How to measure AI adoption at work: metrics beyond login counts
AI adoption metrics measure whether people actually use AI tools in their work, how deeply, and with what result. Login counts and seat utilization overstate adoption; the useful measures are weekly active use by team, the share of eligible tasks done with AI, time in AI tools, and the outcomes that follow.
This guide defines each metric and shows how to collect it, then covers how to measure adoption by team, where the data comes from, a benchmark table to fill in for each team and a 90-day measurement plan. Every statistic comes from an independent survey or an official source, listed at the end. Last checked October 23, 2026.
What are AI adoption metrics?
AI adoption metrics are the numbers that show whether AI tools have become part of how work gets done: who uses them, how often and how deeply, for which tasks, and what changed as a result. They sit in four layers, and each answers a different question.
| Layer | Question it answers | Example metrics |
|---|---|---|
| Access | Can people use AI? | Seats assigned, share of staff with an approved tool and training |
| Use | Do they use it? | Weekly active use rate |
| Depth | Is it part of their work? | Frequent use rate, time in AI tools, use cases per user, eligible task coverage |
| Outcome | Did anything improve? | Time saved, change in the team's own output measure |
Most companies measure the first layer and report it as adoption. Tracking all four shows where adoption stalls: people with access who don't use AI need a reason and training, occasional users need more use cases, and frequent users with no change in outcomes need a look at what the time goes into.
How are AI usage metrics different from AI adoption metrics?
AI usage metrics count activity: sign-ins, sessions, prompts or time in AI tools. AI adoption metrics ask whether that activity has become part of the work, so they combine usage with depth and outcomes. A team can show high usage and low adoption when a few enthusiasts account for most of the activity.
Why do login counts and seat utilization mislead?
Login counts and seat utilization measure access, not adoption: one sign-in a month counts the same as daily use, and none of it shows whether work changed. They also miss the AI people use outside the tools you license.
- The measure you pick changes the answer. In Gallup's May 2026 survey of 22,573 employed US adults, 52% used AI in their role at all, 30% used it a few times a week or more and 15% daily: one workforce, three very different adoption rates.
- A used seat isn't a regular user. A seat counts as used whether its holder opened the tool once or worked in it every day. Weekly measures are stricter, and lower: in the third quarter of 2026, 34.8% of employed US adults aged 18 to 64 had used generative AI for work in the previous week (Federal Reserve Bank of St. Louis, Real-Time Population Survey).
- License data misses unapproved tools. In the KPMG and University of Melbourne study of more than 48,000 people in 47 countries, 57% of employees said they hide their use of AI, and almost half admitted using it in ways that contravene company policies. Personal accounts and free tools never appear in your license reports; our guide to shadow AI covers how to find them.
- None of it measures outcomes. Use is the middle of the story. Without an outcome measure, high usage can't be told apart from busywork.
Which AI adoption metrics matter? The 8 to track
These eight metrics cover use, depth and outcomes. The targets are our editorial starting points for a first 90 days, not industry benchmarks: published benchmarks for most of these metrics don't exist, so set your own from your baseline and raise them each quarter.
"Eligible people" means the people whose roles include tasks you've listed as suitable for AI and who have access to an approved tool.
| Metric | Definition | How to collect it | A reasonable starting target |
|---|---|---|---|
| 1. Weekly active use rate | Share of eligible people who used an approved AI tool at least once in the week | Active-user reports in the AI tool's admin console, single sign-on logs, or app and website analytics on work computers | 60% of eligible people by day 90 |
| 2. Frequent use rate | Share of weekly active users who use AI on three or more days a week | Daily active users from the admin console, or a weekly pulse question | Half of weekly active users |
| 3. Time in AI tools | Hours per active user per week in AI apps and websites | App and website analytics grouped into an AI category; session data from admin consoles where available | No fixed number: rising during rollout, then steady; compare teams that do similar work |
| 4. Use cases per user | Number of distinct kinds of task a person uses AI for | A short quarterly survey with a checklist of your use cases | Three or more for frequent users |
| 5. Eligible task coverage | Share of the tasks you listed as suitable for AI that are now done with AI | A task list per team, then a survey or a sample of completed work | Half of the listed tasks by day 90 |
| 6. Time saved per active user | Hours a week saved on the measured tasks | Timed samples of the same task with and without AI; survey answers only as a cross-check | Above the break-even hours from the AI ROI calculator |
| 7. Outcome change | Change in the team's own output measure since the baseline | The system where the measure already lives: help desk, CRM, code repository or month-end close calendar | A change larger than the usual month-to-month swing |
| 8. Approved-tool share | Share of time in AI tools spent in approved tools | App and website analytics compared with your approved AI tools register | 90% of AI time in approved tools |
Why use cases per user matters: among US employees who use AI, the share who say it has had a somewhat or extremely positive effect on their productivity rises from 45% of those who use it for one or two purposes to 66% for three or four, 78% for five or six and 90% for seven or more (Gallup, May 2026). Breadth is where use turns into results, which is why the starting target is three.
