Gartner’s AI Chief Blasts Current AI Tools: “AI Is Not Doing Its Job Today”
Erick Brethenoux, Gartner’s global chief of AI research, slammed today’s AI at the firm’s Data & Analytics Summit in Sydney. He called out AI-generated meeting summaries as useless and questioned why AI just doesn’t do the work instead of listing tasks.
Brethenoux said: 
“AI is not doing its job today and should leave us alone”
“I don’t have time to do the five actions in the summary,”
“I know what I have to do.”
He wants AI to actually perform tedious tasks automatically, not just summarize or list action items.
He pointed to US healthcare company Vizient, where the CTO asked employees what tasks annoyed them most. The company automated those chores, and got instant adoption with no pushback. Employees then suggested more ideas for automating work.
Brethenoux calls this approach "Empathy AI." Another example is a real estate firm automating 17 tenant assessment steps in parallel instead of sequentially, saving huge time and effort.
He’s skeptical of the hype around personal AI agents that are supposed to juggle calendars and decisions independently across an enterprise.
Brethenoux challenged vendors on this:
“Now you have 50,000 agents running around the enterprise,”
“How do you orchestrate this? How do they negotiate?”
He said vendors are silent on how agents would handle competing priorities—say between an employee’s boss and family.
He says the real challenge is a software engineering problem: what agents can see, do, and control, and how they communicate autonomously.
Brethenoux blasted the sloppy use of AI jargon, blaming hype on mixing "AI agents" and "generative AI" under fuzzy definitions.
He shared a quote from Albert Camus:
“To misname things is to contribute to the world’s miseries.”
The AI “agentic nirvana” promised by vendors, he warns, is not close. The tech still can’t handle complex tasks without serious systems engineering behind it.
