Executive Communication, Handover and Saying No
Report progress executives read, hand over a system the customer can run, and decline work that should not be built.
Executive Communication, Handover and Saying No
The last mile of an engagement is communication, and it decides whether good engineering turns into a funded, adopted system.
Three jobs make up that last mile:
Why it matters
Technically successful projects still fail when sponsors lose confidence, nobody owns the system after handover, or the team builds something that should have been stopped. These are communication failures, and they are preventable.
Executives read little and decide a lot. An update that buries the decision in a list of completed tasks wastes their attention, and the decision gets made without your evidence.
Handover is where many AI pilots quietly die: the system works on the day you leave, then the model changes, the data drifts, and nobody knows how to tell. And the FDE who recommends stopping a weak use case earns more trust than one who ships it.
How to do it
Send a weekly one-page update: metrics against target, what changed, risks with owners, and decisions needed (template below).
Build demos around the decision the sponsor must make, using real cases from the evaluation set, including a failure.
Start handover in the first third of the engagement, not the last week: runbooks, monitoring, the evaluation set, named owners and a support path.
Run the system with the customer's team operating it before you leave.
When recommending no, show the evidence, offer an alternative, and state what would change the answer.
Close with a one-page memo: go, change or stop, against the original criteria.
A weekly update an executive reads
# Claims triage assistant: week 4 of 6
**Status: on track for the 4 December decision.**
| Metric | Baseline | Target | This week |
|---|---|---|---|
| Triage time per claim | 9.0 min | 6.0 min | 6.4 min (down from 7.1) |
| Category accuracy | n/a | 92% | 93% |
| Fraud-flag accuracy | n/a | 100% | 97%: 1 miss, cause found, fix in test |
**Changed this week:** added the policy-number lookup; handlers now see the claim history beside the draft.
**Risks:** the claims system export for October is late (owner: Customer IT, due Friday).
**Decision needed:** approve 3 hours of senior-handler time next week to review the fraud-flag fix.Worked example: "it works, but nobody uses it"
The evaluation meets the target, but only 4 of 14 handlers use the assistant regularly.
A weak update says: "Accuracy 93%, ahead of target. Adoption is lower than expected."
A strong update says: "The assistant meets its accuracy target, but most handlers are not using it, so we are not yet saving time. Interviews show they do not trust the fraud flags after one early miss. Recommendation: two weeks showing handlers the evidence behind each flag, then re-measure. Decision needed: approve moving the end date by two weeks."
The second version tells the sponsor what is true, why, and what to decide.
In a real engagement
Never surprise a sponsor with bad news in a meeting with their boss; tell them first, privately, with a plan.
Write the post-POC recommendation as a one-page memo with a clear go, change or stop. Attach the evidence, but lead with the decision.
At handover, make the evaluation set, the runbooks and the monitoring the customer's property, with a named owner for each. If no one will own the system, say so in the memo: it is a reason not to go to production yet.
Saying no is easier with an alternative: "not full auto-approval, but triage and drafting, which the data supports today".
Challenge
Your evaluation shows the system meets the target, but the frontline team does not trust it and usage is low. Write the three-line executive summary for this week's update.
Checklist
Production checklist
0 of 8 checked
Knowledge check
What should a weekly executive update lead with?
Complete
You have finished “Executive Communication, Handover and Saying No”. Continue to the next chapter or practise in the FDE simulator.
Mark this chapter as complete
Track your progress and unlock the next chapter.