Free data tool · U.S. Census Bureau
Which industries use AI? Latest U.S. Census data
Real numbers for every industry, company size, Florida and the Miami metro area, straight from the Census Bureau, with the margin of error beside each figure.
Data through . Census files imported .
Census now labels this topic “Super Intelligence or SI (formerly referred to as Artificial Intelligence or AI)”. We say “AI” because it is what people search for. Why SI?
Share of U.S. businesses using AI
Biweekly Census results since November 17, 2025, the first collection with the current question wording.
View the data as a table
| Collection ended | Using now | Expected |
|---|---|---|
| October 18, 2026 | 24.6% | 28.6% |
| October 4, 2026 | 23.8% | 27.7% |
| September 20, 2026 | 23.8% | 27.6% |
| September 6, 2026 | 23.2% | 27.3% |
| August 23, 2026 | 22.4% | 25.9% |
| August 9, 2026 | 21.8% | 25.9% |
| July 26, 2026 | 21.5% | 24.3% |
| July 12, 2026 | 21.7% | 24.5% |
| June 28, 2026 | 20.6% | 23.7% |
| June 14, 2026 | 20.6% | 23.9% |
| May 31, 2026 | 20.6% | 23.7% |
| May 17, 2026 | 19.5% | 22.7% |
| May 3, 2026 | 19.8% | 22.8% |
| April 19, 2026 | 19.8% | 23.0% |
| April 5, 2026 | 19.1% | 22.5% |
| March 22, 2026 | 18.9% | 22.3% |
| March 8, 2026 | 18.2% | 21.8% |
| February 22, 2026 | 17.5% | 21.0% |
| February 8, 2026 | 18.9% | 22.1% |
| January 25, 2026 | 18.2% | 22.1% |
| January 11, 2026 | 17.7% | 22.4% |
| December 28, 2025 | 17.8% | 21.3% |
| December 14, 2025 | 17.2% | 20.9% |
| November 30, 2025 | 17.3% | 21.1% |
All U.S. businesses
24.6%
using AI now, ±0.5 points
28.6% expect to within 6 months
Florida
Above U.S.26.8%
using AI now, ±2.1 points
32.3% expect to within 6 months
Miami metro area
Above U.S.28.7%
using AI now, ±3.3 points
34.3% expect to within 6 months
At a glance
- 24.6% of U.S. employer businesses reported using AI in the last two weeks (collection ending October 18, 2026), and 28.6% expect to use it within six months.
- That is up from 17.3% in November 2025, the first period with the current survey wording.
- Among industries measured within plus or minus 5 points, rates run from 10.7% in Accommodation and Food Services to 46.4% in Information.
- Larger firms use AI more: 43.6% of firms with 250 or more employees, against 24.3% of firms with 1 to 4.
- Florida is at 26.8% (plus or minus 2.1 points). The Miami metro area is at 28.7% (plus or minus 3.3 points), above the national rate.
Explore the data
Who is using AI, and who expects to?
Census collection ending October 18, 2026. Every figure shows its 95 percent margin of error so you can see how much to trust it.
Industries compared
Every industry sector Census reports, with the U.S. rate marked on each bar.
19 of 19 industries. Select a name to see what it means for you.
