What challenges did Avantia Group face before AI implementation?

Insurance fraud detection with unstructured data was inefficient, impacting service speed and fairness. A need to improve fraud detection rates and handle claims more efficiently while preserving high service quality for customers.

What was the impact of the AI solution?

Fraud detection rate increased from 2% to 12%, detecting six times more potential fraud indicators. Saved an estimated £1.2 million annually through efficiency and cost avoidance in claims handling. Enhanced customer service by enabling agents to focus more on customers and less on manual fraud detection tasks. Positioned Avantia to scale efficiently while maintaining quality service as customer base grows to target one million insured properties.

Source

GCP

Avantia Group Case Study

Use case typeFraud detectionUpdated Jun 13, 2026

Avantia Group, a UK-based home insurance company, implemented an AI-powered fraud detection and claims analysis platform named 'Holmes' using Google Cloud tools including Vertex AI, BigQuery, and Gemini models. Facing challenges in detecting insurance claims fraud while maintaining fast and fair customer service, Avantia used generative AI to process unstructured data such as call recordings, documents, and photos. Holmes analyzes thousands of data points per claim, identifies coverage, flags inconsistencies and potential fraud, and provides claims handlers with a clear summary swiftly. The system increased fraud detection rates from 2% to 12%, detecting six times more potential fraud indicators and enabling faster, cost-saving claims handling. Avantia projects savings of approximately £1.2 million annually from efficiency gains and plans to further integrate AI deeply into claims operations for scalable, customer-focused service enhancement.

Organization: Avantia Group
Industry: Insurance
Location: United Kingdom

Reported outcomes

+12%

timeTime & speed

+2%time

Strategic outcomes

New product / capabilityBuilt AI-powered fraud detection platformNew product / capabilityEnabled generative AI claims analysisCustomer experience & trustImproved customer service qualityScale & capacityPositioned for scalable growth

Catalog median for time & speed deployments: +80% across 203 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 3

1Fraud Detection
2Claims Management
3Generative AI

Insurance fraud detection with unstructured data was inefficient, impacting service speed and fairness.
A need to improve fraud detection rates and handle claims more efficiently while preserving high service quality for customers.

Developed 'Holmes' platform using Google Cloud's Vertex AI for AI orchestration, BigQuery for data integration, and Gemini generative AI models for analytic and summarization tasks.
Adapted existing ML systems to safely deploy generative AI for claims fraud detection and analysis, supporting unstructured data.
Built a unified analysis context by consolidating diverse data sources into BigQuery and layered AI reasoning with Vertex AI.

Technologies

Google Cloud Vertex AI BigQuery Gemini 2.5 Flash Gemini 2.5 Pro

Fraud detection rate increased from 2% to 12%, detecting six times more potential fraud indicators.
Saved an estimated £1.2 million annually through efficiency and cost avoidance in claims handling.
Enhanced customer service by enabling agents to focus more on customers and less on manual fraud detection tasks.
Positioned Avantia to scale efficiently while maintaining quality service as customer base grows to target one million insured properties.

Architecture

Holmes integrates data from multiple internal systems into BigQuery for structured analysis context, orchestrates AI workflows with Vertex AI, uses Gemini 2.5 Flash for summary and scoring, and Gemini 2.5 Pro for core analytical tasks.

Sources & evidence1

Groundedness: 5/5

Primary source

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

Explore related AI use cases

Customer evidence