What the research found
Kyndryl, an IT infrastructure services company, commissioned Edelman Intelligence to survey 3,700 business and technology leaders in 21 countries, 200 of them in Brazil, between July and September 2026. AI has left the lab: only 49% still describe their organization as experimenting, 13 points fewer than in 2025, and 45% say innovation often stalls after proof of concept (down from 57%). The road is not clear, though. 24% had an AI investment fail and 21% canceled a major initiative.
The bottleneck sits in the foundation. Although 71% consider their IT estate modernized, two in three CIOs say it must be rebuilt to support AI at scale. Four in ten admit their data is not ready, and 46% say data silos limit how far projects can grow. For 38%, their AI strategy emerged by accident rather than by design.
The most useful finding is about people. Among companies that redesigned roles around AI and run formal change management, 69% report positive returns, versus 53% of the rest. Even so, only 13% plan to rebuild processes and operating models around AI. Most prefer to embed it where returns are clearest (40%) or use it as a productivity layer on current work (35%).
In the Brazilian breakdown reported by Forbes Brasil, 91.5% say AI is already in operations beyond pilots and 59% say their technology foundation needs significant re-architecture. These figures do not appear in the public report.
What other research shows
Deloitte reached a similar picture by a different route. In its State of AI in the Enterprise survey, released on October 29, 2026 and based on 3,500 leaders in 25 countries, 81% say AI already speeds up decisions and execution. Only 25%, however, use it to deeply transform the business, and 28% report revenue gains. There is a gap similar to Kyndryl's, too: 60% have redesigned jobs around AI at least moderately, but only 15% have actually changed who holds decision-making authority. And nearly one in four companies say it is unclear who is accountable when AI produces a bad outcome.
In Brazil, adoption is growing faster than the conditions to sustain it. IBGE's Pintec Semestral survey showed that the share of industrial companies with 100 or more employees using AI rose from 16.9% in 2022 to 41.9% in 2024. Among those already using advanced digital technologies, 78.6% cite high costs as an obstacle and 54.2% cite a lack of qualified staff. In the World Digital and AI Ranking 2026, by IMD with Fundação Dom Cabral, the country ranked 54th of 70 economies, two places lower than the previous edition. In the sub-indicators, it is 65th in AI adoption and 61st in digital and AI governance.
Our take
The returns in Kyndryl's survey are self-reported, and the company sells infrastructure modernization. Read it with that in mind. But when sources of different kinds, executive surveys and official statistics, point the same way, the message is hard to ignore. The pilot works. Scaling runs into scattered data, legacy systems, unprepared people and no clear process owner.
What stood out to us was the 16-point gap between companies that changed roles and change management and those that did not. Deloitte shows the same signal another way: many companies redesigned jobs, few changed who decides. That is not fixed by buying another tool. It is process, role and governance work, and it is usually the part that gets postponed.
The drop in risk readiness also concerns us. More AI in production means more things that can fail, and contingency plans have not kept up. For Brazilian companies, which start from a weak position on AI governance in the international ranking, that weighs more.
Questions for your next meeting
- Which AI initiatives are in production today, and how do we measure the return of each one, with numbers rather than impressions?
- Is our data ready for the next stage, or does each department keep its own version of the truth?
- When we put AI into a process, did anyone redesign roles and define who decides, or did we just switch the tool on?
- If an AI system fails or gets something wrong tomorrow, who notices, who is accountable and what is plan B?
Sources
- Readiness Reframed: 2026-27 Readiness Report, Kyndryl (Edelman Intelligence survey of 3,700 leaders in 21 countries), November 10, 2026.
- Enterprise AI Enters an ROI Reckoning: More Use, Uneven Value, Deloitte, October 29, 2026.
- De 2022 a 2024, percentual de empresas industriais utilizando inteligência artificial subiu de 16,9% para 41,9% (From 2022 to 2024, the share of industrial companies using artificial intelligence rose from 16.9% to 41.9%), IBGE, Pintec Semestral 2024, September 24, 2025.
- Brasil no World Digital and AI Ranking 2026 (IMD e Fundação Dom Cabral) (Brazil in the World Digital and AI Ranking 2026), Forbes Brasil, November 11, 2026.
- 91,5% das empresas já levam IA à operação, mas 59% precisam rever a TI (91.5% of companies already bring AI into operations, but 59% need to rethink IT), Forbes Brasil, November 13, 2026.
The figures come from the sources cited. The selection and analysis are by Mapa Consultoria Empresarial.
Want to discuss how this applies to your company?
