Igloo's AI push cuts claims response to one minute, lifts sales across Southeast Asia

Igloo's AI push cuts claims to one minute and boosts sales across Southeast Asia

Igloo is putting AI at the core of its business and the early numbers are strong. Claims response times dropped from as long as three days to one minute. Accuracy hit 98%, and 70% of manual claims work is now automated.

On the commercial side, AI-powered product recommendations reduced customer drop-offs by 24.7%. Seven in ten customers pick an AI-suggested plan, and add-on purchases are up 2.2 times. The company reports 200% net dollar retention and is targeting EBITDA break-even by 2026.

What changed

• Claims: Response time down to one minute (from up to three days), 98% adjudication accuracy, 70% reduction in manual work.
• Sales: 24.7% fewer drop-offs, 70% of customers choose AI recommendations, 2.2x increase in add-on attachment.
• Strategy: AI integrated across product management, sales, and claims; goal to be fully AI-native by 2026.
• Financials: 200% net dollar retention; EBITDA break-even expected by 2026.

Why this matters for insurers, MGAs, and distribution partners

• Faster claims improve satisfaction and reduce cost to serve.
• Better recommendations lift conversion, cross-sell, and retention without adding headcount.
• Operational accuracy (98%) lowers leakage risk and shortens cycle times.
• Clear KPI impact (NDR and EBITDA) strengthens the business case for AI investment.

A practical playbook you can use now

• Start with high-volume, rules-heavy workflows: First notice of loss, simple claims, and add-on offers at checkout.
• Deploy triage and recommendation engines: Use propensity models to route claims and surface the next-best offer.
• Keep humans in the loop: Set clear thresholds for auto-approval, escalation, and audit.
• Measure what matters: Response time, adjudication accuracy, manual touch rate, drop-off delta, attach-rate lift, and NDR.
• Align with governance: Document data sources, monitor drift, and apply responsible AI guidelines such as MAS's FEAT principles (learn more).

Benchmarks to watch

• FNOL-to-first-response time near real-time (minutes, not days).
• Adjudication accuracy approaching 98% with ongoing QA.
• Manual handling reduced by ~70% in targeted flows.
• Checkout drop-offs down ~25% with recommendation systems.
• Add-on attach rates growing 2x or better.
• Net dollar retention trending above 150% for embedded/distribution-led models.

Looking to 2026

• Operating model: AI-native processes from product to claims, with clear human oversight.
• Partner integrations: Embedded distribution and real-time data sharing across the value chain.
• Regulatory alignment: Consistent controls, explainability, and monitoring at scale.
• Talent and training: Upskill product, claims, and sales teams to build, test, and operate AI programs (see role-based AI courses).

The takeaway: AI is already moving the needle on speed, accuracy, and revenue. If you can prove similar gains on a narrow slice of your book, you'll have the momentum-and the budget-to scale it across the line.