Bosses say AI agents are live while most firms still experiment
Reviewed by Helen Jones
A Teradata survey of 1,000 global senior technology and data leaders, published around 29 September 2026, found 69% of C-suite executives say their organisation is already operating with agentic AI, compared with 57% of VPs. Agentic AI here means software that can take multi-step actions with less constant prompting, not a single chat reply. On Teradata’s Agentic AI Maturity Index, 68% of organisations sit in Experimenting and Developing. Twenty-five percent have reached Building, and 7% have reached Operationalizing. Sixty-two percent of leaders say they judge AI on enterprise-wide return, not individual speed. Only 30% use margin improvement as a success measure, and 63% report no more than an emerging or minimal positive return so far. The stages and returns are self-reported in vendor research.
Until now, many leadership conversations treated ‘we have agents’ as proof the organisation had moved past one-off prompting. That claim sat next to real work that still lived in blank chats, personal tricks, and heavy checking. The gap between C-suite and VP answers already hinted that the people closest to delivery were less sure anything was actually running. The survey does not prove agents failed. It shows most firms have not left the experiment stage, even while they talk about enterprise return. Measurement still rarely touches margin. For anyone trying to turn judgement into a repeatable AI workflow, the useful signal is this: live agents are being counted before context, review steps, and team handoffs exist.
Analysis
Treat executive ‘agents are live’ talk as a trap, not a rollout plan, until a task has a named owner, a context pack, and a human review step. Pick one recurring piece of work this week and write it as a one-page workflow blueprint: inputs, privacy boundary, quality check, and who signs off before anyone else runs it.
Source note
Pulse published by Collab365 Spaces, reviewed by Helen Jones on . Cite as "Bosses say AI agents are live while most firms still experiment", Collab365 Spaces. 1 source referenced.