# AI-assisted research synthesis brief

Outcome Driven · Ashley Kays Rabbitt

Define the question, the material, and the checks before you hand anything to an AI tool. Keep customer details out unless you have approval, and verify every claim against the source. Treat the output as a draft to examine, not as a finding.

## 1. What question are you trying to answer?

One question you could defend an answer to.

[Your notes]

## 2. What source material will you use?

List the notes, transcripts or tickets, and where each came from.

[Your notes]

## 3. What must be removed or approved before using a tool?

Names, contact details, account data, anything under agreement.

[Your notes]

## 4. Which tool and settings will you use, and why?

Note how it handles your data and who else can see it.

[Your notes]

## 5. What will you ask it to do?

Request themes with supporting quotes, not conclusions. Name what to do with uncertainty.

[Your notes]

## 6. How will you verify its output against the source?

Spot-check quotes, counts and themes by hand.

[Your notes]

## 7. Where might the output be wrong or too tidy?

Look for missing minority views and claims with no source.

[Your notes]

## 8. What will you decide, and who makes the call?

A person owns the conclusion.

[Your notes]

## 9. What will you save about how this was done?

Prompt, date, version, checks and corrections, for the next person.

[Your notes]