AI in Brazilian companies: 17% use it, Cetic.br shows

AI use in Brazilian companies reached 17% in 2025, according to Cetic.br/NIC.br’s TIC Empresas 2025, which interviewed 4,174 companies with ten or more employees. The share had been 13% in both 2023 and 2024, and the 2025 jump came from both ends: large companies went from 38% to 50%, and small ones from 10% to 15%. In industrial companies with 100 or more employees, IBGE measured 41.9% AI use in 2024, up from 16.9% in 2022. Among companies that use AI, marketing and sales is the most cited purpose, and among those that do not, the lack of trained people shares first place with the lack of knowledge about which AI fits the business. My reading of these numbers is that the gap between Brazilian companies that use AI and those that do not depends less on budget and technology than on people who know how to choose the use, run it and measure it.

At Iasbeck&Co we implement CRM and AI agents in B2B operations and train the teams that work with them, and boards often ask me for a reliable number to take to their directors. I gathered the official surveys here, with the sample, year and definition of each, and my reading for executives, team leaders and HR. The implementation roadmap is in AI for companies.

How many companies use AI in Brazil?

The answer depends on the survey. TIC Empresas covers companies with ten or more employees across sectors, PINTEC Semestral covers industrial companies with 100 or more employees, and the São Paulo survey run for the OECD covers medium and large companies in manufacturing and technology in that state.

The overall picture: TIC Empresas 2025

TIC Empresas reached its 17th edition in 2025. Fieldwork was done by phone between February 2025 and January 2026, and Cetic.br estimates that 93,475 companies used AI applications in the period. The results table carries a note that matters for the reading: the indicator was collected only among companies that have IT specialists, an IT area or an IT department, and Cetic.br reports the result over the total of companies. A company with nobody in IT counts as a non-user, even if the owner uses an AI assistant every day.

Segment (TIC Empresas 2025)	Used AI in 2025
Total (10 or more employees)	17%
Small (10 to 49 employees)	15%
Medium (50 to 249 employees)	32%
Large (250 employees or more)	50%
Information and communication	49%
Professional, real estate and administrative services	24%
Retail and wholesale trade	17%
Manufacturing	14%
Accommodation and food services	10%

Source: Cetic.br/NIC.br, table H9 of TIC Empresas 2025.

Industry: IBGE’s PINTEC Semestral 2024

PINTEC Semestral is an experimental IBGE survey run with ABDI and UFRJ, based on a sample of 1,731 companies that represents the 10,167 mining and manufacturing companies with 100 or more employees. The respondent fills in the questionnaire online, with 2024 as the reference year. In that universe, 4,261 companies used AI in at least one business area, or 41.9%, and AI was the advanced digital technology that grew the most since 2022. By size, the PINTEC Semestral 2024 publication shows 57.5% use among companies with 500 or more employees, 42.5% among those with 250 to 499 and 36.1% among those with 100 to 249.

The figure seems to contradict the 14% that TIC Empresas found in manufacturing, and the difference comes from size: PINTEC does not survey small manufacturers, where most of the country’s companies are.

The most conservative cut: São Paulo for the OECD

In 2023, the SEADE Foundation and Cetic.br applied in São Paulo the same questionnaire that the OECD, BCG and INSEAD used with companies in G7 countries. The OECD report published in 2025 shows that, among 2,561 medium and large manufacturing and technology services companies active in the state, only 7% used any type of AI, with fieldwork between February and July 2023. Only 167 companies were users, too small a base to read by size or sector. The date explains part of the distance from the other sources, because generative AI had barely reached companies.

Why do large companies use AI three times more than small ones?

TIC Empresas shows 50% use among large companies and 15% among small ones, and Eurostat finds the same pattern in Europe, with 55.03% against 17% in 2025. Eurostat itself raises three possible explanations: the complexity of implementing AI, scale, which lets large companies get more from the investment, and cost, easier to absorb for those who bill more. I add one that I see in projects: a large company has someone whose job includes choosing the tool, configuring it and following its use, while in a small company that decision falls to the owner, between selling, hiring and paying suppliers.

The way companies adopt confirms this reading. According to TIC Empresas 2025, 80% of companies that used AI bought ready-to-use solutions, and 54% hired external suppliers to modify or develop the systems. In-house development is more common among large companies, and 37% of them said the systems they use were developed by their own staff. AI reaches Brazilian companies through purchase, and what separates a useful purchase from a forgotten license is having someone in-house who decides where to use it and checks the result.

What do companies use AI for, and where do sales and service show up?

The most cited purpose among Brazilian companies that use AI is marketing or sales, at 44%, according to table H10 of TIC Empresas 2025. Among small users the share rises to 46%, and among large ones it stays at 42%, behind digital security (47%). In industrial companies with 100 or more employees, IBGE found the same pattern with a different yardstick: among companies that used AI, 87.9% applied it in administration, 75.2% in sales and 73.1% in product, process and service development, ahead of production (52.0%). In São Paulo, customer-facing services led, with 49% of users.