Why the rollout matters as much as the tool: in a late-2024 survey of 25,000 workers in 11 occupations across 7,000 Danish workplaces, about 40% had used AI chatbots at work where the employer neither encouraged use nor provided tools or training, against 93% where the employer encouraged use, provided an enterprise chatbot and trained people (Humlum and Vestergaard, NBER working paper, revised March 2026).
How do you measure AI adoption by team and role?
Calculate every metric per team, with that team's eligible people as the denominator and its own output measure, then compare teams that do similar work. Report by team, not by person: the goal is to find where adoption stalls and why, and a ranking of individuals invites people to perform for the metric instead of the work.
A company average hides the differences. In Gallup's May 2026 survey, daily AI use was 42% in technology, 27% in finance and 22% in professional services, while daily use in each of the other seven industries Gallup reports ranged from 9% to 15%. Managers make a difference too: in Gallup's February 2026 survey of 23,717 employed US adults, among employees whose organizations make AI tools available, 78% of those who strongly agree their manager supports AI use were frequent users, against 44% of those who don't.
Three rules keep team comparisons fair:
- Use the right denominator. Count only eligible people, so a team where half the roles have no AI tasks isn't marked down for it.
- Pick one outcome per team. For example: tickets resolved per agent-hour in support, days from first meeting to proposal in sales, pull request cycle time in engineering, days to close the month in finance, drafts shipped per week in marketing.
- Compare like with like. Compare each team with its own baseline and with teams that do similar work, not engineering with sales.
A benchmark table you can fill in
Published AI adoption benchmarks by team are scarce, so the most useful benchmark is your own baseline. Copy this table into a spreadsheet or print this page, fill in each team's number before rollout, and set a target for day 90 using the starting targets above.
| Team | Metric | Your number | Target |
|---|---|---|---|
| Whole company | Weekly active use rate | ||
| Whole company | Approved-tool share of AI time | ||
| Customer support | Weekly active use rate | ||
| Customer support | Time in AI tools per active user per week | ||
| Customer support | Tickets resolved per agent-hour | ||
| Sales | Weekly active use rate | ||
| Sales | Use cases per user | ||
| Sales | Days from first meeting to proposal | ||
| Engineering | Frequent use rate | ||
| Engineering | Time in AI tools per active user per week | ||
| Engineering | Pull request cycle time | ||
| Finance | Eligible task coverage | ||
| Finance | Days to close the month | ||
| Marketing | Time saved per active user | ||
| Marketing | Drafts shipped per week | ||
| Your team |
Where does AI adoption data come from?
Four methods supply most AI adoption data: the admin dashboards of the AI tools you license, single sign-on and identity logs, surveys, and app and website analytics on work computers. Each sees a different slice, so most companies combine two or three, and take outcome data from the systems each team already works in.
