What Gartner said
In June 2025, Gartner predicted that over 40% of agentic AI projects will be cancelled by the end of 2027, due to escalating costs, unclear business value or inadequate risk controls. This is a forecast from an analyst firm, not a measured result.
Gartner also described "agent washing": vendors rebranding existing products, such as chatbots, robotic process automation and AI assistants, as agents without real agentic capabilities. It estimated that only about 130 of the thousands of agentic AI vendors are real.
Why it is relevant
An AI agent is software that can take actions towards a goal, not just answer questions. Because it acts, mistakes cost more than a wrong answer on a screen. Each of Gartner's three causes can be checked before you buy.
We build an AI agent ourselves, called Boafo, so read the rest with that in mind.
Questions to ask a vendor
- What can it do without a person? Ask for the list, and for what stops it. Sensitive actions should go to someone for approval.
- Show me an action, not a conversation. If it can only reply, it may be a chatbot with a new name.
- What does it cost as usage grows? Costs that rise with every action are hard to predict from a pilot.
- What measurable result will we see, and when? Agree the number before you start.
- What can we see afterwards? You should be able to review what it did and why.