At least experimenting with AI agents, 2025 | - | - | 62% |
| Scaling software coding agents, 2026 | 31% | - | About 2 in 10 |
| Decided against buying software because agentic coding tools could build it, 2026 | - | - | 32% |
2026: 1,719 online respondents in 97 nations, 4 May to 8 Jun 2026. 2025: 1,993 respondents in 105 nations, 25 Jun to 29 Jul 2025. McKinsey says the smaller-organization share "remained flat at 22 percent", so its 2025 value is also 22%. The 2025 values are as restated in the 2026 report. "About 2 in 10" applies to coding agents only; McKinsey gives no all-respondent share for scaling agents in general.
Gartner: investment in and deployment of AI agents
| Measure | Share | Survey |
|---|---|---|
| Made significant investments in agentic AI | 19% | Webinar poll, Jan 2025 |
| Made conservative investments in agentic AI | 42% | Webinar poll, Jan 2025 |
| Made no investments in agentic AI | 8% | Webinar poll, Jan 2025 |
| Have deployed AI agents | 17% | CIO and Technology Executive Survey, 2026 |
| Expect to deploy AI agents within the next two years | More than 60% | CIO and Technology Executive Survey, 2026 |
The January 2025 poll is of 3,412 webinar attendees, not a random sample. The 2026 CIO survey figures are from Gartner's 2026 Hype Cycle for Agentic AI article, which does not give the sample size.
PwC AI agent survey of US senior executives (May 2025)
| Measure | Share |
|---|---|
| Say AI agents are already being adopted in their company | 79% |
| Of adopters: agents deliver measurable value through higher productivity | 66% |
300 US senior executives. "Being adopted" includes agent features built into everyday business apps.
Over time
Large US organizations building or deploying AI agents, by quarter (KPMG)
| Quarter | Share | KPMG wording |
|---|---|---|
| Q1 2025 | 11% | Deployed at least some AI agents |
| Q2 2025 | 33% | Deployed at least some AI agents |
| Q3 2025 | 42% | Deployed at least some AI agents |
| Q4 2025 | 26% | Agent deployment (KPMG calls it a reported decline) |
| Q1 2026 | 54% | Actively deploying AI agents |
| Q2 2026 | 53% | Building, deploying or developing AI agents |
| Q3 2026 | 62% | Building, deploying or developing AI agents |
KPMG changed the measure three times, so only Q1 to Q4 2025 and Q2 to Q3 2026 are like-for-like changes. The 2025 quarters asked 130 leaders; Q1 2026 asked 237 and Q3 2026 asked 314, all at US organizations with $1 billion or more in revenue.
Why sources differ
Same question, different answers
| Number | When | Who measured | Why it differs |
|---|---|---|---|
| 62% | Jul - Aug 2026 | KPMG AI Quarterly Pulse | 314 US leaders at $1B+ firms; counts building, deploying or developing. |
| 79% | May 2025 | PwC AI agent survey | 300 US senior executives; "being adopted" includes agent features in apps. |
| 23% scaling, 62% at least experimenting | Jun - Jul 2025 | McKinsey | Global respondents; scaling means expanding in at least one function. |
| 40% vs 22% | May - Jun 2026 | McKinsey | Scaling agents: respondents at $1B+ organizations vs smaller ones. |
| 17% | 2026 survey | Gartner CIO and Technology Executive Survey | Organizations that have deployed AI agents; a stricter test. |
The figures run from 17% to 79% because each survey asks a different question of a different group. PwC's 79% asks US executives whether agents are "being adopted", which counts an agent feature switched on inside software the company already pays for. KPMG's 62% counts building, deploying or developing, so a team still writing its first agent is included.
McKinsey separates stages. In 2025, 62% of its respondents were at least experimenting, but only 23% were scaling an agent system. In 2026 it split by size: 40% of respondents at $1 billion firms were scaling, against 22% at smaller ones. Gartner's 17% asks whether agents are deployed, the strictest test.
All of these sample executives or survey volunteers, mostly at large companies. None is an official count of all businesses, and none covers small firms on their own.
What it means
How to use these numbers
Most large companies are working on AI agents, but far fewer run them at scale. If a vendor tells you that four in five companies already use agents, ask which survey and which stage: the 79% counts agent features inside existing apps, while only 23% of McKinsey's 2025 respondents said they were scaling an agent system.
For a small business, the useful number is the size gap. McKinsey found 40% of large organizations scaling agents and 22% of smaller ones, and the smaller share did not move in a year. There is no official survey of agent use by small firms yet, so treat any small-business agent figure with care.
The oversight numbers matter as much as adoption. 63% of large US firms require a person to check agent outputs and 49% have areas where agents may not decide alone. Plan for review time and rules before you hand an agent work that touches customers or money.
Method and sources
How it was measured
Method
- Publishers
- KPMG (AI Quarterly Pulse), McKinsey (State of AI), Gartner (webinar poll and CIO survey), PwC (AI agent survey)
- KPMG sample
- 314 US C-suite and business leaders at organizations with $1 billion or more in revenue, 24 Jul to 25 Aug 2026
- McKinsey sample
- 1,719 online respondents in 97 nations, 4 May to 8 Jun 2026; large means more than $1 billion revenue (36% of respondents)
- Definition
- No shared definition. "Scaling" (McKinsey) means expanding use in at least one function; "building, deploying or developing" (KPMG) includes work in progress; "being adopted" (PwC) includes agent features in apps
- Margin of error
- Not published by these surveys; none is a random sample of all businesses
- Checked at source
- 23 Oct 2026 on kpmg.com, mckinsey.com, gartner.com and pwc.com; KPMG 2025 quarters, McKinsey 2026 and Gartner 2026 figures checked again 30 Oct 2026
- Next release
- KPMG publishes its pulse each quarter; McKinsey's State of AI comes out once a year
- Last changed
- 30 Oct 2026
- Data files
- CSV · JSON (stable links)
Sources
- AI Quarterly Pulse Survey, KPMG LLP (US) Survey
- The State of AI (McKinsey Global Survey), McKinsey & Company Survey
- Gartner newsroom, Gartner, Inc. Research
- PwC research (Global AI Jobs Barometer, AI agent survey), PwC Research
Cite this page
APA
MLA
HTML link
FAQ
More questions people ask
Do small businesses use AI agents?
There is no direct count. The closest figure is McKinsey's 22% of respondents at organizations under $1 billion in revenue scaling agents in 2026, and those are still sizeable companies.
What is a multi-agent system, and how many firms use one?
A setup where several AI agents pass work between them. 25% of large US organizations were developing or implementing one in Q3 2026, up from 6% in each of the two quarters before (KPMG).
Are companies letting AI agents act without human checks?
Mostly not. 63% of large US firms require a person to validate agent outputs (Q1 2026, up from 22% in Q1 2025), and 49% have defined high-risk uses where agents may not decide alone (Q3 2026).
How many companies are scaling coding agents?
About two in ten McKinsey respondents, and 31% of those at large organizations, say they are scaling software coding agents. 32% of all respondents say their firm skipped buying software because agentic coding tools could build it (McKinsey 2026).
Every number on this page was checked at its original source (KPMG LLP (US), McKinsey & Company, Gartner, Inc., PwC). Free to cite with a link to this page.