← All case studies

AIG: AI Underwriting + Claims Assistance at scale

Production GenAI assistance for underwriting and claims workflows, designed for adoption, governance, and measurable operational impact.

At a glance

Challenge
Adopt GenAI in regulated underwriting and claims workflows.
Scope
Enterprise rollout across high-volume underwriting and claims journeys.
Timeline
Multi-quarter phased production delivery.
Measurable outcomes
Cycle-time reduction, higher throughput, and sustained user adoption.
Industry
Insurance
Scope
Enterprise workflows
Focus
Underwriting + Claims enablement
Role
Technical lead (BCG) — day-to-day technology leadership, architecture, delivery execution, stakeholder alignment
Approach
Secure enterprise GenAI patterns, workflow-integrated UX, evaluation loops, and risk-aware rollout

Results

  • Reduced cycle time for key workflows
  • Increased throughput and user productivity
  • Scaled assistance across major underwriting and claims journeys
  • Built foundations for reuse across teams

Context

AIG needed AI assistance embedded directly into real underwriting and claims workflows.

The target was not a demo, but a system teams would trust and use at enterprise scale.

The problem

Enterprise constraints were real: security, privacy, governance, and integration complexity.

Success also required measurable outcomes and broad cross-functional adoption.

  • Balance compliance and velocity.
  • Integrate with existing workflows and systems.
  • Prove operational impact, not just model quality.

The approach

We focused on workflow-first design and tight integration with existing systems.

Delivery used staged rollout and evaluation loops so the team could measure usage, improve quality, and expand coverage without increasing risk.

  • Workflow-integrated UX over standalone copilots.
  • Evaluation metrics tied to real operating outcomes.
  • Phased deployment with governance checkpoints.

What I led

I owned day-to-day technical leadership across architecture, integration, and execution.

I translated business goals into an executable delivery plan and drove cross-functional alignment through production realities.

Results

The program delivered production AI assistance that improved speed and usability in underwriting and claims contexts.

A platform approach also enabled expansion across additional business workflows.

Lessons

Enterprise GenAI succeeds when embedded into core workflows and measured like a product.

Governance and iteration need to be built into delivery from day one.

← All case studies