Conning's Mazumdar: AI Poised to Fill Insurance Talent Gap as Baby Boomers Retire

Conning's Mazumdar: AI set to reshape insurance workforce amid looming talent deficit

Baby boomers are retiring. In insurance, that means a wave of expertise leaving the building. Manu Mazumdar, head of data analytics and insurance technology at Conning, says artificial intelligence will be critical to fill the gaps and keep operations moving.

This is less about hype and more about throughput. You need the same-or better-service levels with fewer tenured specialists. AI gives you leverage where the labor market won't.

Where AI will close gaps first

• Underwriting triage: Intake, risk summarization, and appetite matching to reduce cycle time and improve hit ratios.
• Claims intake and routing: FNOL capture, document extraction, and next-best action to lower leakage and LAE.
• Pricing and reserving support: Data prep, scenario testing, and narrative explanations for actuaries and product teams.
• Operations: Endorsement processing, reconciliations, compliance checks, and customer email/chat handling.
• Distribution enablement: Producer support, quote comparison, and coverage summarization for faster binds.

What "good" looks like

• Data foundation: Clean policy, claims, billing, and document data; clear lineage; controlled access.
• Human-in-the-loop: AI drafts, humans decide-especially for material risk and customer outcomes.
• Model governance: Versioning, testing, monitoring, and documented decision boundaries.
• Change management: Work instructions updated, controls embedded, training delivered before go-live.

90-day action plan

• Weeks 1-2: Identify workflows most exposed to retirements (underwriting support, complex claims, compliance).
• Weeks 3-4: Baseline metrics: cycle time, touch count, LAE, leakage, rework, service levels.
• Weeks 5-8: Pilot two use cases end-to-end (e.g., claims document extraction and UW triage) with embedded controls.
• Weeks 9-12: Train users, finalize SOPs, track lift, and prepare scale plan to adjacent lines or regions.

Roles that will grow

• AI product owners: Business-first leaders who tie models to outcomes and controls.
• Claims and UW analysts with AI skills: Configure prompts, review outputs, and manage exceptions.
• Data stewards: Keep inputs trusted and auditable.
• Model risk managers: Validate, stress test, and monitor performance and fairness.
• Knowledge engineers: Capture retiring experts' playbooks into reusable guidelines and retrieval systems.

Risk, compliance, and audit

Define where AI may recommend versus decide. Log prompts, sources, and outputs for audit. Test for bias, hallucination, and drift, and keep a clear override path.

Align with emerging principles so regulators and auditors see discipline, not experiments.

• NAIC AI Principles

Core metrics to track

• Efficiency: Cycle time, touches per transaction, queue age, intake throughput.
• Quality: Rework, error rate, leakage, indemnity accuracy, reserve adequacy signals.
• Growth: Quote-to-bind, producer response time, coverage accuracy.
• People: Case load per FTE, engagement, time-to-proficiency for new hires.

Build skills now

Upskill current teams on AI-assisted underwriting, claims, and model oversight. New tools matter, but trained staff turn them into results.

For a structured path by role, see curated options at Complete AI Training - Courses by Job.

The practical takeaway

AI won't replace your people. Teams that pair expert judgment with AI assistance will outperform, especially as retirements accelerate.

Pick high-friction workflows, put guardrails in place, and measure lift. The firms that move now will keep service levels high while the hiring market tightens.