Disruptive technologies are sometimes mistakenly written off as misfires when they first emerge. Instant-film icon Polaroid ignored the advent of digital cameras, leading to its bankruptcy in 2001 and later reinvention as a successful digital-first company. But an opposite dynamic can also happen. New technologies can grow like wildfire, only to go up in flames. At its peak in 2010, Blackberry commanded more than half of the U.S. market for smartphones and 20% of the global market.
One year ago, according to Google Trends, you could count the amount of time agentic AI was searched on one hand. Go to any day in June 2025 and you’ll see a completely different story. It has topped 100 on the Google index as some executives in the business pump it up as the next big thing in AI and some executive call for pumping the brakes until infrastructure and workflow issues are sorted.
The burning question now concerns the newest incarnation of artificial intelligence (AI). How will the journey of agentic AI, which can autonomously complete tasks and take actions independent of human involvement, unfold? The question has grown more acute amid rising skepticism over whether the technology can live up to its promise, produce valid outcomes and be used ethically.
A forthcoming PYMNTS Intelligence report reveals that while almost all (CFOs) at enterprise-level companies are familiar with agentic AI, just 15% are even contemplating putting it to work. Companies are in the early stages of evaluation, trials and proof of concept, not wholescale adoption. Agentic AI may be everywhere as a topic, but it’s nowhere near part of the business firmament.
An Idea, Not a Project
What’s clear in the data is that the companies exploring agentic AI have already gone deep with their use of generative AI, a less-but-still-advanced technology like ChatGPT that is used to create content, including reports, field customer service queries, code software and analyze data.
In other words, a firm’s exploration of agentic AI closely tracks how far it has already advanced with gen AI. Among firms highly automated firms through gen AI, two-thirds are weighing agentic AI adoption, viewing it as a “logical next step.” Meanwhile, at companies still relying heavily on manual workflows, only 11% are considering agentic AI.
Paradoxically, even as gen AI’s use cases surge — with its application in product innovation up 21% and fraud detection up 31% since March 2024, the PYMNTS report shows — overall investment in Gen AI is slowing. Only 27% of enterprises are boosting their gen AI budgets this year, a sharp drop from 53% last year. This contraction likely spills over to agentic AI, which demands even greater confidence and trust before companies commit time, budgets, and leadership.
Read the forthcoming report to learn how agentic AI is a popular but controversial without a clear timeline for implementation.
Read more:
Agentic AI Systems Can Misbehave if Cornered, Anthropic Says
MIT Looks at How AI Agents Can Learn to Reason Like Humans
Gen AI’s Evolving Role in the Enterprise Reset
The post For 85% of Enterprise CFOs, Agentic AI Remains a Question Mark appeared first on PYMNTS.com.