Turn customer discovery into a pilot plan
Practise a discovery conversation that produces a measurable outcome, a data inventory, an explicit scope and a first delivery milestone.
What you'll be able to do
- Map a customer's current workflow and identify the owner of each handoff
- Separate a measured baseline from a proposed improvement
- Write a pilot agreement with success criteria and unresolved assumptions
Before this: the-fde-role
Northline's support director asks for an AI assistant that handles warranty questions. Start by asking a support employee to show one recent ticket. Record what triggered the work, which systems they opened, what they checked and where another person became involved. Repeat with an exception case. The difference often reveals the first useful scope boundary.
Run a short discovery session
For this practice exercise, allow 30 minutes. Spend 5 agreeing on the outcome, 10 walking the current workflow, 10 checking data and exceptions, and 5 agreeing on the next step. These durations are a rehearsal budget you can adapt to the meeting.
The time budget totals 30 minutes. Reserve the final five minutes to agree on a deliverable, an owner and a review date.
Ask which warranty questions take longest, which mistakes cause rework, and who can confirm the applicable policy. Check whether employees must match a serial number, a purchase date or a sales region. Identify how they decide which policy version applies.
Calculate a baseline
Suppose observation finds 240 eligible tickets per day, with 12 minutes of handling time each. The proposed pilot aims for 8 minutes. That is 4 minutes saved per eligible ticket, or 960 minutes across a full day at complete adoption. At an assumed 60% adoption, the arithmetic becomes 576 minutes, or 9.6 hours of gross daily capacity.
This is a planning estimate. Confirm the volume and time with the customer, then measure the result during the pilot. Review time, rework, operating cost and work shifted onto another team all affect the realized benefit. Recovered capacity may be used to clear a backlog or improve service.
Produce a one-page pilot agreement
| Field | Northline's illustrative agreement |
|---|---|
| User and task | Support employees drafting answers to one product family's warranty questions |
| Inputs | Approved policies and ticket metadata available to that employee |
| Output | A draft with policy links and an escalation option |
| First milestone | Replay historical tickets and review the drafts with a policy owner |
| Pilot outcome | Measure handling time alongside correctness and rework |
| Operational owner | Support lead with a named engineering contact |
| Expansion decision | Review quality, usage and incidents together before widening scope |
Record assumptions beside this agreement. For example, the policy owner still needs to confirm how regional exceptions are represented, and IT still needs to confirm a supported ticket API. Give each uncertainty an owner and a next check.
Handle a change in the room
Prompt. The sponsor adds automated refund approval to the first milestone. The finance owner is absent. Respond as the FDE.
Worked answer. Explain the current draft-answer milestone and ask what refund outcome the sponsor needs. Capture the proposed action, financial limits and required approval policy. Offer to evaluate refund recommendations with finance using historical examples, while the agreed pilot continues through employee review. Schedule the decision with the accountable owner and record how it changes the delivery estimate.
Practice. Write five discovery questions for a slow employee-onboarding process. Your questions should reveal a daily user, a business metric, a data owner, an exception and an operational constraint. Then produce a pilot agreement from your answers.
Practise this stage with ChannelPulse
Add a target job to Your interviews, then use interviewer research and a practice session to rehearse your discovery questions. Set up your interview. Practise the conversation.
Reading and attribution
The discovery and stakeholder themes draw on Tech Mahato's business-consulting guide and bicced's customer and interview-round guides. The handbook is licensed under CC BY 4.0. This lesson uses original questions, a fictional customer and illustrative calculations. Consult the source links for the authors' own examples.