The problem

What the customer was up against.

A tier-1 carrier with 42M subscribers was bleeding ~3.2% monthly churn in its consumer mobile base — concentrated in price-sensitive prepaid and family-plan segments. Existing retention was reactive (act when the customer called to cancel) and undertargeted (one-size-fits-all offers). Retention teams couldn't differentiate customer value or context at the moment of intervention.

The solution

What xyner built.

Deployed xyner agents across at-risk identification, proactive intervention, in-conversation retention and post-churn win-back — fully integrated with the BSS, network behaviour signals, customer-comms platform and offer engine.

The outcomes

Measured impact.

Annualised churn in the targeted segment cut 18 percentage points; retention conversation handle time down 35%; offer-acceptance rate up 22 points; $128M annualised retained revenue.

Executive summary

At a glance.

Situation

A tier-1 carrier with 42M subscribers was bleeding ~3.2% monthly churn in its consumer mobile base — concentrated in price-sensitive prepaid and family-plan segments. Existing retention was reactive (act when the customer called to cancel) and undertargeted (one-size-fits-all offers). Retention teams couldn't differentiate customer value or context at the moment of intervention.

Intervention

Deployed xyner agents across at-risk identification, proactive intervention, in-conversation retention and post-churn win-back — fully integrated with the BSS, network behaviour signals, customer-comms platform and offer engine.

Outcome

Annualised churn in the targeted segment cut 18 percentage points; retention conversation handle time down 35%; offer-acceptance rate up 22 points; $128M annualised retained revenue.

Industry

Telecom · Consumer mobile

A tier-1 mobile carrier in North America

Scope

United States

At-risk segment identification, intervention dispatch, retention conversations, win-back

Duration

9 weeks pilot, 7 months full rollout

From contract signature to full rollout.

Architecture

What the deployment actually looks like.

Carrier-scale customer-base analytics requires near-real-time signal aggregation across BSS, network usage and customer-care touchpoints. The deployment uses streaming data feeds with bounded latency from signal to intervention.

At-Risk Detection Agent

Combines BSS signals, network-behaviour signals (data-usage trends, location changes, secondary-SIM detection), and care-touchpoint signals to score churn risk per subscriber.

Intervention Dispatch Agent

Selects the right intervention per customer — proactive outreach, retention conversation, family-plan offer, network-performance investigation — and dispatches via the right channel.

Retention Conversation Agent

Joins inbound retention calls and proactive-outreach conversations as a copilot for the agent; surfaces context, suggested offers and the customer's full history in real time.

Offer Engine Integration

Calls the carrier's offer engine for personalised, margin-protected offers; never proposes offers that breach commercial policy.

Network & CRM integration

Pulls real-time network-performance data for the customer's location to address signal-quality concerns substantively, not deflectively.

Outcome Capture

Every conversation outcome captured; offer-acceptance, customer reasoning, and post-intervention behaviour fed back to recalibrate scoring and offers.

Implementation timeline

How the rollout sequenced.

A 9-week pilot covered one customer segment (price-sensitive prepaid) before expanding across segments and geographies.

Weeks 1-3

Streaming foundations

Deploy data plane; integrate Kafka streams from BSS, network-behaviour platform and care-touchpoint platform; configure RBAC inheritance from carrier IdP.

Weeks 4-5

Agent & model configuration

Configure four agents; calibrate churn-risk model against historical data; integrate offer engine and customer-comms platform.

Weeks 6-7

Shadow mode

Agents score and recommend interventions in shadow; retention teams compare against current process; thresholds calibrated.

Weeks 8-9

Prepaid pilot live

Live for prepaid segment; metrics reviewed daily; weekly business review with VP Retention and VP Customer Care.

Months 3-5

Postpaid + family-plan rollout

Add postpaid and family-plan segments with segment-specific intervention designs.

Months 6-7

Win-back + cross-channel expansion

Post-churn win-back agent introduced; rollout to chat, email and proactive SMS interventions.

Governance & controls

How the deployment is governed.

Telecom carries customer-data, CPNI, customer-fairness and increasingly state-level AI consumer-protection governance.

CPNI & customer-data discipline

Customer-proprietary-network-information accessed only for the permitted retention purpose; access logged per CPNI requirements.

Offer fairness

Offers tested for fairness across customer segments and protected attributes; results reviewed quarterly by the carrier's fairness committee.

Customer-conversation guardrails

Agents (as copilots) never pressure customers, never use misleading framing, always preserve clear paths to standard contract terms.

Vulnerable-customer protections

Vulnerable-customer signals trigger enhanced support pathways; offers and retention conversations adjusted accordingly.

Audit-grade trail

Every conversation, every offer, every outcome captured for regulatory and internal audit.

What other enterprises can learn

Three transferable lessons.

Three lessons for other consumer carriers tackling churn with agentic AI.

1

Proactive beats reactive — at the segment level

Reaching customers before they called to cancel was the single biggest lever. Reactive-only programmes leave most of the retention value on the table.

2

Network context made the conversation honest

When the agent (as copilot) surfaced the customer's actual network experience, the retention conversation became substantive rather than defensive. Carriers underestimate how much trust this builds.

3

Offer-engine integration is non-negotiable

Agents that can't personalise the offer become deflection bots. Integration with the carrier's existing offer engine is what made the work move the P&L.

We stopped treating churn as a customer-service problem and started treating it as a customer-trust problem. The platform helped us have honest conversations at scale.
VP Retention & Customer Loyalty, tier-1 US mobile carrier

Reference call available through your xyner account team; high-level case included in a major US telecom-industry analyst report.

Talk to a partner

Could the same outcome work in your environment?

Tell us your sector. A senior xyner partner will walk you through a tailored plan.