Weeks 1: access, remit, and baseline
Week one covers access grants, the remit, the targets, and the recorded baseline. It is short when the brief was written before purchase and long when it was not, which is the main variable in the whole timeline.
The baseline is the piece that cannot be recovered later. Record output volume, cycle time, and current human hours before the first run, because after it starts nobody remembers the previous state accurately.
Weeks 2 to 3: close review and send-backs
Weeks two and three are close review. Expect to reject a meaningful share of proposals and to spend real time writing specific reasons for each rejection.
This phase feels like the deployment is going badly and is usually going well. Send-backs with specific reasons are how the standard transfers; a period with no rejections at this stage more often means nobody is reading than that the work is perfect.
Weeks 4 to 6: remit tuning
Weeks four to six are for tuning the remit from the pattern in your send-backs rather than from impressions. Group the rejections and find the repeats: the missing element, the misjudged tone, the wrong priority.
This is the least visible and most valuable phase. Nothing dramatic happens, output looks similar week to week, and the underlying rejection rate falls because the instructions finally match the intent.
Weeks 6 to 9: first promotions
Weeks six to nine are when the first autonomy promotions land. Pick the class of action approved most often without amendment, promote only that class inside explicit limits, and keep every other gate.
Expect two or three narrow promotions across this period rather than one broad one. Each should be small enough that reversing it would be uneventful, because the ability to reverse is what makes the grant reasonable.
- Week 1: access, remit, targets, recorded baseline
- Weeks 2-3: close review, specific send-backs
- Weeks 4-6: remit tuning from send-back patterns
- Weeks 6-9: first narrow autonomy promotions
- Weeks 10-12: measure against baseline, consider a second role
Weeks 10 to 12: measurement and the second role
Weeks ten to twelve are for measuring against the recorded baseline and deciding about a second role. Compare output volume, cycle time, and human hours, and be honest about the review time the arrangement consumes.
If the numbers support expansion, write the one-page operating model before adding anything: owner, remit, targets, approval routing, access scope, autonomy level, review rhythm. That page is what makes the second role a two-week deployment rather than a repeat of this twelve-week one.
What makes a timeline slip?
Timelines slip for three reasons, and none of them are technical. An undefined process, meaning the remit cannot be written sharply. An unavailable reviewer, meaning the queue backs up. And scope creep, meaning a second function is added before the first is measured.
All three are management problems with management fixes. The technical work in a Digital FTE deployment is genuinely short; the organisational work is where the calendar goes.
What does the timeline cost in human hours, week by week?
As a labelled outline with your own figures to fill in: week one costs the manager a few hours for connections with scoped credentials, the remit and escalation list, the Brand Brain review, and the baseline record, plus an hour of security's time on the connections. Weeks two and three cost the most approval time, perhaps fifteen minutes a day, because every class of action is gated and the manager is reading proposals closely and sending back with reasons. Weeks four to six cost half an hour a week of remit tuning from the send-back patterns and less approval time as the remit settles.
Weeks six to nine cost the monthly review, about thirty minutes, at which the first class of action is promoted from a clean record, and approval time falls further. Weeks ten to twelve cost the measurement comparison against the baseline and, if a second role is added, a shorter version of week one, because the platform pre-fills what it already knows. Recording these hours as they happen is what makes the twelve-week comparison honest: the human time is part of the cost, and its downward trend is part of the result.