What challenges did AXA Switzerland face before AI implementation?

Complex queries could take an entire weekend in the legacy environment. Siloed systems prevented real-time decision making across CRM, ERP, ecommerce, and operational data. Human specialists spent about 25 minutes reading complex medical documents to extract key information. The company needed automated triage to determine which claims required expert evaluation.

What was the impact of the AI solution?

Reduced query and processing times by over 95%, from days to minutes or seconds. Turned a 25-minute medical-document review task into a seconds-long AI workflow. Uncovered a 120,000 Swiss franc false claim in the first week of deployment of a fraud detection ruleset. Improved damage expense management, pricing structures, and marketing through data-based insights. The company described the Smart Data projects as among the most profitable initiatives in AXA Switzerland's history.

Source

GCP

AXA Switzerland uses BigQuery and Vertex AI/Gemini for AI-assisted claims triage and faster document extraction

Use case typeClaims automationUpdated Jun 13, 2026

AXA Switzerland is the country's leading insurer, providing car, liability, health, life insurance, and pension solutions to many Swiss companies. The organization was constrained by legacy infrastructure, slow complex queries, and manual reading of medical documents, which limited real-time decision making and efficient claims handling. By consolidating enterprise data in BigQuery and using Vertex AI and Gemini for medical document extraction and claims triage, AXA Switzerland modernized its insurance operations on Google Cloud.

Organization: AXA Switzerland
Industry: Insurance
Location: Switzerland
Published: January 2024

Reported outcomes

−95%

timeTime & speed

Strategic outcomes

New product / capabilityAutomated claims triage and document extractionRisk & complianceImproved fraud detection and false-claim identificationBetter decisions & insightImproved pricing and expense decisionsSpeed & agilityEnabled faster claims and query handling

Catalog median for time & speed deployments: −60% across 727 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 4

1Claims triage automation
2Document understanding
3Fraud detection

Complex queries could take an entire weekend in the legacy environment.
Siloed systems prevented real-time decision making across CRM, ERP, ecommerce, and operational data.
Human specialists spent about 25 minutes reading complex medical documents to extract key information.
The company needed automated triage to determine which claims required expert evaluation.

AXA Switzerland migrated and unified enterprise data with BigQuery.
It used Vertex AI and Gemini for AI-driven text understanding and medical document extraction in seconds.
The company implemented automated claims triage to route only the right cases to experts and applied fraud-detection rulesets to identify false claims.

Technologies

BigQuery Vertex AI Gemini Cloud Run Dialogflow

Reduced query and processing times by over 95%, from days to minutes or seconds.
Turned a 25-minute medical-document review task into a seconds-long AI workflow.
Uncovered a 120,000 Swiss franc false claim in the first week of deployment of a fraud detection ruleset.
Improved damage expense management, pricing structures, and marketing through data-based insights.
The company described the Smart Data projects as among the most profitable initiatives in AXA Switzerland's history.

Architecture

AXA Switzerland unified enterprise data in BigQuery and used Vertex AI and Gemini for document understanding and claims triage. The solution operates in a Google Cloud environment and includes automated fraud-detection rulesets and natural-language interactions over the enterprise data foundation.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jan 1, 2024Publisher: Google CloudEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence