Research Debrief Template: Turn Interviews into Findings and Decisions

A research debrief template turns individual interview material into a disciplined team review. Its purpose is not to manufacture consensus. It creates a traceable path from source evidence to provisional pattern, finding, implication and decision.

The debrief begins with usable session records. Prepare them with the note-taking and outlining guide and keep quotations, observations and interpretations distinct.

Choose the right debrief level

Use two related formats:

• Session debrief: ten to twenty minutes soon after one interview, focused on context, strongest evidence, surprises and follow-up.
• Cross-session debrief: a longer comparison across participants, focused on patterns, differences, confidence and research implications.

Do not wait until every interview is complete to notice that a question is confusing. Do not declare a cross-study finding after one memorable session. Each level has a different job.

Re-anchor the team on the research decision

Open with the decision the study is meant to inform and the research questions. This keeps the discussion from becoming a general collection of interesting comments.

Write these at the top of the board or document:

Decision this research informs:
Research questions:
Scope and participant criteria:
Evidence collected so far:
Known limitations:
Decision owner:

The decision owner does not own the facts. Everyone should be able to challenge an interpretation by returning to the evidence.

Copy this research debrief template

RESEARCH DEBRIEF

Study:
Debrief type: session / cross-session
Date:
Facilitator:
Participants:

1. RESEARCH DECISION
Decision to inform:
Questions in scope:

2. EVIDENCE
Session reference:
Observed behavior / verified statement:
Source location:
Relevant research question:

3. INTERPRETATION
What this may mean:
Alternative explanation:
Confidence and limitation:

4. COMPARISON
Supporting examples:
Contradictions / negative cases:
Contexts where pattern does not apply:

5. OUTPUT
Provisional finding:
Implication:
Decision or experiment:
Owner and review date:
Open research question:

Keep source links or stable references beside each evidence item. A wall of anonymous sticky notes is difficult to audit later.

Start with silent evidence review

Give participants time to read the relevant notes before discussion. Silent review reduces anchoring by the first speaker and gives quieter team members a chance to notice contradictory evidence.

Ask each person to mark:

• evidence relevant to the research question;
• a surprise or changed assumption;
• a contradiction;
• a missing context;
• an interpretation that needs source checking.

Use research repository structures to keep source references available without exposing unnecessary participant data.

Build patterns from evidence, not labels

Cluster evidence by behavior, trigger, constraint or outcome. Avoid beginning with a favored solution category and forcing every observation beneath it.

For each proposed pattern, complete this table:

Frequency in a small qualitative sample is descriptive of that sample, not a population estimate. Avoid percentages that imply statistical precision the study does not provide.

Preserve contradictions and negative cases

Contradictions are often the most useful part of a debrief. Two participants may use the same workaround for different reasons, or one participant may describe a process differently from what they demonstrate.

Do not resolve disagreement through majority vote. Ask whether the apparent conflict comes from role, environment, task frequency, experience, accessibility, policy or a difference in the question asked. Record the conditions under which each account may be true.

The objective summary guide helps reduce material while retaining qualifications and disagreement.

Distinguish findings from implications

A finding describes a supported pattern in the evidence. An implication describes what the organization may do because of it. A feature idea is neither evidence nor automatically the correct implication.

Example:

Evidence: Four administrators checked an older record before approving an exception.
Provisional finding: Exception approval depends on historical context not visible in the current view.
Possible implication: Test whether showing prior exceptions improves review confidence.
Not justified: Users need an AI approval engine.

This separation prevents the debrief from laundering a stakeholder preference into a research conclusion.

Give every decision an owner and review point

The debrief may produce a design experiment, an additional research question, a documentation change or a decision not to proceed. Record who owns the action and when it will be reviewed.

Use a decision log for consequential choices so later teams can see the evidence, alternatives and assumptions available at the time. Use meeting follow-up to communicate actions without attaching raw participant material to broad channels.

Handle recordings and sensitive evidence carefully

Only use recordings collected through an approved process with clear notice, agreement, appropriate access and retention. A debrief audience should not automatically receive access to raw audio or complete transcripts. Share the minimum evidence required for the legitimate research purpose.

When authorized research sessions are captured with agreement, Kuno can support visible recording and draft notes for human verification before a debrief. Researchers remain responsible for participant rights and every interpretation. Explore Kuno

If a quotation contains sensitive or identifying context, consider whether a faithful paraphrase or restricted source reference can support the analysis with less exposure. Follow the approved research and privacy plan.

Use AI as an assistant, not an evidence owner

AI can sort authorized text, suggest candidate clusters or format the debrief record. It can also erase qualifiers, merge distinct participants and produce a coherent theme unsupported by the source.

Require a researcher to verify every source link and quotation. Preserve the uncategorized and contradictory items. Do not ask a model to infer emotion, identity, health, honesty or personality, and do not let generated confidence scores substitute for methodological judgment.

Facilitate disagreement productively

When interpretations conflict, use a fixed sequence:

1. Restate the research question.
2. Show the source evidence.
3. Name each interpretation.
4. Identify assumptions behind each view.
5. State what evidence could discriminate between them.
6. Record the current decision and uncertainty.

Seniority should not determine which participant evidence is real. The facilitator protects inquiry, time and psychological safety while preventing circular debate.

Produce a concise debrief output

Close with a one-page output containing:

• the research decision and scope;
• three to five supported findings at most;
• important contradictions and limitations;
• implications clearly labelled as implications;
• decisions, owners and review dates;
• open questions and planned evidence collection;
• links to restricted source material where authorized.

Store this output in the UX research repository so future teams can trace it to sessions instead of copying findings without context.

Run a final evidence check

Before distributing the debrief, confirm:

• Every finding links to identifiable source evidence.
• Quotations are verified and contextualized.
• Interpretations and implications are labelled.
• Negative cases remain visible.
• Claims do not exceed the sample or method.
• Personal information is minimized.
• Actions have owners and review dates.
• Access and retention follow the approved plan.

A clean narrative is less valuable than a traceable one. Kuno can help create reviewable draft notes from consented capture, while the research team owns source checking, synthesis and decisions. See Kuno