Question 2Discovery and scoping
How do you decide what to build first when a customer wants ten things?
FDEs have limited weeks per customer. This tests judgment about value, effort and risk.
“Run discovery and translate ambiguous business problems into clear requirements, scope, and acceptance criteria”Senior Forward Deployed Engineer, Applied AI · Juniper Square
The answerFour steps
- 1
Score each request by business value and by effort, with the customer in the room
Say it like this
“I'd put the list on a shared screen with their ops lead and score each one: hours saved per week, and rough days to build.”
Why it scores
Scores with the customer, not for them. They own the value side; you own the effort.
- 2
Pick one item that proves value fast and unblocks later work
Say it like this
“I'd start with the one that pays back fast and sets up the rest, like syncing their inventory data, because the forecast and the dashboard both need it.”
Why it scores
Picks the item that unblocks the others. Shows sequencing, not just ranking.
- 3
Write down what's out of scope so it doesn't creep back in
Say it like this
“Then I write the parked items into the project doc with one line on why, and the customer signs it…”
Why it scores
Parks scope in writing. This is how FDEs protect a six-week engagement.
- 4
Revisit the list after the first release, with real usage data
Say it like this
“After the first release, we watch real usage for two weeks and re-score the list.”
Why it scores
Re-ranks with real usage. Early guesses about value are often wrong.
StrongAll together
About a minute, out loud. The marked parts are the four steps; hover one to find it above.
Say a warehouse customer sends ten asks, from a new dashboard to an AI demand forecast. I'd put the list on a shared screen with their ops lead and score each one: hours saved per week, and rough days to build.1 Usually two or three stand out. I'd start with the one that pays back fast and sets up the rest, like syncing their inventory data, because the forecast and the dashboard both need it.2 Then I write the parked items into the project doc with one line on why, and the customer signs it,3 so they don't come back as “quick asks” in week three. After the first release, we watch real usage for two weeks and re-score the list.4
WeakThe answer that fails
Sounds fine. Common. Scores low.
“I'd ask the customer which ones matter most to them and start at the top of their list, moving fast so we can get through as many as possible this quarter.”
Interviewer's note
Hands the judgment back to the customer. No view on effort, dependencies, or what to drop.
PracticeTwo minutes, then they push back
Your turn
How do you decide what to build first when a customer wants ten things?
Answer like you're in the room: about two minutes, no notes. Say it out loud, even alone. It feels different from thinking it. Then check your answer against what interviewers score.