| Method | What it measures well | What it misses |
|---|---|---|
| Admin dashboards of the AI tools you license | Active users, sessions and sometimes feature use inside those tools | AI tools you don't license, personal accounts, AI features inside other apps, and outcomes |
| Single sign-on and identity logs | Who signs in to approved AI tools with a company account, and how often | What happens after sign-in, tools used with personal accounts, and time spent |
| Surveys and pulse questions | Use cases, reasons, barriers, perceived time saved, and use on personal devices | Accuracy: self-reported time savings can be far off, and people underreport use they hide |
| App and website analytics on work computers | Time in every AI app and website on work computers, approved or not, by team, on or off the office network | Personal phones and computers, and what people type into AI tools, if the analytics record usage only |
Surveys are best for the why, not the how much. In a randomized trial of early-2025 AI tools, 16 experienced open-source developers took 19% longer to finish their tasks when AI was allowed, yet afterwards they estimated that AI had sped them up by 20% (METR, July 2025). METR said in February 2026 that those results no longer reflect current AI tools, but the gap between felt and measured speed is the reason to time tasks rather than ask about them.
What does a 90-day AI adoption measurement plan look like?
A 90-day plan sets a baseline in the first two weeks, launches with targets and rules, reviews use by team every week, and compares outcomes with the baseline at the end. By day 90 you should know which teams use AI, for what, and whether it paid off.
- Days 1 to 14: baseline. List the AI-suitable tasks for each team and who is eligible. Record each team's outcome measure for the previous month and its current time in AI tools, approved or not, and run a short survey on current use. The AI readiness assessment shows the gaps in data, process, skills and governance to close before you start.
- Days 15 to 30: launch with rules. Give eligible people an approved tool, training and a short policy; the AI acceptable use policy template covers approved tools, data that must never go into AI, disclosure and human review. Fill in the benchmark table with each team's starting numbers and day 90 targets.
- Days 31 to 60: weekly review by team. Track weekly active use, frequent use and time in AI tools for each team. Where a team lags, ask its manager and a few people why: missing use cases, no time to learn, or a tool that doesn't fit the work. Start timing samples of the tasks you're measuring.
- Days 61 to 90: outcomes and decisions. Compare each team's outcome measure with its baseline, estimate the time saved that became output, and run the numbers through the AI ROI calculator. Decide team by team: expand, change the tool or the training, or stop paying for seats nobody uses.
After day 90, review the same metrics every quarter: new AI tools, and new AI features in the apps you already use, keep arriving.
How does Intelogos measure AI tool usage?
Intelogos is a workforce analytics platform that shows how work time is spent across apps and websites, without screenshots or keystroke logging, and lets managers ask questions about it in plain language. It records the AI apps and websites people use on work computers like any other tool and groups them in a built-in AI time category for AI tools and assistants. On the Tools page, you can filter to the AI category and to a department to see which AI tools that department uses and for how long, and the Time Categories chart on the Dashboard and Profile pages shows time in AI tools next to other kinds of work.
That gives you two of the eight metrics from measured use instead of survey answers: time in AI tools by department, and the approved-tool share, by comparing the AI tools on the Tools page with your approved AI tools register. Pick a date range before a rollout and one after it to see where time in AI tools grew and where it didn't. On the AI Intelligence and Enterprise plans, managers can also ask these questions in plain language with Ask AI; the AI adoption prompts for managers are ready to copy.
What it doesn't capture matters as much. Intelogos doesn't take screenshots. Intelogos doesn't record what people type; it records only whether keyboard or mouse input happened. So it shows that a department spent hours in an AI tool, not what anyone pasted into it. It covers work computers with the Intelogos agent installed, not personal phones. An AI tool Intelogos doesn't recognize yet appears in the Other time category until someone moves it to AI on the Tools page. The Tools page is available to owners, and to admins and managers with the App Management permission.
Frequently Asked Questions
What is the current AI adoption rate?
About half of US workers use AI at work, depending on how it's measured. In May 2026, 52% of US employees used AI in their role, 30% a few times a week or more and 15% daily (Gallup, 22,573 employed adults). In March 2026, about 56% of workers had used AI on the job for at least one of 11 tasks (US Census Bureau, Household Trends and Outlook Pulse Survey). In the third quarter of 2026, 34.8% of employed adults aged 18 to 64 used generative AI for work in the previous week (Federal Reserve Bank of St. Louis). Among businesses, 19.8% of US firms used AI in a business function as of May 3, 2026, and 37% of firms with 250 or more employees (US Census Bureau, Business Trends and Outlook Survey).
