Sign one: the work is continuous and never finishes
A function is a candidate for a Digital FTE when the work is continuous rather than episodic. Content that should ship weekly, reconciliation that runs monthly, tickets that arrive daily, outbound that needs constant motion.
Continuous work suits function ownership because there is always a next thing to decide. Episodic work, however valuable, tends to be better served by a person with a sharp tool, since the coordination overhead that a Digital FTE removes barely exists.
Sign two: one person is the bottleneck
The clearest signal is a function whose throughput drops when a single person is unavailable. If content stops when the marketer is on holiday, or reconciliation slips when the controller is busy, the constraint is one person's attention rather than the difficulty of the work.
That is exactly the constraint function ownership moves. The work does not become easier, but it stops depending on one calendar.
Sign three: the evidence exists and lives somewhere
A function is ready when its evidence lives in a system of record that can be read: a ledger, a search console, an ad platform, a help desk, a CRM. The Digital FTE needs somewhere to look for truth and something to be measured against.
Work that is judged entirely on taste, with no external evidence, is much harder to delegate this way. It can still be assisted, but the ownership argument weakens considerably.
- Continuous work that never finishes
- One person whose absence stops the function
- Evidence that lives in a readable system of record
- A backlog that never clears despite everyone being busy
- Work that queues overnight or over weekends
- Quality that varies with who happens to do the task
- A function everyone agrees matters but nobody owns fully
Counter-sign one: the process is undefined
A function is not ready when the business has never agreed how the work should be done. A Digital FTE inherits the clarity of its remit; it does not resolve ambiguity, it acts confidently inside it.
The tell is that two people in the same team would produce visibly different work and both would defend it as correct. Fix that first. The definition work has value on its own, and it is the prerequisite rather than the obstacle.
Counter-sign two: the value is mostly relational
A function is a poor fit when most of its value comes from relationships and reading people. Key account management, negotiation, internal politics, and anything where the point is that a specific person showed up.
Parts of those functions still benefit — the preparation, the research, the follow-up, the record-keeping. The relationship itself does not delegate, and a vendor suggesting otherwise is overselling.
Counter-sign three: nobody has time to review
A function is not ready when no one has capacity to review the output, and this is the most commonly ignored counter-sign. A Digital FTE produces proposals; proposals need decisions; decisions need a person with time and authority.
If the honest answer is that the one qualified reviewer is already underwater, solve that first. Deploying into a function with no review capacity produces an approvals queue that nobody clears, which is worse than the original backlog because it also looks like progress.
What is the quickest honest test?
The quickest honest test is two weeks of time recording by the people doing the function, sorted into two columns: work with a checkable right answer that could be done from evidence, and work that needs judgement, a relationship, or a decision about a person. If the first column holds most of the hours and the evidence for it lives in systems a Digital FTE can connect to, the signs are present. If the second column dominates, the function needs people, and an AI employee would spend its time escalating.
The two weeks also produce the baseline that any later pilot needs: hours by task, outcomes, and the quality measures you have. That is why the test is worth doing even when the answer turns out to be no; the record is useful either way. The pilot guide on ftes.ai starts from exactly this baseline, and the signs described above are the reasons a function's first column tends to be large: continuous work, a single bottleneck, and evidence that already exists somewhere.