BLOG
Comup Perspectives
Data and judgment for IT and customer-experience leaders: what AI is already achieving in real service operations.
Data and judgment for IT and customer-experience leaders: what AI is already achieving in real service operations.
Gartner said it plainly: over 40% of agentic AI projects will be cancelled by the end of 2027. That is not a forecast against the technology — the agents work. It is a forecast against the way most organizations are choosing where to deploy them.
of agentic AI projects will be cancelled by 2027, according to Gartner: escalating costs, unclear business value and inadequate risk controls.
The reasons Gartner cites are not technical. They are managerial: projects born out of fashion, pilots that never defined which number they were supposed to move, and budgets that grow with no baseline to compare against. An agent with no measurable value to produce is an expense with a good slide deck.
Service operations show it clearly: the agent gets installed on top of a broken process, a poor knowledge base and categories nobody respects — and the result is automating the disorder, only faster and with more confidence.
The projects that make it did not start with the tool. They started with a pain that had a number: a first-response time, a first-contact resolution rate, a cost per ticket. And they committed to one simple rule: at day 90 there is a number — or the course is corrected, or the project is cancelled with dignity before the budget cancels it.
Agentic AI will not replace IT teams. It will replace projects with no measurable purpose. The question for your organization is not whether to adopt agents — it is whether your first use case can prove its value in 90 days. Let's talk about how to choose it.