Xect AI is an outage customer-actions agent: for a hub, terminal or F1 cut, or a planned maintenance window, it identifies the affected addresses, notifies the customers it can reach with an ETA when one is available, holds tickets confirmed related to the outage and defers the truck rolls that won't help, and issues credits when it's over, so an outage costs the network team the repair, not the call center.
Outage awareness today starts the moment a customer calls to report it, then propagates queue by queue until someone maps the ticket to a location. Xect AI's agents use incident and service-impact data to scope the fault or the feed event first, coordinate the customer update, the dispatch decision and the applicable credit calculation, and maintain a shared incident record, closing many incidents before that first call would come in.
An alarm, a planned-maintenance ticket, a utility outage feed or a burst of ONT dying-gasp messages all land as the same kind of event. Any one of them is enough to open an incident and start the clock.
Hub, terminal or F1 topology translates directly into the list of affected addresses, with no manual lookup required. Each address carries its channel preference and its priority-account flag, so the notification and the escalation are already scoped.
The agent notifies affected addresses with an ETA when the incident type and topology support one, updating it as new information arrives, and updates the IVR and chatbot banners. It holds tickets confirmed related to the outage for that address, apart from unrelated open faults, and defers the truck rolls that would only reach a known outage, while priority accounts get escalated for a direct call.
Network faults are confirmed restored per ONT rather than assumed for the whole area; power-related outages close when the utility reports restoration. Credits get computed from each customer's actual downtime, and the incident then replays in the Process Miner so the next one runs faster.
Each agent owns one step in the cut-to-credit path above, whether the cause is a fiber cut or a neighborhood power outage.
Turns a hub, terminal or F1 fault into an address list
Reads the topology behind whatever failed, whether that's a hub, a terminal or a single F1 leg, and resolves it into the exact addresses affected before anyone drafts a notification.
Notifies, updates every channel, and holds work that won't help
Pushes the notification to affected addresses with an ETA when one can be estimated, updates the IVR and chatbot banners so a call already in progress hears the same story, and holds tickets confirmed related to the outage rather than every open ticket at that address.
Tells a dying network apart from a dark house
Correlates a cluster of ONT dying-gasp signals against the utility's own outage feed to confirm the cause is commercial power, not the network, before a truck gets dispatched to a house where power, not the network, is the fault.
Turns measured downtime into a traceable credit calculation
Measures downtime per ONT rather than estimating it for the whole outage, applies the credit policy automatically, and leaves a record precise enough that a customer calling to dispute it gets an answer instead of a promise to check.
Topology scoping turns straight into the address list and the send, with no manual pull in between.
The Power Outage Manager tells a commercial power loss from a network fault before a dispatch decision gets made.
Customers who are notified see an ETA when one is available, often before they'd think to call in.
Downtime gets measured per device, not estimated for the whole neighborhood.
An outage should cost you the repair, not the call center.
A hub, terminal or F1 cut scopes to the addresses behind it, and the same notify-hold-credit flow runs end to end.
See the industry pageA node outage scopes to every address on that node, and the same agents notify, hold and credit them the same way.
Talk to us about cable & HFCA tower power loss scopes to every subscriber served from that tower, distinguished from a network fault the same way a dark house is.
Talk to us about fixed wirelessThe utility outage feed is the second half of the Power Outage Manager's correlation, confirming a dark house is commercial power, not the network.
Talk to us about utilitiesA municipal network runs the same cut-to-credit flow, sized for a network that answers to its residents directly.
Talk to us about municipal broadbandOwns detection and restoration, and needs every alarm, maintenance ticket and dying-gasp cluster to register as one incident instead of five separate alerts.
Owns the queue during an outage, and needs call volume to fall because customers already know, not because agents were told to read a script.
Owns dispatch, and needs a truck rolled only when the fault is real, with priority accounts escalated ahead of the rest instead of lost in the same queue.
Owns outage reporting and credit-policy compliance, and needs restoration timing and credit calculations to hold up as evidence instead of being reconstructed after the fact.
Book a 30-minute demo and run a simulated F1 cut through notification, ticket holds and credits on your topology.