Same-day claims, 1 million app users, smarter sales calls: AI redefining Singapore's banks and insurers

AI that actually moves the needle for insurers: faster claims, smarter sales, leaner ops

Real results are coming out of AI pilots in APAC insurance and banking. Etiqa is releasing 60% of travel claims on the same day and has grown its Etiqa Plus app to 1 million onboarded customers. Banks and tech providers are also proving AI use cases in voice intent analytics, content summarisation, and gen AI assistants that cut costs and raise satisfaction.

Claims assessment: from weeks to same day at Etiqa

By integrating AI into travel claim assessment, Etiqa now clears 60% of claims within a day. According to CTO Dennis Liu, customers used to wait weeks or months for outcomes. Faster decisions lifted mobile adoption too, with the Etiqa Plus app crossing the 1 million customer mark.

• Target quick wins: straight-through processing for low-risk, well-documented claims.
• Measure the basics: same-day release rate, average time-to-decision, resubmission rate, and NPS after payout.
• Feed outcomes back into the model to tighten fraud flags and reduce false positives.

Voice intent analytics: fewer calls, better timing

DataKlout Ai Solutions analyses tone during live calls to identify intent and interest. Founder and CEO Puneet Sinha says this lets teams prioritise real opportunities, trim unnecessary follow-ups, and apply the same approach in collections, customer service, and fraud monitoring.

• Run a pilot on one call queue. Compare conversion, average handle time, and call attempts per sale.
• Get consent, update disclosures, and run bias checks on outcomes by segment.
• Use insights to adjust scripts and call timing rather than pushing volume.

From long reports to bite-sized insights at Maybank

Maybank Singapore uses AI to turn CIO quarterly outlooks into short, clear stories that everyday customers can read. Helena Ooi reports higher engagement across social channels and more accessible guidance without waiting for a relationship manager.

• Apply the same approach to policy summaries, coverage explainers, and renewal nudges.
• Tag content by product, risk, and life event to fuel personalised notifications.
• Keep compliance in the loop and version-control every output.

Gen AI assistants: 50%+ agent workload reduction

Nucleus Software is handling inbound and outbound queries for clients using a gen AI assistant. Pradeep Malik notes a 50%-60% reduction in agent workload, 40%-50% lower operating costs, and a higher customer happiness index when routine queries are automated and complex cases escalate to humans.

• Set clear deflection targets and strict escalation rules.
• Ground the assistant in approved knowledge bases; log every response.
• Track containment rate, first contact resolution, and post-contact CSAT.

Risk, compliance, and controls

Voice analytics and automated decisions demand strong guardrails. Align with established frameworks and document every model's purpose, data lineage, monitoring plan, and human oversight.

• Review AI fairness and data use principles from MAS FEAT: MAS FEAT principles.
• Adopt an AI risk framework with controls and testing: NIST AI Risk Management Framework.
• Secure consent for voice analytics and maintain transparent opt-out paths.
• Run periodic bias, drift, and false-positive reviews; keep audit logs.

90-day action plan for insurance leaders

• Pick one claim type and deploy AI triage to lift same-day releases.
• Pilot tonal intent on a single sales or collections line; A/B test outcomes.
• Convert one long advisory into micro-stories; push via app and email.
• Launch a gen AI assistant for FAQs with human handoff; measure containment.
• Baseline KPIs now; report gains weekly with a clear rollback path if needed.
• Upskill your team on prompt design, evaluation, and guardrails: AI courses by job.

The thread across these cases is simple: focus on one process, define the metric that matters, and ship. Speed, clarity, and control beat big-bang promises every time.