Axios C-Suite: No one knows who should manage AI agents
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Brendan Lynch/Axios. Stock: Getty Images
AI agents are increasingly joining our teams with most of what a new hire gets: a boss, a specific job to do, often a name others know and a record of the work they're doing.
Why it matters: Every CEO faces a very tough dilemma: How do we manage agents? And who's responsible when they screw up?
The big picture: We can't hand vital work to imperfect agents without clear rules for who's in charge of them. This creates a new gray area between HR and tech. Someone needs to be accountable.
Your company might not be here yet, but it soon will be. And lots of companies are pumping out new products to speed this up.
- Atlassian, maker of Jira and Confluence, began rolling out agents that work next to your people, each with its own profile and a work record everyone can see. CEO Mike Cannon-Brookes wrote, "Every agent has a clear owner, a distinct profile, and real-time attribution."
- Cisco unveiled plans to let teams invite agents into Webex meetings and calls.
- Workday, the HR software giant, agreed this week to buy C1.ai so it can govern what both employees and agents can access, from the day they join to the day they leave.
Yes, but: The two people who should own this — your head of HR and your head of IT — probably aren't on the same page.
- Only 38% of CHROs and CIOs share an understanding of how AI affects the future of work, per Gartner.
- "The greatest risk for CHROs is being confined to supporting a transformation they should co-lead," according to Gartner senior director analyst Rachel Juley.
- Think about it: Is an agent pure tech, or a de facto worker doing human tasks next to humans? Truth is, it's a hybrid that demands new thinking.
What we can do as CEOs:
- Name one boss for agents. One executive owns agent governance, with HR and IT both at the table. Job one is identifying every agent you have. 82% of enterprises found AI agents they didn't know about in the past year.
- Make every agent's owner answerable for its work. Give owners full visibility into what their agents do and responsibility for making sure it's doing it correctly and safely.
- Match oversight to risk. A human signs off on anything you can't undo: money, customers, people decisions.
- Know who pulls the plug, and test it. Set a clear standard for each task. Give someone the power and accountability to pull back when agents fall short.
Get ahead of this: Most middle managers struggle with managing people. Imagine them managing people and AI. This will require next-level management training.
📈 If you're a CEO or on a CEO's team: Ask to join Jim's new weekly Axios C-Suite newsletter.