What are examples of AI adoption metrics?
The weekly active use rate, the frequent use rate, time in AI tools, use cases per user, eligible task coverage, time saved per active user, the change in the team's own output measure, and the share of AI time spent in approved tools. Seats assigned and login counts are access metrics: they show who could use AI, not who does.
What is the 10/20/70 rule in AI?
It's a guideline from the consulting firm BCG for where AI effort should go: 10% of resources to algorithms, 20% to technology and data, and 70% to people and processes. BCG says the companies leading on AI follow it, based on its 2024 survey of about 1,000 senior executives in 59 countries. For measurement, the 70% is the point: adoption metrics track the people and process part, which seat counts don't show.
How often should you measure AI adoption?
Measure use weekly during a rollout and monthly after it: weekly active use, frequent use and time in AI tools. Measure use cases, task coverage, time saved and outcomes quarterly, because they move slowly and need a baseline to compare with.
Sources
Every statistic in this guide comes from the sources below. All were checked on October 23, 2026.
- Gallup, "Organizational AI Adoption Jumps Six Points", July 20, 2026: Gallup Panel web survey of 22,573 employed US adults, May 6 to 20, 2026, margin of error ±0.9 percentage points. 52% use AI in their role, 30% a few times a week or more and 15% daily; the share reporting a positive effect on productivity by number of purposes (45%, 66%, 78%, 90%).
- Gallup, "Artificial Intelligence" indicator: daily AI use by industry as of May 2026 (technology 42%, finance 27%, professional services 22%, the other seven industries 9% to 15%).
- Gallup, "AI in the Workplace: What Separates Adopters and Holdouts", April 12, 2026: Gallup Panel web survey of 23,717 employed US adults, February 4 to 19, 2026, margin of error ±0.9 percentage points. Frequent use by manager support (78% against 44%) among employees in organizations that make AI tools available.
- Federal Reserve Bank of St. Louis, FRED, "Generative Artificial Intelligence, Use Last Week for Work: Employed Adults", updated October 20, 2026: the Real-Time Population Survey's Generative Artificial Intelligence Adoption Tracker by Alexander Bick, Adam Blandin and David Deming, a nationally representative online survey of US adults aged 18 to 64. 34.8% in the third quarter of 2026.
- US Census Bureau, "About a Third of Workers Who Used AI in the Last Week Said They Completed Tasks One to Two Hours Faster", August 11, 2026, figures revised September 24, 2026: the March 2026 Household Trends and Outlook Pulse Survey, March 13 to 30, 2026, about 12,521 respondents. About 56% of workers have used AI on the job for at least one of 11 tasks.
- US Census Bureau, "Large Firms With at Least 20 Employees Biggest AI Users", May 26, 2026: the Business Trends and Outlook Survey, a biweekly, nationally representative survey of US businesses. 19.8% of businesses used AI in any business function as of May 3, 2026; 37% of firms with at least 250 employees. The question changed in November 2025 from AI use in producing goods or services to AI use in any business function.
- KPMG and the University of Melbourne, "Trust, attitudes and use of artificial intelligence: A global study 2025", April 2025: more than 48,000 people in 47 countries, November 2024 to January 2025. 57% hide their use of AI; almost half admit using AI in ways that contravene company policies.
- Anders Humlum and Emilie Vestergaard, "Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI," NBER Working Paper 33777, May 2025, revised March 2026: surveys of AI chatbot adoption in 11 occupations in Denmark, run with Statistics Denmark, the latest in late 2024 with 25,000 workers across 7,000 workplaces. About 40% use without employer initiatives and 93% with encouraged use, an enterprise chatbot and training.
- METR, "Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity", July 10, 2025: randomized trial with 16 experienced open-source developers and 246 tasks. And METR, "We are Changing our Developer Productivity Experiment Design", February 24, 2026.
- BCG, "AI Adoption in 2024: 74% of Companies Struggle to Achieve and Scale Value", October 24, 2024: survey of about 1,000 CxOs and senior executives in 59 countries, with the 10% algorithms, 20% technology and data, 70% people and processes rule.