How to read this
- Above U.S. above the U.S. rate, beyond the margin of error
- Similar to U.S. within the margin of error
- Below U.S. below it
- Too imprecise to rank margin wider than 8 points
- U.S. rate on each bar
| Compared with U.S. | Position | ||||
|---|---|---|---|---|---|
| NAICS 51 | 46.4% ±4.0 | 51.1% ±4.1 | +4.7 pts | Above U.S. | Top quarter |
| NAICS 54 | 43.8% ±1.8 | 48.8% ±1.5 | +5.0 pts | Above U.S. | Top quarter |
| NAICS 52 | 40.5% ±3.0 | 47.4% ±2.7 | +6.9 pts | Above U.S. | Top quarter |
| NAICS 55 | 36.4% ±24.6 | 38.8% ±21.6 | +2.4 pts | Unclear | Too imprecise to rank |
| NAICS 61 | 33.4% ±3.8 | 36.7% ±3.8 | +3.3 pts | Above U.S. | Top quarter |
| NAICS 53 | 29.0% ±2.0 | 35.4% ±2.3 | +6.4 pts | Above U.S. | Top quarter |
| NAICS 62 | 27.6% ±1.7 | 32.9% ±2.3 | +5.3 pts | Above U.S. | Middle |
| NAICS 31-33 | 22.9% ±2.6 | 27.0% ±2.1 | +4.1 pts | Similar to U.S. | Middle |
| NAICS 56 | 22.7% ±2.8 | 25.5% ±2.6 | +2.8 pts | Similar to U.S. | Middle |
| NAICS 42 | 21.4% ±1.7 | 25.7% ±1.9 | +4.3 pts | Below U.S. | Middle |
| NAICS 71 | 21.4% ±2.5 | 24.6% ±2.5 | +3.2 pts | Below U.S. | Middle |
| NAICS 44-45 | 18.5% ±1.4 | 22.5% ±1.7 | +4.0 pts | Below U.S. | Middle |
| NAICS 21 | 16.8% ±5.8 | 16.6% ±5.6 | -0.2 pts | Below U.S. | Middle |
| NAICS 22 | 16.7% ±11.3 | 13.4% ±9.9 | -3.3 pts | Unclear | Too imprecise to rank |
| NAICS 23 | 16.1% ±1.5 | 19.1% ±1.6 | +3.0 pts | Below U.S. | Bottom quarter |
| NAICS 81 | 14.6% ±2.0 | 16.5% ±2.1 | +1.9 pts | Below U.S. | Bottom quarter |
| NAICS 48-49 | 13.5% ±3.0 | 16.0% ±4.1 | +2.5 pts | Below U.S. | Bottom quarter |
| NAICS 72 | 10.7% ±2.1 | 13.9% ±2.0 | +3.2 pts | Below U.S. | Bottom quarter |
| NAICS 11 | 6.0% ±5.1 | 9.0% ±5.0 | +3.0 pts | Below U.S. | Bottom quarter |
- 46.4% ±4.0Expected 51.1%Above U.S.Top quarter
- 43.8% ±1.8Expected 48.8%Above U.S.Top quarter
- 40.5% ±3.0Expected 47.4%Above U.S.Top quarter
- 36.4% ±24.6Expected 38.8%UnclearToo imprecise to rank
- 33.4% ±3.8Expected 36.7%Above U.S.Top quarter
- 29.0% ±2.0Expected 35.4%Above U.S.Top quarter
- 27.6% ±1.7Expected 32.9%Above U.S.Middle
- 22.9% ±2.6Expected 27.0%Similar to U.S.Middle
“±” is the 95 percent margin of error. “Expected change” is expected minus current use: what businesses told Census they expect, not a forecast. Position is a band (top quarter, middle half, bottom quarter) among industries with a reliable estimate. Neighbouring industries are often statistically close.
Find your industry
What does this mean for your business?
Pick an industry to see its latest numbers, how it compares, and practical places to start.
Professional, Scientific, and Technical Services NAICS 54
Using AI in the last two weeks
43.8%
±1.8 points at 95% confidence
Expect to use it within six months
48.8%
±1.5 points at 95% confidence
That is above the U.S. rate of 24.6%. It sits in the top quarter of the 17 industries Census measures precisely enough to rank. Neighbouring industries are statistically close, so read this as a band, not an exact position.
- Miami metro area, all industries
- 28.7% ±3.3 using AI now
- Florida, all industries
- 26.8% ±2.1 using AI now
Where professional, scientific, and technical services businesses often start
- Research and report drafting
- Client communication templates
- Data analysis and charts
We can set up secure AI workflows for your consulting or technical team.
- Microsoft 365 and Copilot plans
- Google Workspace and Gemini
- Cloud services
- AI acceptable-use policy guide
- IT policy templates
- Security compliance guide
Readiness guidance is based on iFeeltech's small business IT experience. It is not part of the Census data.
What businesses do with it
Where AI is used, and why some say no
From Census's AI supplement, collected November 17, 2025 to February 8, 2026. It is a separate, one-time survey with its own questions and period, so these figures will not match the headline numbers above.
- Used AI in any business function in the last two weeks
- 17.9%
- 71.3% said no and 10.8% did not know.
- Expect to use AI in any business function in the next six months
- 21.6%
- Employees used AI in work-related tasks in the last six months
- 22.6%
- Employees used generative AI in work-related tasks in the last six months
- 20.8%
Business functions using AI
Base: all U.S. employer businesses, any use of AI in the function over the last six months.