Purpose (companies that used AI, TIC Empresas 2025)	Total	Small	Large
Marketing or sales	44%	46%	42%
Organization of business administration processes	35%	38%	40%
Business management	34%	36%	33%
Digital security	34%	32%	47%
Human resources or recruitment	24%	21%	35%
Production processes	23%	22%	28%
Logistics	22%	21%	31%

Source: Cetic.br/NIC.br, table H10. The base is the total of companies that used AI, and a company can mark more than one purpose.

In TIC Empresas, the most cited application is the automation of processes and workflows, at 68% of users, and the fastest growing were written language analysis, from 22% to 38%, and natural language generation, from 20% to 30%, the territory of generative AI. In PINTEC, IBGE measured the degree of use, and AI appears in a localized way in most companies: only 12.9% of industrial companies using AI in sales say the use is comprehensive in that area. My reading is that AI entered the commercial area in the hands of individuals, like the salesperson who writes emails with an assistant, and has not yet become a process of the area, with rules, CRM data and an indicator. The state of AI in companies shows the same pattern in international surveys.

Is Brazil behind Europe in AI use?

Eurostat measured 19.95% AI use among EU companies with ten or more employees in 2025, up from 13.48% in 2024, in a survey of 157,000 of the bloc’s 1.53 million companies. The distance from Brazil’s 17% is smaller than the difference between European countries, which ranges from 5.21% in Romania to 42.03% in Denmark. In information and communication, Europe reaches 62.52% and Brazil 49%.

The purposes also look alike. Among European companies that use AI, 34.70% apply it to marketing or sales and 31.05% to organizing business administration processes or management, the two most cited purposes. The comparison has limits, because the Brazilian question only reaches companies with an IT area and the European one goes to all. I use these numbers to show that both sides start in the same place, without measuring who is ahead.

What holds back AI adoption in Brazilian companies?

The lack of prepared people appears at or near the top in every survey. In TIC Empresas 2025, among companies that did not use AI, the most cited reasons were the lack of trained people to use these technologies (14%) and the lack of knowledge about AI technologies suited to the company’s activities (14%), followed by cost, incompatibility with existing systems and the view that AI is not useful for the business, all at 13%. The percentages look low because the base is the total of companies that did not use AI. My reading is that the same rule as the usage indicator applies, with the question asked only to companies with an IT area, so the ranking of reasons says more than each percentage. Among large non-users the shares rise: cost at 42%, lack of trained people at 36% and incompatibility at 34%.

Reason for not using AI (TIC Empresas 2025)	Total	Small	Medium	Large
Lack of trained people	14%	12%	24%	36%
Lack of knowledge about suitable AI	14%	13%	26%	33%
Costs seem too high	13%	11%	25%	42%
Incompatibility with existing systems	13%	12%	19%	34%
AI is not useful for the company	13%	12%	14%	24%
Data availability or quality	10%	10%	16%	28%
Data protection and privacy	10%	9%	18%	27%
Lack of clarity about legal consequences	9%	9%	14%	22%

Source: Cetic.br/NIC.br, table H13. Base: total of companies that did not use AI.

PINTEC asks about advanced digital technologies as a group, without isolating AI, and the result points the same way. Among industrial companies that adopted any of these technologies, high costs were the most cited difficulty (78.6%), followed by the lack of qualified staff in the company (54.2%) and security and privacy risks (47.2%). Among those that adopted none, the lack of qualified staff was the reason that grew the most, from 46.4% in 2022 to 60.6% in 2024. In Europe, among companies that considered using AI and did not, lack of expertise was the reason for 70.89%, ahead of lack of legal clarity (52.52%).

Why the lack of qualified people is the barrier training solves

Cost and system incompatibility depend on budget and IT projects, which commercial leaders almost never control. The lack of trained people and the lack of knowledge about which AI fits the company are barriers of a different kind, because the company solves both by training the team it already has. Together they hold the top two places in TIC Empresas, and part of the cost reason also fades when someone knows how to estimate the return, the subject of AI ROI.

The São Paulo data for the OECD shows the difference in stance. Among São Paulo companies that use AI, 38% cited employee training as a way to adopt the technology, against 75% in G7 companies, and only 17% created a leadership role or a dedicated AI team, against around half in the G7. In the report chapter on G7 countries, the OECD notes that around 20% of companies fail to find qualified candidates when they try to hire and that many have trouble understanding which skills they need. A company that does not know what to ask of the market does not know what to teach in-house either.

Training does not solve everything the surveys call qualification, because building your own model requires specialized hiring. For most Brazilian companies, which buy ready-made solutions, the missing skill is using well what is already contracted: choosing the process, writing the instruction, checking the output and measuring the effect.

