Turnover risk is a forward-looking insight that identifies which groups of employees are more likely to see departures over the next 12 months. This is surfaced as a Multi-signal Intelligence™ insight.

It's different from the historical turnover data in Retention Insights, which shows exits that have already happened. Turnover risk is about what might happen next.

When Culture Amp surfaces a turnover risk insight, it means we've identified a demographic cohort in your organization - for example, employees in a particular team, location, or tenure band - whose survey responses and other data signals point to a higher likelihood of departure. See below for further details on how turnover risk is calculated.

Insights are provided straight to Coach so you can spot where people are at high risk of leaving and start a guided conversation with Coach about what to do next. Open it and Coach walks you through what the data shows and how you might take action. Review how Coach helps you act on these signals here.

To begin monitoring for turnover risk, you require Multi-signal Intelligence™ configured on your account with the following sources:

We also recommend a closed survey that includes a question representing 'intent to stay', configured for Multi-signal Intelligence. Coach uses this question to explore what is linked to intent to stay for a group at risk.

For details on configuring Multi-signal Intelligence™ and configuration of specific data sources, see the Multi-signal Intelligence setup and administration guide here.

The insight appears as a card in the insights carousel when a group at elevated turnover risk is identified. Groups at low risk, or where risk cannot be assessed, will not be shown. If no qualifying group can be identified, no card will be shown.

The insight will be shown to Account Administrators and users with an Insights role in the AI Coach side pane. It appears in two places:

These are the only two places the insight appears. You will not see it in Coach surfaces tied to a specific survey report, performance review cycle, or other context. For how the carousel works, see Where are insights shown?.

Turnover Risk insights stay in the carousel for as long as your data continues to identify the group. Opening an insight does not remove it, and you see the same insights on any browser or device. For full details, see When insights appear and disappear.

Insights are checked each time you start a new Coach conversation, using the latest data available for your account. The insight card shows automatically when a group at risk is identified.

When there is a qualifying group at risk in your data, the card will show:

The wording on the card is produced directly from your data. Coach does not write it. The same data always produces the same card, so two people looking at the same signal see the same group.

The entire card is clickable and clicking it will start a conversation with Coach.

Use the insight when you want to understand a group at risk and decide what to do about it:

After you open the insight, Coach starts a conversation focused on the group named on the card. Because you have already seen the group, Coach goes straight to why it stands out:

Coach will recommend possible actions, framed around what you can do directly, what you can equip your leaders to do, and what you can raise upward. You can keep chatting to dig deeper, ask follow-up questions, or focus on one part of the group. Coach builds on the conversation as you go. Review how Coach helps you act on these signals here.

Turnover risk is calculated using a combination of signals available within your Culture Amp account. Inputs such as 'Intent to Stay' survey responses, tenure and manager changes (and many more from your employee data) are combined in a model that calculates a risk score at the individual employee level.

Rather than presenting a risk score for every possible way your workforce could be filtered, Culture Amp automatically identifies your most significant turnover risk hotspots - the demographic cohorts where the signal is strongest and most worth a leader's attention.

A hotspot is a group of employees defined by shared attributes - such as a particular department, or a combination of location and tenure band - that the model identifies as having an elevated level of turnover risk (high or medium) compared with your organization overall. Hotspots are built from the demographics added in Settings > Insights, using one demographic or a combination of two.

To ensure the signal is reliable and that individuals cannot be identified, we only surface hotspots that include at least 25 employees. Hotspots are ranked by significance, so the most important ones surface first in your Coach conversation. Find out more on the science behind turnover risk here.

When analyzing a hotspot, you'll see predictors - the signals most associated with the elevated risk for that specific group.

A predictor might be that the group gave notably low Intent to Stay responses compared with the rest of your organization, or that its tenure profile is concentrated in a band the model links to higher turnover. Predictors are a starting point for understanding why a group has been flagged - they show the input signals most influential to the model for that group, and can point toward the conversations worth having. In Coach conversations, predictors are called contributing signals.

At Culture Amp we've studied turnover for years, and "intent to stay" (I see myself still working here in two years' time) is the strongest signal we've found, though not the only one. When people disagree with this question, they tend to leave sooner, consistently enough to treat it as an early warning. Because intent to stay sentiment shifts before an exit happens, you can act while the outcome is still open.
​

We read intent to stay alongside other signals our people science research links to turnover, such as tenure and recent manager changes, to find where risk is concentrated and what's driving it. The result describes conditions across a group, not which people will leave. Three things are worth keeping in mind:

Finding the signal is data science. A score on its own doesn't tell you what's happening or what to do, and that's where people science comes in: to explain what the signal means and guide your response.
Coach brings that guidance into a conversation about the flagged group, supporting you in three ways.

​Explains what the signal means. Coach starts with the predictors, the factors most tied to the group's risk, so you see why it was flagged, not just how risky it is. It weighs those factors against what you know about the group and checks them against survey data like intent to stay and engagement scores. When the evidence is thin, it says so rather than forcing an explanation. That's the difference between a number and an insight you can use.

​Recommends the right next step. Coach draws on Culture Amp's retention research to suggest moves suited to the group, not a generic checklist. For a group at turnover risk, for example, it might suggest you:

Coach sorts these into what you can do directly, what you can equip your managers to do, and what to raise with senior leaders, so the response matches your span of control.

​Keeps the insight responsible. Risk is shown at the group level only, never as certainty about a named person, and always as a starting point for your judgment rather than a verdict. This protects your people's confidentiality and lets you trust what you're seeing.

Turnover risk is a useful signal, but it has important limits:

💬 Need help? Open a Support Conversation to speak with our Product Support team.