- Sales and marketing14.3%
- Strategy and business development12.4%
- Information technology11.4%
- Research and development11.2%
- Public relations and communication9.3%
- Management and administration8.1%
- Customer service7.6%
- Legal and compliance7.1%
What staff use generative AI for
Base: businesses whose employees used generative AI in work-related tasks.
- Writing or editing documents, emails, or communications85.4%
- Searching for information or technical help49.9%
- Interpreting, analyzing, translating, or summarizing documents44.6%
- Information processing, paperwork, or filing34.7%
- Developing or researching new projects, processes, or products30.0%
- Data analysis or visualization19.1%
- Tutoring, training, or learning16.1%
- Customer support15.5%
Why businesses do not plan to use AI
Base: businesses that do not plan to use AI in the next six months.
- AI is not applicable to this business61.6%
- Lack of knowledge on the capabilities of AI22.0%
- Concerns about privacy/security20.7%
- AI is not a mature enough technology yet13.0%
- Concerns about bias8.6%
- Lack of skilled workforce7.1%
- Too expensive6.9%
- Lack of required data5.1%
What businesses changed to use AI
Base: businesses that used AI in any business function.
- Developed new workflows15.4%
- Trained current staff to use AI15.0%
- Purchased cloud services or cloud storage8.5%
- Purchased computing power or specialized equipment or software8.2%
- Changed data collection or data management practices6.8%
- Used vendors or consulting services to install or integrate AI3.8%
- Hired staff trained in AI1.3%
- Other4.8%
- This business did not make any changes to use AI64.3%
How AI changed the work
Base: businesses that used AI in any business function.
- Supplement or enhance a task performed by an employee43.7%
- Introduce a new task not previously done by an employee10.6%
- Perform a task previously done by an employee10.1%
- None of the above51.5%
Effect on total employment
Base: businesses that used AI in any business function.
- No change95.7%
- Increased2.3%
- Decreased2.0%
The longest lists show their top answers; the JSON download has every answer. Businesses could choose more than one answer where a question allows it, so percentages in a card can add up to more than 100.
Before you roll anything out
An eight-step AI readiness checklist
Practical steps we walk small businesses through. This is our guidance, not Census data.
Check who can see which files
AI tools can surface anything a person has access to. Tidy SharePoint, OneDrive and Google Drive permissions first. SharePoint permissions audit
Write a short acceptable-use policy
Say which tools are approved and what staff may and may not paste into them. AI acceptable-use policy guide
Turn on multi-factor authentication everywhere
AI-assisted phishing is more convincing. MFA limits the damage when a password leaks. Business Security Score
Use a password manager
Shared logins for AI tools are a common weak spot. Give each person their own.
Decide what data is sensitive
Client records, financials, health information and contracts should have explicit rules before any AI tool touches them.
Name the approved tools
One or two vetted tools with business terms beat a dozen personal accounts. Microsoft 365 and Copilot plans
Train staff on AI-driven scams
Deepfake voice and email requests for payments are now realistic. Agree a call-back rule. AI vishing and deepfake BEC guide
Check endpoint protection and logging
Make sure every laptop is patched, protected and reporting before adding new software.
Readiness guidance is based on iFeeltech's small business IT experience. It is not part of the Census data.
Methodology and data
How these numbers are made
Source. The U.S. Census Bureau’s Business Trends and Outlook Survey (BTOS), a nationally representative survey of U.S. employer businesses collected in two-week periods. Every percentage on this page is a Census figure. We do not model, estimate or fill gaps.
What the numbers mean. “Using AI now” is the share of businesses that used AI in the last two weeks. “Expected” is the share that expect to use it in the next six months. The expected change is expected minus current, which is what businesses told Census they expect, not a forecast.
Latest period. The collection ending October 18, 2026. The trend chart starts at November 17, 2025 because that is when Census began using the current question wording.
Question wording changes.
- November 17, 2025: Census reworded the core question to ask about AI use "in any business function" instead of "in producing goods or services".
- October 19, 2026: Census updated the instrument wording and relabelled the topic “Super Intelligence or SI (formerly referred to as Artificial Intelligence or AI)”, citing Executive Order 14434. This page does not yet contain data from that collection. We keep the term “AI” because it is what people search for.
Trends are not strictly comparable across a wording change.
