---3.png){kind=logo}
Identifies Risk Early
Surfaces a member dropping into amber on the Keepme Score with weeks of runway.
Meet Ember
Ember is the retention agent for Antares. It identifies members at risk of leaving, intervenes with the right action at the right time, and engages directly when a member chooses to cancel.
What Ember Does
Intervenes at the right moment
Engages the at-risk member through WhatsApp, in-app, or at checkin.
Wave Leisure
Your Member App Assistant
Type your message...
Engages when members cancel
Reactive cancellation save function surfaces the most relevant alternative so the member can stay.
Cancellations
JungleGym Member App
⌂
Type a message...
➤
BC
BigCityGyms
Active now
Aa
➤
🏋
FitLife Gym
AI Support Assistant
Type your message...
➤
Why this matters
Most fitness operators lose around a third of their members every year. By the time a member reaches the cancellation form, the decision is largely made with no real save conversation. The real opportunity is earlier, when the member's behaviour is starting to shift but they haven't yet decided to leave. That's where Ember intervenes.
Introducing Keepme Score
Every member has a Keepme Score, a 0 to 100 number that reflects their relationship with your business. The member's Keepme score is generated by Pulse the intelligence layer underneath Antares.
Green
Healthy engagement. The member is active, behaviour is stable, and no intervention is needed.
Amber
Behaviour is drifting. This is the moment where intervention has the highest expected value.
Red
The member has already disengaged. Recovery is harder and requires more aggressive action.
Pulse runs predictive models against your member data continuously. The score updates as the member's behaviour changes. Ember uses the score to know who to engage, when, and how. You set the rules for what happens at each band. ✓ Different interventions for amber than for red ✓Different cadences for high-tenure members than for new members. The platform does the prediction. You stay in control of what happens next.
The Math
Economics That Scale With Your Member Base.
The cancellation save economics are immediate and measurable. When a member chooses to cancel, Ember either saves them or processes the cancellation cleanly. The numbers below are conservative and based on what most operators see in the first months of running Ember. The example below uses a conservative 10% save rate, an average monthly fee of $50, and assumes a saved member stays for an additional six months beyond the point they would have cancelled.
Six additional months is a floor, not a forecast. Average member tenure in the industry runs well above that, and the actual retained value of a saved member is materially higher than the table suggests. The numbers above are what the operator can defend without argument. The real economics are better. The numbers move further when the save rate is higher, when the operator's average fee is higher, when actual remaining tenure is modelled, or when acquisition cost replacement is factored in. They move further still when the reason data Ember captures is used to fix the systemic drivers of attrition upstream.
The proactive layer compounds beyond the cancellation save numbers. By identifying members at risk months before they would otherwise leave, Ember reduces the cancellation volume the save function sees. The full retention impact compounds across a 12-month horizon. The operators who get the most value from Ember are the ones who deploy both layers, with the proactive layer working upstream and the cancellation save layer catching what remains.
Where Ember Sits in Antares + What It Captures
Ember runs on the same Antares platform as every other agent in the suite. Configure once, and the same operating logic carries across every agent.
Operator control
The operator controls what Ember is allowed to offer, including freezes, tier changes, location switches, discounts, concessions, and escalation rules. Ember does not invent offers.
Shared knowledge base
Because Ember shares the same knowledge base, brand voice, policies, and integrations as the rest of the Antares agents, there is no separate cancellation system to manage.
Real retention
insights
insights
Ember records the member’s risk score, captures what they said in their own words, and classifies it into a structured reason. Operators see the real drivers of cancellation and retention, not just the limited answers from a dropdown form.
Offer and outcome
data
data
Every conversation shows what Ember offered, whether the member accepted it, and what happened next. This makes it clear which interventions are working and where new save options may be needed.
Patterns across the estate
Ember helps operators see whether cancellation reasons are isolated incidents or part of a wider pattern. Those patterns do not stay trapped in cancellation. When Ember identifies why members are leaving, that data feeds back into Nova, helping it qualify leads more intelligently at acquisition. The agents learn from each other.
Better member economics
Ember turns cancellation conversations into lifecycle data. Each save, downgrade, pause, and cancellation can be viewed alongside lead source, tier, tenure, location, and revenue. That gives operators stronger evidence for LTV modelling, sharper retention decisions, and a clearer view of which members stay, leave, and why.
Cross-agent intelligence
Reason data Ember captures feeds back into how Nova qualifies at acquisition. When Ember identifies that members are leaving for a particular reason, Nova adjusts how it qualifies leads at the top of the funnel. The agents learn from each other.
Pulse Predictive Layer
Ember runs on Pulse, the intelligence layer underneath Antares. Pulse generates the Keepme Score, surfaces behavioural patterns, and predicts where Ember should intervene next. Every Ember conversation feeds back into Pulse, which improves predictions over time.
What You Configure
Most of what governs Ember is set at the Antares platform level. The knowledge base, the brand voice, the integrations, and the policies are shared with every other agent. Ember-specific configuration:
- The offer set: which interventions Ember is permitted to make, and within what limits
- The escalation rules: which conversations route to a human, and to which human
- The cancellation completion logic: what happens when save is not the right outcome
- The Keepme Score bands and what triggers engagement at each band
- The intervention rules per band including which channels, which messages, which offers
- The channels Ember uses to engage at-risk members
Every Ember conversation is recorded, reviewable, and auditable. Nothing the agent does is a black box.