ZenotiHelp

AI Retention Manager overview

ZTZenoti Docs Team Updated October 25, 2026 4 min readGuide

The AI Retention Manager is Zenoti's AI-powered guest churn-risk dashboard. It learns each guest's visit, spend, and booking rhythm, then surfaces which guests are drifting before they stop visiting β€” so your team can act at the right moment rather than after the fact.

Before you begin

  • The AI Retention Manager is accessed from the Guest menu in the Zenoti left navigation. It is an embedded dashboard; there is no separate login screen. If your session is not valid, the dashboard shows an error state rather than a sign-in prompt.
  • If your organization has been onboarded, the dashboard shows a training screen while the Retention Manager builds its predictions. The screen states "Usually ready within 24 hours." Risk data appears once training is complete.
  • Data displayed in the dashboard is near-real-time with a maximum delay of 24 hours.

Permission: AI Retention Manager security object β€” View to open the dashboard and read it; Edit to create target segments, refresh member lists, and send guests to AI Lead Manager. Set per role at Configuration > Security > Security roles.

How it works

The AI Retention Manager evaluates each guest against their own historical rhythm β€” not a fixed calendar β€” and assigns them to one of four bands based on their likelihood of stopping visits. This band assignment drives every element of the Overview page.

When an employee opens the AI Retention Manager for the first time, the dashboard loads directly on the Overview page, on the Repeat Guests tab. The in-product Help drawer is available on every visit (not the first), and walks through the "Start here" routine, what each widget is for, and what the four retention bands mean.

If the organization's data has not yet finished processing, the dashboard shows a training screen instead of risk data.

The four retention bands

The Retention Manager places every guest in exactly one of four bands:

Band Meaning Suggested action
Critical risk of churn Guest is leaving now Reach out today
High risk of churn Guest is slipping Check in soon
Monitor Early sign worth watching Keep an eye on this guest
Healthy Guest is on track No action needed

Note: The five internal risk levels (VERY HIGH through VERY LOW) are never shown in the dashboard. Guests are always described using the four operator-facing bands above.

Tabs: Repeat Guests and First time Visitors

The Overview page offers two guest-population tabs:

  • Repeat Guests β€” guests who have visited more than once. The summary card is labeled Retention Score.
  • First time Visitors β€” guests who have visited only once. The summary card is labeled First time Guest Retention Score, showing the share of first-time guests expected to return for a second visit on their own.

Both tabs present the same set of widgets, scoped to their respective guest population.

Action Bar navigation

The Action Bar renders four primary section buttons in this order: Overview, Guests, Proof of value, Reports. A fifth button β€” How the Retention Manager works β€” is right-aligned and opens the explainer view.

Key features

The Overview page is organized around the following Widgets:

  • Retention Score β€” Out of every 100 guests, how many look set to keep visiting. The score counts guests in the Monitor and Healthy bands. Use this as a size-of-the-problem check before opening the Retention Groups. On the First time Visitors tab the equivalent widget is the First time Guest Retention Score.

  • Retention Groups β€” Shows how many guests sit in each of the four bands: Critical risk of churn, High risk of churn, Monitor, and Healthy. Selecting a band row steers the About this group panel (docked beside the rows) to that band, so you can read the signals driving that group without leaving the page.

  • About this group panel β€” Displays insight for whichever band is selected in Retention Groups. Navigate between insights using Previous insight and Next insight.

  • **Total guest count** β€” The count of guests in scope for the current tab and center selection.
  • Movement tracker β€” Shows how many guests moved between bands over a rolling 30-day window, and the average time from first signal to band change. If movement data is not yet available, the widget displays "Movement data is not available yet."

  • Guest Retention Strategies β€” Presents strategic actions tied to retention risk, with the option to create a target segment and hand off to the Zenoti campaign builder.

  • Top services & providers β€” Highlights which services and providers are most associated with retained guests.

  • **Client watch list** β€” Your guests ranked by leaving risk. Reach out to the top of the list first. Sortable by **Sort: Risk (high to low)**, **Sort: Risk (low to high)**, and **Sort: Guest name (A-Z)**. Searchable by guest name or ID.

In-product Help drawer

The Help drawer is available on every visit from the Help button in the dashboard header. It covers:

  • Start here β€” a daily routine for working the overview.
  • Finding your way around β€” what each widget is for.
  • Common questions β€” answers to frequent questions about reading the page.

The drawer also links to How the Retention Manager works for a deeper explanation of how guests are placed in each band and where the tool has known limits.

Prediction signals

The How the Retention Manager works view lists the signals the Retention Manager uses to score guests, ranked by how much each signal moves the risk score. These signals apply across the whole organization's predictions β€” they are not specific to a single center.

While signals are being prepared for a newly trained model, the dashboard shows: "The live signal list for your business is being prepared. Everything above is up to date. Come back shortly to see which signals the Retention Manager uses."

Note: "Likely to stop visiting" is an estimate based on each guest's recent visit, spend, and booking patterns. Use it as a prompt to reach out, not a certainty.

Was this article helpful?
Your feedback helps us improve our documentation.