Gartner has confirmed what I have been watching unfold in real time across the projects we are running at Intradiegetic.
By 2028, more than half of all enterprises will have stopped paying for AI copilots. Simply put the market is moving past them toward agentic AI platforms that don’t assist humans with tasks, but execute workflows autonomously on behalf of the organisation.
Some see this as a software update, but it’s not. It is an entirely new operating OS. It is a structural shift in what intelligence means inside an enterprise.
A few months ago, I wrote about the restructuring projects we were running across the Middle East and Europe, communication departments rebuilt around agents, not around people. Headcount reductions between 30 and 70 percent. Not through gradual optimisation, but through the rapid displacement of traditional specialist roles by systems that now manage content calendars, stakeholder mapping, analytics, and campaign design without continuous human supervision.
The Gartner report puts numbers to what I have been describing on the ground. Only 5% of enterprise applications incorporated AI agents in 2025. That figure is projected to reach 40% by the end of 2026. And by 2028, at least 15% of day-to-day work decisions will be made autonomously, up from zero percent in 2024.
No surprise here, but what is striking is the warning buried inside the forecast.
Gartner expects more than 40% of agentic AI projects to be canceled by 2027 because organisations are deploying it without clear business value, without adequate risk governance, and without understanding the difference between genuine agentic architecture and what Gartner is now calling “agent washing”: the rebranding of existing chatbots and automation tools as if they were something fundamentally new.
Only around 130 vendors, out of thousands claiming agentic capability, are actually building the real thing.
This is the moment where haste becomes expensive. The pressure I see from leadership teams is to move fast and show AI on the balance sheet. And the temptation, as I wrote recently about sycophancy in AI systems, is to hear what you want to hear, to choose the comfortable answer over the honest one. To call a chatbot an agent and declare the transformation complete.
At Intradiegetic, the work we do is precisely in that gap: treating communication, decision-making, and information flow as an operating system, one that needs to be redesigned from the architecture up, not patched at the surface. Human judgment is not the bottleneck in that model. It is the thing worth protecting. And right now, it is the scarcest resource in the room.
The copilot era is ending faster than most organisations realise. What comes next will either be built on clear thinking about value, governance, and human oversight — or it will be built on hype, and it will be canceled before 2028.
Leave a Reply
You must be logged in to post a comment.