Home/Case studies

Case study · Agentic Process Outsourcing · LATAM

Rebuilding A&H and Auto operations around AI agents

How a leading global insurer in Mexico designed an AI-native managed-service layer across the policy lifecycle without replacing its existing core.

200+active products
10connected operating modules in the scoped design
98%+contracted automated-reconciliation target
13 monthsplanned phased deployment

Case studies

From operating challenge to controlled execution

01

The operating challenge

Growth was outpacing an operating model built around spreadsheets, fragmented systems and manual handoffs.

  • The underwriting team processed roughly 80 quotes per month, with product rules managed inconsistently in Excel.
  • Only 10% of submissions closed on the first offer, limiting speed and commercial conversion.
  • Collections, reconciliation and CFDI processes lacked a centralized automated workflow.
  • Fragmented data restricted real-time portfolio and P&L visibility and slowed product launches.
02

The operating model

Gangkhar designed a managed-service layer around the insurer’s existing core, combining Sherpa+ infrastructure with Sherpa+Lens intelligence and governed human oversight.

  • Ten connected modules cover product configuration, underwriting, policy administration, billing, commissions, servicing, sponsor onboarding, interfaces, notifications and fraud detection.
  • Predictive models surface cancellation risk and commercial opportunities while governed routines automate bounded actions.
  • Material decisions remain under human control; every workflow is measurable, auditable and explainable.
03

Designed service levels

The programme was designed around measurable operating targets that can be governed contractually and monitored in production.

  • Individual issuance in five minutes or less; bulk issuance of up to 1,000 policies in 30 minutes or less.
  • Payment confirmation within ten minutes and automated reconciliation of at least 98%.
  • CFDI generation within 15 minutes after payment, 99.9% platform availability and critical support response within one hour.
04

Phased deployment

A roughly 13-month rollout sequences discovery, integration, migration, controlled launches and operational stabilization to reduce transition risk.

  • The architecture avoids a core replacement and keeps the programme connected to the insurer’s existing technology estate.
  • The intended impact is higher first-offer conversion, shorter commission cycles, real-time P&L visibility and faster time to market.
Back to In practice

Discuss a similar deployment

Could this operating model work for your business?

Map the product, participants, controls and first deployment with the Gangkhar team.
Request a meeting