In 2025, agentic artificial intelligence (AI) was more a shiny big idea than a cold hard reality. Companies talked breathlessly everywhere about the potential of the rapidly evolving technology. But nearly all were skeptical or cautious, content to absorb news and ideas about the latest iteration of AI through a flood of research reports, thought leadership pieces and posts on LinkedIn and Substack.
Like a river with still pockets, flowing eddies and rushing currents, the still phase is giving way. Recent PYMNTS survey data shows a sharp shift from sideline interest to real-world action — and it’s happening fast. On Wednesday, enterprise planning platform Board said it was launching tools and services for finance, supply chain and merchandising planning that were developed with Microsoft Foundry using Microsoft Azure’s agentic AI.
For the most recent edition of the CAIO Report, PYMNTS Intelligence surveyed 60 chief product officers at U.S. businesses bringing in at least $1 billion a year. The study, completed last November, explored how they’re thinking about and using agentic AI.
Last August, more than half of companies (52%) said they were just “considering” or “exploring” agentic AI. By November, that number had plunged to 30%. In other words, a big chunk of the enterprise market moved out of the window-shopping phase.
What replaced the passive interest is hands-on implementation. In November, nearly 1 in 4 CPOs reported that they were either piloting agentic AI or fully using it in production processes, up from just 3% in August. Actual usage and piloting were evenly split: Just over 1 in 8, or 12%, are testing agents out, and another 12% have already made them part of their daily operations.
The upshot is that in just three months, the number of enterprises using or testing agentic AI spiked sevenfold.
But while some companies are sprinting ahead, others remain harder to sway. About half of companies seem totally uninterested in agentic AI, which can autonomously make real-world decisions and perform tasks with minimal human oversight.
In November, 45% of CPOs said they didn’t have any plans to adopt the technology any time soon. That portion might be a steep drop-off from June, when 90% said the same, but it’s entirely unchanged from August. As with polarizing foods like broccoli, there’s not a lot of middle ground: Companies are either racing toward autonomous systems or sitting agentic AI out — for now — altogether.
It’s not just one or two industries embracing the technology. Adoption is rising across every sector. Goods and manufacturing enterprises saw one of the biggest jumps in implementation, going from almost no usage last August to nearly 1 in 5 firms using or piloting the technology by November. Services companies followed a similar path, with adoption jumping fivefold, from 4.3% in August to 25% three months later. Technology companies lead the pack, with 31% piloting or using AI agents.
One the one hand, the agentic AI digital divide between industries is shrinking quickly. On the other hand, the divide between enthusiasts and skeptics remains entrenched.
Some companies are showing off their agentic AI ventures.
Earlier this month, grocery giant Kroger said it would introduce Google’s Gemini agent tools to allow shoppers to convert a “I want to prepare vegan tomato soup” request into a shopping cart of ingredients through a single click. Walmart announced an agentic commerce partnership with Google that will take consumers from online shopping to purchase to delivery.
Rideshare and delivery firm Uber is using agentic tools it developed internally to improve employees’ workplace communications in Slack. The tool has reduced what it called “incorrect advice” in Slack messages by 60%.
BMW is using a multi-agent system to boost efficiency across its manufacturing plants, where specialized autonomous actors collaborate. Paper goods manufacturing giant Georgia-Pacific is doing the same to modernize operations across its plants.
In financial services, JPMorgan Chase has started deploying agents internally to automate labor-intensive tasks, with one agent able to produce an investment banking deck in 30 seconds.
“The broad vision that we’re working towards is one where the JPMorgan Chase of the future is going to be a fully AI-connected enterprise,” Chief Data Analytics Officer Derek Waldron told CNBC during a demonstration.
Capital One is using AI agents as “digital concierges” for car buyers and dealerships that allows them to compare vehicles and schedule appointments with sales representatives. No allowing an agentic bot to buy a car for you just yet.