A survey of 200 CIOs, CTOs, and IT directors across the U.S., Canada, and Europe finds a quarter (25%) expect to reach full-scale orchestration of artificial intelligence (AI) this year, with AI providing the connective tissue between tools, teams and processes.
Conducted by Zapier, a provider of an IT automation platform, the survey finds another 43% anticipate reaching an agentic AI stage, where autonomous systems work across functions with minimal human input.
Collectively, the survey results indicate more than two thirds of IT leaders (68%) expect their organization to achieve some level of advanced adoption of AI in 2026. A full 84% said they are confident they will have solid proof of a return on investment in AI by 2026, with 54% prioritizing measurable productivity improvements, while 22% are planning to verify financial savings.
Additionally, 83% of respondents said their organization will demand that AI error rates stay at 5% or below for high-stakes operations. A total of 71% also identified “human-in-the-loop” approvals as their top governance priority for 2026.
The challenge now is achieving those goals in a way that responsibly scales AI by ensuring the proper orchestration and governance frameworks are in place, says Charles Crawford, senior product marketer at Zapier. In fact, 70% of respondents said they now view AI governance as a strategic differentiator, the survey finds.
Ultimately, organizations will need to carefully navigate how to insert probabilistic outputs generated by an AI model that are never the same twice into deterministic workflows that ultimately require a task to be completed the same way every time, he adds.
Too many employees who have adopted AI are simply copying and pasting incorrect AI outputs, otherwise known as AI workslop, into business processes without much careful examination, notes Crawford. Understanding the disparity that might exist between the desired outcome and the output generated by an AI model requires humans to understand that the AI model is, at the end of the day, making a guess based on the data that has been exposed to it, he adds. “It’s a nuanced distinction,” says Crawford.
Regardless of that distinction, there is no going back, with nearly three quarters (74%) noting AI budgets would be among the last to be cut during an economic downturn. More than two-thirds (69%) said their organization plans to invest $1 million or more in AI over the next year, with more than half planning to spend over $5 million.
At the same time, however, while 71% said AI will reshape teams, only 21% anticipate headcount reductions. For example, 65% of respondents in 2026 plan to hire AI automation specialists, followed by AI platform engineers (64%) to maintain AI infrastructure.
According to the survey, nearly half (46%) also plan to link promotions and pay directly to an employee’s ability to operate responsibly within AI-driven systems.
At this juncture, there is little doubt that AI will eventually transform almost every business workflow imaginable. The only thing left to determine now is not just to what degree but also the actual total cost. After all, if the cost of tokens consumed to drive an AI workflow everyday exceeds the value of the workflow, it won’t be too long before many enterprise IT organizations find their AI investments, much like a real estate investment gone wrong, are upside down.