Margins of error. Each figure shows a 95 percent margin of error, 1.96 times the Census standard error. When a margin is wider than 8 points the number is shown but not ranked, and it is not compared with the national rate. Comparisons use both standard errors, so a gap smaller than the combined margin is called “similar”.
How the position labels work
Among industries with a reliable estimate, the top quarter are labelled “Top quarter”, the bottom quarter “Bottom quarter”, and the rest “Middle”. Neighbouring industries are often statistically close, so read the label as a band, not an exact rank. The range sentence in the findings only names industries measured within 5 points.
How places are defined
Florida is Census’s state estimate. The Miami metro area is Census’s “Miami-Fort Lauderdale-Pompano Beach, FL MSA”, one of the 25 largest metro areas it publishes. Smaller samples mean wider margins.
Business functions, tasks and reasons
These come from Census’s one-time AI supplement (November 17, 2025 to February 8, 2026). We show the percentages exactly as Census tabulates them, for the group named on each card. Census’s working paper CES-WP-26-25 reports some function shares on different bases, so its percentages will not match these cards.
What iFeeltech adds
The position labels, plain-language comparisons, the Miami view and the readiness guidance. The guidance is our small business IT experience, not Census data.
Questions about AI adoption
What percent of U.S. businesses use AI?
24.6% of U.S. employer businesses said they used AI in the last two weeks, in the Census Bureau's survey collection ending October 18, 2026. 28.6% expect to use it within six months. The margin of error is about 0.5 points.
Which industries use AI the most?
Among industries measured within plus or minus 5 points, the highest current rates are Information (46.4%), Professional, Scientific, and Technical Services (43.8%), and Finance and Insurance (40.5%). The lowest is Accommodation and Food Services (10.7%). Neighbouring industries are often statistically close, so treat the order as a band rather than an exact ranking. Every rate has a margin of error, shown in the explorer above.
How many small businesses use AI?
24.3% of firms with 1 to 4 employees and 22.6% of firms with 5 to 9 employees used AI in the last two weeks, against 43.6% of firms with 250 or more employees.
Are Miami and Florida businesses using AI?
Census puts Florida at 26.8% and the Miami-Fort Lauderdale-Pompano Beach metro area at 28.7% (plus or minus 3.3 points), against 24.6% nationally. These cover all industries together.
What do businesses use AI for?
In Census's AI supplement, collected November 17, 2025 to February 8, 2026, the most common business functions were sales and marketing (14.3% of all firms), strategy and business development (12.4% of all firms), and information technology (11.4% of all firms). Among firms whose employees used generative AI, 85.4% used it for writing or editing documents, emails or communications.
Why do some businesses not plan to use AI?
Among firms that do not plan to use AI in the next six months, the top reasons were "ai is not applicable to this business" (61.6%), "lack of knowledge on the capabilities of ai" (22.0%), and "concerns about privacy/security" (20.7%). Firms could pick more than one reason.
Why does the Census Bureau say "SI" instead of AI?
Census now labels this topic "Super Intelligence or SI (formerly referred to as Artificial Intelligence or AI)" and says its survey wording changed beginning with the October 19, 2026 collection because of Executive Order 14434. We use "AI" because it is the term most people search for. The latest figures cover the collection ending October 18, 2026, before that wording change.
How current is this data?
The latest Census collection shown here ended October 18, 2026. Census publishes new results about every two weeks, and this page is refreshed from the official Census files when a new release appears.
Why do other reports show different percentages?
Surveys differ in question wording, time period and who is counted. Census's own AI supplement found 17.9% of businesses used AI in any business function in the last two weeks, collected November 2025 to February 2026 (71.3% said no and 10.8% did not know). The 24.6% above is the regular biweekly measure for the latest period. Compare figures only when they share a question, a period and a base.
How accurate are these numbers?
They are survey estimates, not a count. Each figure shows a 95 percent margin of error (1.96 times the Census standard error). If a margin is wider than 8 points, we show the number but do not rank it or compare it with the national rate.
Can I use this data in my own work?
Yes. The numbers come from the U.S. Census Bureau, and you can download them as CSV or JSON from this page. Please credit the Census Bureau and link to this page.
Thinking about AI for your team?
We can help you start safely
We review access, security and policies so AI tools help your staff without exposing client data. Talk to a Miami IT consultant, no pressure.