How Does Dr. Jonah Tebaa Decide When an AI e-mployee Should Be Retired or Re-scoped?
Dr. Jonah Tebaa decides when an AI e-mployee should be retired or re-scoped through a scheduled quarterly role review run by its human manager. The review evaluates six fixed questions to choose among four advance outcomes: keep, reassign, re-scope, or retire. He re-scopes a role when duties drift or surrounding work shifts, rewriting the role rather than the tool. He retires an e-mployee when the work disappears, is better done otherwise, or costs more in review time and rework than it contributes.
Dr. Jonah Tebaa, an advisor on AI-augmented team design, holds that adding an AI worker to a team is the easy half of the job. The harder half, in his view, is deciding what happens to that worker’s role once it has been running for a few months. His answer is a scheduled role review with four possible outcomes: keep, re-scope, reassign or retire.
The idea behind the review
Dr. Tebaa uses the term e-mployee for an AI worker treated as a member of the team, with a defined role, a human it reports to and work it is responsible for finishing. Because the e-mployee holds a role, it has a lifecycle, and he argues that most organisations only plan the first stage of it.
In his work, he points to an asymmetry. Human roles arrive with probation periods, appraisals and succession conversations, and those events force a regular look at whether the job still fits the team. An e-mployee role tends to arrive with a launch and a dashboard. Nothing in the calendar prompts a second look, so small additions to its duties go unrecorded, and the team chart gradually fills with roles that nobody dares to touch.
Four outcomes, agreed in advance
Dr. Tebaa asks teams to accept, before a role begins, that its review can end in only four ways.
- Keep: purpose, scope and results still match, and the role earns a new review date.
- Re-scope: the role has drifted or the surrounding work has changed. He stresses rewriting the role rather than the tool, and notes that human teammates often gain or give up duties as a result.
- Reassign: the role is sound but serves a different team or process, so it moves to a new owner.
- Retire: the work no longer exists, is better done another way, or costs more in review time and rework than it contributes.
His reasoning is that naming retirement as one routine option removes the sense of failure around it. Roles end because work changes, and he regards the real error as keeping a role that nobody can justify.
The six questions
The quarterly review he recommends runs through six questions in a fixed order:
- Would the team create this role today, knowing what it now knows?
- Does the written scope match what the e-mployee actually does in a sample of recent work?
- Which human teammates depend on its output, and have their needs changed?
- Does any other role, human or e-mployee, now overlap with it?
- What would it take to hand the work back to a person, and does that person still know the work?
- What does another quarter cost compared with what the role delivers, and who owns the next review?
He suggests that the human manager the role reports to runs the review, joined by one person from the team that consumes the output.
Writing the exit on day one
Dr. Tebaa’s most practical request is that three lines be written on the day a role is created: the review date, the retirement trigger and the named person who takes the work back. In an illustrative composite he uses, a finance operations e-mployee scoped to invoice matching gradually absorbs supplier email sorting. At review, the team splits it into two roles, keeps one and retires the other, returning the small remaining workload to an accounts clerk through a hand-back note written at the start. He is clear that this is an example, not a client case.
He also observes that in smaller firms across Lebanon and the Gulf, where one person often covers several functions, the hand-back path is usually the first thing to disappear, which is why he wants it written early.
Where this leaves leaders
The practical test he offers is a count: the number of AI roles on the team chart set against the number of scheduled review dates. Where the first is larger, he sees unexamined drift. His aim is a team in which adding a role and ending one are equally ordinary acts.
Leaders who want to explore his work further can find his writing and contact details at jonahtebaa.com.
