Canadian insurers step up GenAI adoption: from pilot to production
Canadian banks and insurers are moving GenAI into core operations. The latest survey from KPMG Canada shows 90%+ of leaders view GenAI as critical to competitive advantage, and 86% are investing despite economic pressure. Insurers are ahead of banks on integration, with more reporting "advanced" adoption across key workflows.
The shift isn't theory. Better models, clear productivity gains, and higher customer expectations are pulling GenAI into day-to-day work. Leaders report impact across market analysis, pricing and underwriting, fraud detection, customer engagement, and back-office operations.
Why this matters for insurers
Insurers have the data, the digital customer base, and the operational leverage to turn GenAI into measurable results. Think faster underwriting, cleaner risk signals, tighter fraud controls, and lower loss adjustment expense. The lesson from early movers: treat GenAI like infrastructure, not an experiment.
As one industry leader put it, "Generative AI is now a strategic imperative." That mindset shift is what's separating pilots from production.
Where GenAI is delivering value now
- Market analysis: Processing large volumes of market and competitor data to spot emerging risks and growth opportunities.
- Personalization: Real-time, context-aware recommendations (e.g., coverage options in digital flows) that lift conversion and retention.
- Customer engagement: Conversational AI handling routine inquiries, triaging complex cases, and guiding agents with next best actions.
- Risk and fraud: Layering GenAI on existing engines to flag anomalies, improve alerts, and support more granular stress testing.
- Operations: Document summarization, policy and contract review, claims assessment support, and internal knowledge search.
Leaders in banking are seeing a step-change in productivity and deeper engagement. Insurers can push even further by wiring GenAI into underwriting and claims decisioning with human oversight.
Insurers ahead on integration
Around 30% of insurers report advanced adoption-GenAI fully integrated across core workflows. Banks tend to be in partial rollout, focused on productivity and sales first. Sector strengths include rich (and increasingly structured) datasets and a culture that supports practical innovation.
Consumer behavior helps. Strong digital uptake in direct-to-consumer insurance creates a ready channel for GenAI-enabled experiences. The caveat: strategies must match your risk appetite and maturity, not your peer's press release.
Pressure points you need to fix
- Data quality: ~30% cite fragmented or poorly governed data undermining outputs. No clean data, no credible AI.
- Cyber and privacy: 95% worry about breaches or misuse; 60% report a past incident. Secure architectures and strict access controls are non-negotiable.
- Governance: Many frameworks are immature-unclear ownership, slow decisions. Align to model risk and AI controls early. See OSFI's model risk guidance for direction: OSFI Guideline E-23.
- Platform choices and talent: Nearly half are investing in talent; 44% in advanced platforms. Roughly 60% are modernizing core IT to embed GenAI into workflows.
- ROI timing: Most expect returns within 1-3 years, tracking productivity and adoption before revenue.
A 90-day plan for insurance leaders
- Pick 2-3 workflows with clear payback: Claims FNOL triage, subrogation prediction, underwriting assist for small commercial, or fraud alert review.
- Run data sprints: Define critical datasets, fix lineage and quality rules, and lock schemas. Create feedback loops from users back to data owners.
- Stand up guardrails: RACI for model ownership, approval gates, prompt/content policies, red-teaming, and incident response. Align with OSFI expectations.
- Secure by default: Private model endpoints, secrets management, PII minimization, DLP, and usage monitoring. Log everything for audit.
- Human-in-the-loop: Escalation paths, rationale capture, and audit trails for any decision support. Ship prompt libraries and checklists to frontline teams.
- Measure weekly: Baseline handle time, LAE, quote-to-bind, NPS, STP rates. Track adoption and time saved per task, not just outputs.
Metrics that predict ROI
- Agent and adjuster time saved per case
- Reduction in manual reviews and rework
- Underwriting cycle time and quote-to-bind lift
- Fraud hit rate improvement and false positives down
- Loss ratio movement in pilot cohorts
- Customer containment and CSAT for AI-handled intents
Build vs. buy, without losing your edge
Use advanced platforms for orchestration, security, and monitoring. Keep proprietary data assets, underwriting rules, and risk features in your control. Layer GenAI onto existing risk engines instead of rebuilding everything from scratch.
If you need a deeper library of use cases and playbooks, explore AI for Insurance.
Bottom line
GenAI has moved from interesting to essential. The winners are wiring it into underwriting, claims, fraud, and service-then proving value with clean data, tight controls, and weekly metrics. Start small, move fast with guardrails, and scale what works.
Your membership also unlocks:
