AI Prompts for Exit Interview Analysis

Exit interviews are one of the few times employees tell the truth about why they are leaving — if the interview is designed to let them, run by someone they trust, and analyzed as a set rather than as anecdotes. The value is in the patterns: reasons that cluster by manager, team, tenure band or role family, and the gap between the reason recorded in the HRIS ('better opportunity') and the reason actually given ('I could not see a path here').

These prompts design the interview, code the findings, and find the patterns. They connect to retention analysis, which uses attrition data, and to the survey pages, which hear from the people who stayed.

Before you use these

Have these ready to replace the highlighted [variables]:

• Current exit process: who interviews, when, what is asked, where data goes
• Exit interviews or surveys from the period (de-identified)
• HRIS leaver data: role, tenure, manager/team, recorded reason
• Anonymity threshold and who sees the analysis

The prompts

• 1. Design an exit interview that gets honest answers
• 2. Code exit interviews into consistent reasons
• 3. Find the patterns across leavers

1. Design an exit interview that gets honest answers

Act as a people analytics partner redesigning the exit interview for [company].

Current process: [who interviews, when, questions, where data goes, participation rate]
What we learn and miss: [summary]
Constraints: [HR capacity, locations, languages]

1. Interviewer and timing: who should interview (not the direct manager; HR or a trained neutral), when (after the decision is settled, before the last day, with an optional follow-up after they leave), and how participation is invited.
2. Question guide: opening on confidentiality and purpose; the decision (when they started thinking about leaving, what triggered it, what would have changed their mind); the experience (manager, team, work, growth, recognition, workload, fairness); the destination (what they are going to and what it offers that we did not — phrased without prying); what they would tell leadership; probes that get past politeness.
3. Confidentiality statement in plain language: aggregation, who sees what, how quotes are used.
4. Short exit survey (10 items) for those who decline the interview, aligned to the engagement survey themes so results compare.
5. Recording and coding: how answers are captured consistently for analysis; the reason taxonomy used.
6. Data handling and retention — to verify with policy.

Design for honesty and comparability. An exit interview run by the manager collects politeness.

2. Code exit interviews into consistent reasons

You are coding exit interviews for [company / period].

Interviews: [de-identified notes with role family, tenure band, team size band — no names]
Reason taxonomy: [existing codes, or propose: manager relationship, growth/career, compensation, workload/burnout, role fit, flexibility/location, culture/fairness, personal/external, restructuring]
HRIS recorded reasons: [per leaver]

1. Code each interview: primary reason, secondary reason, with the quote that supports each; confidence (clear / probable / unclear).
2. Trigger vs underlying: separate the event that triggered the decision from the underlying dissatisfaction, where the interview shows both.
3. Compare with the HRIS reason; classify mismatches (e.g. HRIS 'better opportunity', interview 'no path here').
4. Taxonomy gaps: reasons that do not fit the codes; propose additions.
5. 'Would have stayed if': what leavers said would have changed the decision — coded.
6. Quotes worth keeping, checked so they cannot identify the speaker.

Code what was said. Do not infer reasons the leaver did not give.

3. Find the patterns across leavers

Act as a people analytics partner finding patterns in exit interview data for [company / period].

Coded reasons: [per leaver: primary, secondary, trigger, 'would have stayed if', with role family, tenure band, team/group above threshold, level band]
Population: [headcount by the same groups, for comparison]
Anonymity threshold: [n]

1. Reason distribution overall; changes versus the prior period.
2. Concentrations: reasons that cluster in a group (role family, tenure band, level, team above threshold) at a rate materially above the population; state counts and the comparison; do not report groups below the threshold.
3. Tenure patterns: reasons by tenure band (first-year leavers versus long-tenured), and what each implies (hiring/onboarding versus growth/management).
4. Manager-related patterns: reported at team or department level above the threshold, framed as a support need, not an individual verdict.
5. Fixable causes: patterns that point to a specific policy, process or capability gap, with a recommended action and owner; patterns that need more evidence (stay interviews, survey cuts) before acting.
6. Summary for leadership: three findings, three actions, one page.

Report patterns above the threshold only. This analysis informs policy and support, not decisions about individuals.

Worked example

Related prompts

• Employee Retention Analysis
• Engagement Survey Analysis
• Employee Survey Design
• Manager Coaching
• Career Pathways and Ladders
• HR Analytics and People Metrics

Logical next step

After this, most HR teams move on to Employee Retention Analysis.

All Human Resources prompts · Search the full library