Xect AI is an event-driven churn model. It assesses every account's disconnect risk using the signals that precede it: outages, repeat calls, speed shortfalls and help-center searches, then routes eligible accounts to a retention workflow within policy, so outreach starts before the cancel call.
A churn score built from last month's billing exports already missed this week's outage, the third repeat call and the disconnect search that just happened. Xect AI scores every account on the signals as they occur, so retention workflows start while there is still a customer to retain.
Customer 360 events and location intelligence arrive as they happen, not on the next batch run.
Features, including outages, hub troubles per 100, install NPS, competitive changes, bandwidth and payment shifts, drive the model on every update.
The top drivers behind each account's score come back in plain language, so the retention action fits the cause.
Retention actions run as flows, and anything above policy pauses for a human to approve.
Each agent owns one part of the score-to-retention path: reading the signal, updating the score, and running the retention action that fits what caused the risk.
Turns every system's activity into a churn feature
Streams feature events from network, care, billing and digital systems the moment they happen, and adds neighbor and hub context so an account's risk reflects its surroundings, not just its own history.
Updates the score and explains what moved it
Re-scores an account the moment a new event arrives, retrains on a set schedule to stay current, and raises a drift alert when the model's behavior starts to shift.
Matches the offer to the cause, then follows up
Matches the retention offer to the driver behind the risk, such as a service fix for a network issue instead of a blanket discount, since resolving the underlying issue removes the reason for the risk rather than masking it. Books the callback when a human touch is the right play, and feeds the outcome back so the model learns what worked.
Outages, hub troubles, speed variance, NPS, calls, searches and bill changes all feed the score as they happen.
A help-center search for cancellation starts a retention workflow in the same hour it happens, not at the next model run.
Every score comes with the top three reasons behind it, so the retention action fits the cause.
Pilot hubs run the live score and same-hour retention workflow against the quarterly model as a baseline. The comparison is in progress, with results to follow.
Retention that starts before the cancel call is easier to scale than retention that starts after.
A subscriber who calls about a second outage this month, searches the help center for “cancel,” and asks a retention rep about a competitor's rate during that call scores as high risk within the hour, not at next month's model run.
See the industry pageIn a competitive retail energy market, a rate hike, a billing dispute call and an early contract termination request are exactly the pattern the model flags before the switch completes.
Talk to us about utilitiesRenewal risk carries the same pattern as churn: a claim handled poorly, a rate increase and a call asking about the cancellation fee. The model scores these as events and flags the policy before renewal.
See the industry pageA drop in product usage, a failed payment retry and a support ticket about downgrading are the same kind of event-driven churn signal in any subscription business, not just telecom.
Talk to us about SaaSA viewer watching less each week, a failed payment and a “cancel subscription” search in the help center is a candidate for a retention workflow the same day, not at the next billing cycle.
Talk to us about media & streamingOwns retention spend and offer timing, and needs the drivers behind an account's risk before deciding which offer to send.
Owns the churn model's accuracy, and needs features that update with the business instead of a monthly rebuild on stale exports.
Owns the care team's queues, and needs repeat-contact risk visible before the third call comes in, so retention outreach happens upstream of the cancellation request.
Owns the revenue forecast, and needs churn tracked as a live KPI with offer spend broken out by the cause it was meant to fix.
Book a 30-minute demo and see an account's risk score move with its events, then walk through a retention workflow.