Summary of the surveys cited

Source	Reference year	Sample and method	Main finding
Cetic.br/NIC.br, TIC Empresas 2025	2025	4,174 companies with 10 or more employees, phone interviews from Feb 2025 to Jan 2026; AI indicator collected in companies with an IT area or specialist	17% used AI (15% of small, 50% of large); marketing or sales is the most cited purpose (44% of users)
IBGE, PINTEC Semestral 2024	2024	Sample of 1,731 industrial companies representing 10,167 companies with 100 or more employees; phone contact and online self-completion	41.9% used AI (16.9% in 2022); sales area in 75.2% of users; lack of qualified staff rose to 60.6% among those that adopted no advanced technology
OECD, The Adoption of Artificial Intelligence in Firms (São Paulo chapter, SEADE and Cetic.br)	2023	2,561 medium and large manufacturing and ICT companies in São Paulo, 167 AI users; phone and web	7% used AI; 38% of users cite training, against 75% in the G7
OECD, same report (G7 chapter)	2022 and 2023	840 AI-using companies in G7 countries, manufacturing and ICT, sample not nationally representative	Nearly three quarters train employees to adopt AI; around 20% fail to find qualified candidates
Eurostat, Use of artificial intelligence in enterprises	2025	157,000 of the EU’s 1.53 million companies with 10 or more employees, survey by national statistics offices	19.95% used AI (17% of small, 55.03% of large); lack of expertise is the main reason for not using it (70.89%)

What these data do not say

None of these surveys measures results: they show how many companies say they use AI and for what, without saying whether use increased revenue or cut costs. Large companies use more AI, and that does not allow the conclusion that AI made them grow: the reverse path is more likely, with size, budget and an IT team coming before adoption.

Definitions also change from one survey to another. TIC Empresas asks only companies with an IT area or specialist and reports over the total, which tends to underestimate use in small companies, where AI usually comes in through the owner or a salesperson, on a personal account. Comparing 7%, 17% and 41.9% as if they were the same measure leads to wrong conclusions, because each survey looks at a different size and sector.

All answers are self-reported, usually by one person who may not know what other areas use. My hypothesis, which the surveys do not test, is that the salesperson who pays for an assistant with a personal card stays out of the IT manager’s answer. The São Paulo figure, collected in the first half of 2023, portrays another moment of generative AI.

What leadership does with these numbers

1. Pick the right number for your company. If your company is a manufacturer with more than 100 employees, compare yourself with PINTEC; if it is in services or trade, use TIC Empresas for your size and sector.
2. Run your own survey before buying. Ask each area who uses AI, for which task, with which tool and on which account, including personal accounts.
3. Start with sales and marketing, as the market does, and move from individual use to process. Pick a commercial task with volume and an indicator, such as proposal follow-up, and define an owner, an approval rule and a baseline.
4. Treat the lack of prepared people as a budget line. If the company buys ready-made solutions, the missing skill is using and measuring. Set aside part of the AI budget for role-based training, next to the license.
5. Fix the data AI will read. Unavailable or poor-quality data appear as a reason for not using AI in 28% of large non-user companies. In sales, that data lives in the CRM, and the work is described in CRM data quality.
6. Answer legal concerns before they stall the project. Privacy and legal clarity appear in all three sources, and a policy with authorized tools and permitted data answers a good part of them. The LGPD points are in AI and LGPD.

What HR and L&D take from these data

For HR, the most useful data sits in the barriers. The lack of trained people is among the main reasons for not using AI in Brazil and leads in Europe, and PINTEC shows that reason growing in industry between 2022 and 2024. At the same time, São Paulo companies cite training half as often as the G7, according to the OECD. HR has the argument to put AI training in the annual plan, with role-based groups and process usage indicators, and the design of those groups is in AI training for companies.

The official numbers show Brazilian companies buying ready-made AI and using it first in sales and marketing, without anyone in-house who knows how to choose the use and measure the effect. That is the work of the in-company applied AI training, built on the company’s processes and data, with groups for executives, team leaders and sales, marketing and customer service teams. It delivers the prompt library by area and the agent architecture, with what each agent reads, what it runs on its own and who approves the rest.

Frequently asked questions

How many Brazilian companies use artificial intelligence?

According to Cetic.br’s TIC Empresas 2025, 17% of Brazilian companies with ten or more employees used AI in 2025, the equivalent of 93,475 companies. In industrial companies with 100 or more employees, IBGE’s PINTEC Semestral measured 41.9% in 2024.

Do small companies use AI?

Less: 15% of companies with 10 to 49 employees used AI in 2025, against 50% of large ones, according to Cetic.br. The survey only asks companies with an IT area or specialist, which tends to leave out informal use by the owner and the team.

What do Brazilian companies use AI for?

Marketing or sales is the most cited purpose, by 44% of companies that use AI, followed by organizing administrative processes (35%). In industrial companies with 100 or more employees, IBGE found AI in the sales area of 75.2% of users.

What is the main barrier to adopting AI in Brazil?

In TIC Empresas 2025, the lack of trained people and the lack of knowledge about suitable AI tie at the top, each cited by 14% of companies that do not use AI. Among large companies cost leads, at 42%, and in industry the lack of qualified staff grew from 46.4% to 60.6% between 2022 and 2024 among companies that adopted no advanced technology.

Is Brazil behind Europe in AI?

The average gap is small: 17% in Brazil and 19.95% in the European Union in 2025, with questions that are not identical. The difference between European countries is much larger, from 5.21% in Romania to 42.03% in Denmark.

How do you prepare the team to use AI?

Start with a commercial process that has an indicator, train people by the role they play in that process and measure use before and after. Each group’s content should cover how to ask, how to check the answer and who approves what AI produces, with the data and tools the company already has.