As companies put AI agents to work, managers are facing a new question: should overseeing software count like managing people when pay and seniority are on the line? An article carried by The Irish Times describes the debate, and the workload that comes with supervising agents that never clock off.
Watch Desk analysis
What happened
The article says AI agents, which can independently handle tasks such as checking invoices, browsing the web and sending emails, are used in 40 per cent of companies with revenue of $1 billion or more, up from 27 per cent last year, citing McKinsey. It says growth at smaller companies is flat, though still 22 per cent.
Kit Krugman, chief people officer at AI supply-chain firm Altana, says organisations will have to rethink compensation. If employees are overseeing agent teams, she argues, that work could count as management. The article also quotes former Google and Slack executive Brian Elliott saying he limits himself to two or three agents at once because supervising more becomes a substantial cognitive load.
Why it matters
The familiar measure of a manager’s importance is often the number of people reporting to them. Agents complicate that arithmetic: they may not need the same kind of supervision as people, but keeping track of fast-moving systems, silent mistakes and costly token use is still work.
The article describes a live workplace debate, not a settled pay policy. It also points to a tension companies cannot spreadsheet away: automating tasks may save money, while making human oversight more demanding.
Our read
Counting bots as direct reports would be a neat fix to an untidy problem, which is usually a sign to check the small print. The better question is what responsibility the manager actually carries: how many agents they supervise, what decisions those agents make, and who answers when they go quietly wrong. Headcount alone looks like a poor proxy for that job.
What to watch
- Whether employers change management pay or performance measures as agent use grows.
- How firms define responsibility for agents that make mistakes without obvious warning.
- Whether the cost of supervising agents, including token bills, is counted alongside their promised savings.
Discussion spark: If someone supervises dozens of AI agents, should that count towards their pay and seniority like managing a human team, or should compensation depend on responsibility and results instead?
Sources and evidence
- Source update (9 November 2026, 06:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.