Discovery and scoping · Discovery

Three statements from a transcript

Mark three discovery statements a colleague wrote from a team-lead interview and rewrite them so a reviewer who was not in the room could test each one.

critiqueeasyExercise · 3 min · 9 of 64

On this page

A colleague ran the team-lead interview and wrote three statements. Two of them will be the goal line of a scoping doc by Friday unless somebody reads them properly first.

The situation

Your colleague spent 45 minutes with the claims team lead on the first on-site day and sent you three statements for the trip report, with the transcript attached. The team lead reports to the sponsor and will be in the room when the scoping doc is signed. The three statements:

Handlers hate the policy PDFs and would love a search tool.
Triage takes about 10–15 minutes and most of it is finding the wording.
The team lead says a chatbot would save 30%.

The relevant part of the transcript:

FDE:   How long does a claim take to triage?
LEAD:  Depends. Ten, fifteen minutes. The PDF bit is most of it. Everyone
       hates the PDFs. Last Tuesday we had 1,340 come in and I had 26 on
       the floor, two were out, so we were still going at six.
FDE:   Would something like a chatbot help?
LEAD:  Oh, massively. If it could just answer the question, I'd say you'd
       get thirty percent back, easily.
FDE:   Great, that's what the contract says too.

The claims system logs when a claim is opened and when its route is set. Your colleague has not asked for an export. The scoping doc is due on Friday, and the sponsor will read the goal line first and the rest only if the goal line holds their attention.

The question

Which of the three statements would you keep as written, and how do you rewrite the others so that a reviewer who was not in the room could mark each one true or false?

A strong answer · Rubric · Where answers go wrong

Write your answer first, then open

A strong answer

None of the three survives as written, and the second is the closest. Statement 1 records a feeling and a prediction; "everyone hates the PDFs" is a lead, not a finding, and "would love a search tool" was never said. Statement 2 has a number range but no source and no confirmation path. Statement 3 is the colleague's own pitch reflected back: the FDE proposed the chatbot, the lead guessed a figure, and the FDE matched it to the contract.

Rewritten so a reviewer can mark them:

On Tuesday last week 1,340 claims arrived and 26 handlers were on the floor; the floor was still triaging at 18:00 (team lead's recollection; to confirm from the claims system's daily count and open-at-18:00 figure).
Triage takes 10–15 minutes a claim by the team lead's estimate, "most" of it finding the policy wording (estimate only; a week of open-to-route timestamps is requested, and the search step needs document-system timestamps or an afternoon of timed observation).
Not a statement. The team lead's 30% is a reaction to a proposed solution and belongs, if anywhere, as the sponsor's goal in the scoping doc, where it already is. What goes in the trip report instead is what was not asked: whose idea the chatbot was, and what handlers do today when the PDF does not answer the question.

The evidence that reverses this is the log: if a week of open-to-route timestamps gives a median near 15 minutes rather than near 10, the goal line changes before anything else does. The test applied to each line is the one from discovery interviewing: could a sceptical reviewer who was not in the room mark it false, and from what.

Rubric

rejects the opinion statement without discarding the lead inside it
attaches a source and a confirmation path to every number kept
recognises that the chatbot was the interviewer's idea and the 30% the lead's reaction to it, matched to the contract by the interviewer, not a measurement
separates the floor's facts (1,340, 26, 18:00) from the lead's estimates
names the log export as the next step, with the fields it needs

Where answers go wrong

"Keep all three — they are what the customer said." Loses because the first adds a wish nobody voiced and the third is the interviewer's own pitch, agreed with; a reviewer cannot mark "would love" or "would save 30%" false, so neither is a finding.

"Drop statement 1; feelings are not data." Loses because the feeling points at the step to count. "Everyone hates the PDFs" is the reason to time the search step and ask for the document-system log; thrown away, it takes the pilot's hypothesis with it.

"Rewrite statement 3 as: the chatbot will save 30%." Loses because it converts a reaction into a forecast, and the forecast then appears in the scoping doc as a promise the pilot made to itself.