Ask why people are leaving and get a written read with an inline chart: exit reasons grouped by theme, with the notable patterns called out. A snapshot on demand, grounded in your exit-interview data.
For
HR Business Partners
People Analytics Teams
Retention & Talent Teams
HR & People Leaders
Sample prompt to Joy's Assistant
Analyze our [exit interview] data from the last [12 months] and show me the top reasons people are leaving, broken down by theme, with the notable patterns flagged.
What it can look like
An illustrative example. Your output will reflect your own sources, data, and prompts.
Analyze our exit interview data from the last 12 months and show me the top reasons people are leaving, broken down by theme, with the notable patterns flagged.
Reading your exit-interview notes and survey responses for the trailing twelve months, 64 people left, most of them by choice. The reasons group into six themes, and two of the smaller ones are where the fixable problems actually sit, both flagged below.
I read the exit-interview notes and survey responses for the trailing twelve months. 64 exits in that window, 71% of them voluntary, and the reasons sort into six themes. Compensation is the single most-cited reason at 28%, but it is not the whole story, and two of the smaller themes are where the fixable problems sit.
64
Exits analyzed
71%
Voluntary
Compensation
Top reason (28%)
Here is how the voluntary and involuntary reasons break down across the six categories.
64exits
Compensation28%
Career growth24%
Manager relationship19%
Workload/burnout16%
Relocation/life8%
Other5%
Compensation leads at 28%, followed by career growth at 24%, manager relationship at 19%, workload and burnout at 16%, relocation and life events at 8%, and a small remainder of other reasons at 5%. The headline is compensation, but the categories underneath it hide the two patterns that are actually worth acting on.
Career growth clusters in the 2 to 3 year tenure band. Of the 24% who named career growth, the clear majority left between their second and third year, mid-career reps leaving to advance rather than early hires or long-tenured staff. One wrote that they "loved the team but could not see a next step here without waiting for someone above them to move." That is a mid-career ladder problem, not a pay problem, and it is concentrated enough to design a path for.
Manager relationship is concentrated in one department. The 19% who cited their manager are not spread evenly. Over half of that theme comes from a single department, which means the company-wide 19% is really one localized issue plus a normal background rate everywhere else. One departing employee put it plainly: "the work was fine, the manager was the reason I started looking." Address that one manager or department and this theme shrinks noticeably.
Compensation, by contrast, is broad-based. The 28% who named pay are spread across teams, tenure bands, and levels with no single hotspot, which is what you would expect from a market-rate question rather than a fixable pocket. It is real, but it is the slow structural lever, not the quick win.
The takeaway: inside that headline 28% compensation number sit two targeted, fixable problems. A mid-career growth path aimed at the 2 to 3 year band would address most of the 24% career-growth theme, and fixing one manager or department would take a real bite out of the 19% manager theme. Both are smaller, more concrete, and faster to move than the compensation line everyone reaches for first.
Want me to break the manager relationship theme down by department to name the one that's driving it, or cross the career-growth leavers against tenure band to confirm the 2 to 3 year cluster?
From a folder of notes to the pattern behind them
The Exit Interview Themes recipe reads your exit-interview notes and survey responses on demand and returns a written analysis with one inline chart: the reasons people are leaving grouped by theme, counted, and ranked, with the notable patterns flagged. It's produced when you ask, not a board that runs on its own.
Connect your exit data
Give Joy read-only access to your exit-interview notes and survey responses, or add an export with departure records, tenure, and department mapping to the Knowledge Center.
Ask for the themes
Ask for the top reasons people left over a window, grouped by theme. Joy reads the notes and survey data at ask time and sorts the reasons into categories.
Read the analysis and chart
Joy returns a short written read with one inline chart showing the reason split by theme, and flags the patterns worth acting on. The figures in the prose match the chart.
Take it to the room that decides
Copy the read and numbers into your retention review or leadership update. Joy drafts the analysis; you present it. Nothing is exported or scheduled for you.
Make it one click for your team
Save this ask as a custom command on the assistant your team already uses, so anyone can run it in one step.
Make it yours
Grouped By Theme
Sorts dozens of free-text notes into a handful of countable reason categories you can rank.
One Clear Chart
A single inline chart shows the reason split, so the top drivers are unmistakable.
Patterns Flagged
Calls out where a theme clusters by tenure or department instead of leaving it as a flat average.
Slice to the Source
Follow up by department or tenure band and Joy re-reads to trace a theme to its origin.
By Department
Break a theme down by department to see whether a reason is company-wide or localized.
By Tenure Band
Cross the leavers against tenure to separate early-attrition reasons from mid-career ones.
Voluntary Only
Filter to voluntary exits to focus on the reasons you can actually influence.
Compare Windows
Compare this year's theme mix against last to see which reasons are rising or fading.
Frequently Asked Questions
Is this a live exit-analytics dashboard?
No. It's a written analysis produced on demand with one inline chart. When you ask, Joy reads your exit-interview notes and survey data and returns the read for that moment. There's no persistent board refreshing on its own.
How does Joy group free-text notes into themes?
Joy reads the notes and survey responses and sorts each stated reason into a category such as compensation, career growth, or manager relationship, then counts and ranks them. In the example that yields six themes across 64 exits, with compensation the most-cited at 28%.
Can it flag patterns, not just totals?
Yes, that's the point. Beyond the split, Joy calls out where a theme clusters, like career-growth reasons concentrating in the 2 to 3 year tenure band or a manager theme where over half comes from a single department, so you see the fixable problem under the headline number.
Can I break a theme down by department or tenure?
Yes. Ask a follow-up and Joy re-reads the data to break the manager-relationship theme down by department, or cross the career-growth leavers against tenure band, useful for tracing a theme back to the team behind it.
Can Joy send this analysis out on a schedule?
No. Joy drafts the analysis in chat; you copy it into your retention review or leadership deck and share it yourself. Joy doesn't schedule recurring reports or send anything on your behalf.
