Huzzle Labs

Zurich Insurance: AI efforts

Zurich Insurance Group was the standout climber of 2026, rising eight places to #4 in the Evident AI Index for Insurance — the index's single biggest mover. Under a branded AI360 strategy it has gone from scattered pilots to enterprise-wide execution, on a deliberately multi-vendor stack (AWS, Azure, Cytora, Quantexa). Part of InsureBench's overview of AI in insurance; every figure is sourced.

Where Zurich ranks on AI

Zurich jumped from 12th in 2025 to 4th in the 2026 Evident AI Index for Insurance — an eight-place rise that made it the index's biggest mover, behind only Allianz, AXA, and Manulife.1 Evident credited rapid AI-headcount growth and a shift from use cases to disclosed, outcome-bearing deployments.

Zurich was also one of just five insurers accounting for roughly half of the sector's well-documented AI use cases, and one of the few to disclose financial returns from AI in the inaugural 2025 index.2

Evident AI Index 2026
#4of 30
global insurers
Biggest mover — up 8 places from #12 in 2025. Behind Allianz, AXA, Manulife.
Talent
Innovation
Leadership
Transparency
Pillars illustrate Evident's four scoring dimensions; Zurich was the index's top riser.
#4
in the 2026 Evident AI Index, up from #12
10→95%
straight-through processing via Cytora agentic underwriting
£260k
of bogus claims detected per day (Quantexa, Zurich UK, 2024)
~11,000
colleagues enrolled in AI/data upskilling since 2023

The AI build-out, 2023–2026

A multi-cloud foundation, then a branded strategy, an in-house lab, and agentic AI in core underwriting.

Jan 2023

Multi-year AWS migration begins

Zurich starts moving 1,000 applications to AWS public cloud, targeted for 2025.3

Mar 2024

eXchange 2.0 AI platform

A controlled environment to test and launch AI propositions; Zurich says its AI assurance framework already met the EU AI Act.4

Nov 2024

Azure OpenAI in underwriting

Used to process unstructured data (images, emails, multilingual reports); Microsoft cites 200+ AI tools deployed.5

Jun 2025

SageMaker MLOps cuts deploy time

A standardized AWS pipeline cuts average model deployment from ~26 weeks to ~8.6

Oct 2025

Quantexa fraud + Zurich AI Lab

Zurich UK's Quantexa partnership detects ~£260k of bogus claims per day; the Zurich AI Lab launches with ETH Zurich and St. Gallen.7

May 2026

Cytora agentic underwriting scales

Rolled out across 5 countries in 90 days; manual triage cut 75→15 min, straight-through processing from 10% to 95%.8

Jun 2026

Rises to #4 in the Evident AI Index

An eight-place jump — the index's biggest mover.1

Inside the programme

Zurich's "AI360" strategy spans three pillars — AI for Everyone, AI at our Core, AI for the Future — delivered through group platforms and external partners.

Platforms

ZuriChat & Zurich IQ

29 business units

A secure internal gen-AI assistant (ZuriChat) and a multi-cloud LLM gateway (GenAI Lounge), plus the Zurich IQ suite — Guideline IQ (underwriter chatbot) is live in 29 business units.9

Agentic underwriting

Cytora

95% STP

Agentic-AI risk submission processing: straight-through processing from 10% to 95%, digitization accuracy 70–80% → 98%, scaling to 20+ markets.8

Fraud

Quantexa

£260k / day

Real-time decision intelligence across CUE, the Insurance Fraud Register and Companies House data — detecting roughly £260k of bogus claims a day in Zurich UK.7

Research

Zurich AI Lab

Oct 2025

A research collaboration with ETH Zurich and the University of St. Gallen, spanning Zurich, St. Gallen and Singapore — sponsored by CEO Mario Greco.10

"The Zurich AI Lab is our moonshot factory — bringing together business and technology expertise with cutting-edge research to pioneer the next generation of insurance solutions."

Mario Greco, Group CEO, Zurich Insurance — Oct 202510

A multi-vendor, responsibility-first stack

Unlike peers betting on a single model provider, Zurich runs a deliberately multi-cloud, multi-vendor stack — AWS for MLOps, Azure OpenAI for unstructured data, and specialist partners (Cytora, Quantexa) for underwriting and fraud.

It pairs this with a published responsibility framework — the boundary an insurance AI benchmark is built to measure.

Governance at a glance
  • Data & Responsible AI Commitment — five public pledges
  • Four principles — safety, transparency, accountability, reliability
  • EU AI Act — Zurich states its assurance framework was compliant ahead of the law (its own assertion)
  • ~11,000 colleagues in AI/data upskilling since 2023

Why this matters for benchmarking

Zurich is putting agents into the heart of underwriting — Cytora taking straight-through processing to 95% — and AI into claims and fraud. The faster that work is automated, the more it matters whether the models are actually right. That's what InsureBench measures; see the methodology.

Sources

  1. The Insurer — Allianz overtakes AXA; Zurich rises in Evident's AI ranking (Jun 16, 2026). theinsurer.com
  2. Evident Insights — Insurance AI Index 2026 report. evidentinsights.com
  3. Computer Weekly — Zurich plots multiyear move to AWS (Jan 2023). computerweekly.com
  4. Insurance Times — Zurich platform update; EU AI Act readiness (Mar 2024). insurancetimes.co.uk
  5. Microsoft — Zurich Insurance + Azure OpenAI Service (Nov 2024). microsoft.com
  6. AWS — How Zurich is becoming an AI-led insurer with SageMaker (Jun 2025). aws.amazon.com
  7. Zurich UK — Strengthening fraud defence with Quantexa (Oct 17, 2025). zurich.co.uk
  8. Reinsurance News — Zurich scales Cytora AI across underwriting (May 2026). reinsurancene.ws
  9. Zurich — AI at Zurich (AI360, ZuriChat, Zurich IQ). zurich.com
  10. Zurich — Zurich AI Lab launch (Oct 29, 2025). zurich.com
Leaderboard opening 2026. Built by Huzzle Labs.
Get in touch about InsureBench