Fifty-nine percent of Brazilian companies already use AI in customer service. Fifty-four percent use it in marketing and sales. In business strategy and finance, adoption drops to 38%. In HR and sustainability, to 29%. The pattern is not subtle: AI moves fast where the risk is low and the decision is operational — and barely shows up where the decision is strategic. The budget tells the same story: 77% of companies invest up to 2% of their technology budget in AI, and only 9% go above 5%.
The numbers come from Panorama 2026, a survey by Amcham Brazil with Humanizadas, which interviewed 629 executives — most of them senior leadership — at medium and large companies employing more than 592,000 people combined. The most uncomfortable finding isn’t low adoption: it’s that 61% of leaders say they haven’t seen meaningful impact yet, and only 3% have turned AI into new revenue or a real competitive advantage. A company that only deploys AI where the risk is low shouldn’t expect a result that only shows up where the decision is consequential.
Customer service gets AI first because nobody is risking anything
Putting an AI assistant in front of customer questions is a low-risk bet: the worst case is a bad answer, correctable on the spot, with no structural consequence. That’s why customer service and marketing lead adoption — not because they’re where the highest potential return sits, but because they’re where experimenting doesn’t threaten anything important if it fails. A pilot there doesn’t need board sign-off, doesn’t change anyone’s decision process, and can be switched off with zero political cost if it doesn’t work.
That’s tactical use: it solves a visible, point-in-time problem that’s easy to justify in a meeting. It isn’t the same as strategic use — the kind that changes how a company decides what to prioritize, where to allocate capital, or how to weigh risk. And that’s exactly where adoption falls off a cliff.
Strategy and HR are left out because that’s where the real decision lives
Bringing AI into finance, strategy, or HR means touching how a company decides who gets promoted, which investment gets approved, which risk gets accepted. That requires someone with political weight to own the outcome — and unlike a service chatbot, a mistake here doesn’t get fixed with an apology. That’s why adoption in those areas sits at 38% and 29%: it isn’t a shortage of use cases, it’s that the right use case demands a leadership decision most companies haven’t made yet.
That’s the gap the budget reveals without needing any culture survey: whoever only funds the cheap, tactical pilot is, in practice, choosing not to touch the strategic decision — even while listing AI as the number-one priority on this year’s planning slide.
The budget is the lie detector for priority
A company can put “AI” at the top of its strategic priorities slide and still not be prioritizing AI for real — the test isn’t the slide, it’s the budget line. With up to 2% of the technology budget, the money covers a tool license and maybe one isolated pilot; it doesn’t cover legacy-system integration, process redesign, training the people who’ll actually use it, and the follow-up once the pilot leaves the slide deck — exactly the costs that decide whether a strategic use case survives its first committee objection or dies there. Only the 9% that go above 5% have the budget runway for that, and it’s no coincidence they’re also closer to the small group of 3% that already sees new revenue or competitive advantage.
Three questions before asking for more AI budget
As João Paulo Batistella argues — an innovation executive with a background at Ericsson, Telefônica/Vivo, and five years as CEO of EISA — budget is a leadership decision, not an automatic consequence of priority talk: whoever signs off on the number is, in practice, deciding whether AI will touch a strategic process or stay confined to customer service. Before asking for more budget, three questions separate whoever will actually use the increase from whoever will just buy more tactical tooling:
What share of the AI budget goes toward a process that today is decided by a person using judgment — not toward automating a repetitive response? If the answer is zero, the budget is funding comfort, not strategy.
Who signs off on that number, and how often do they revisit the decision? An AI budget set once a year by someone who doesn’t track the result month to month tends to keep protecting the status quo.
If the budget doubled tomorrow, is there a ready strategic-use plan to receive that money — or would the company just buy more licenses of the same tactical tool? Money without a specific plan tends to flow back to the lowest-risk path, which is already saturated.
The survey also points to why even companies that want to go beyond tactical use get stuck: 64% cite lack of technical training on the team, 52% cite the absence of a defined usage strategy, 43% cite poor internal data quality. None of those three barriers gets solved by buying another license — which is why the budget, more than the priority speech, remains the most honest indicator of where AI will actually operate in 2026.
Follow Eleva Technologies for more analysis on technology applied to business: follow @ElevaTechno on X, @elevatechnologies on Instagram or Eleva Technologies on LinkedIn, or learn about the group at elevatec.net/about.
