In mid-2024, Close Brothers encountered a sophisticated fraud. It looked like a dream deal: financing a £135,000 luxury car with a £50,000 deposit. Four months later, the payments stopped. The £50,000 was completely fake. They ended up losing a substantial amount of money. Their manual document review process was taking loads of time and creating workflow bottlenecks but, in the end, it was also ineffective.
How Close Brothers saved £800K by stopping document fraud
Client NameClose Brothers
Client CountryCzech Republic
- Client typeEnterprise
- IndustryFinancial Services, Markets & Insurance
- Application areasFinance & Accounting
- AI technologiesExplainable & Trustworthy AI, Large Language Models (LLMs), Machine Learning, MLOps & Model Monitoring, Natural Language Processing (NLP)
- Business impactsData-Driven Decisions & Planning, Operational Efficiency & Cost Savings
- Data typesDocuments / Semi-structured Data, Image Data
- Delivery modelsProduct / Licensed Software
- DeploymentsCloud
- Key capabilitiesAnomaly, Risk & Fraud Detection, Decision Support & Augmented Analytics
- Project stagesScaling / Expanded Implementation
- Solution formsAPI / Micro-service Interface, Web Portal / Dashboard
Solution Description
Problem description
Solution
Close Brothers Motor Finance embarked on a trial of Resistant AI’s Document Forensics, which immediately spotted 18 additional cases of fraud. Resistant’s automated document fraud check takes only 12 seconds, replacing the previous 15-minute manual fraud assessment process. Ten months after implementation, Close Brothers estimates that they have prevented £800,000 worth of fraud. The dramatic transformation and success in motor finance has driven expansion plans across Close Brothers’ wider business portfolio, including premium finance, savings, invoice finance and asset finance divisions.
Main Users of the Solution
Loan underwriters, fraud analysts
Project timeframe (months)
1
Additional services
- AI strategy and roadmap
- Identification and prioritization of suitable use-cases
- Data collection and pre-processing
- Compliance/regulatory support
Implementation
Project Owner on the Client's Side
Business Unit Manager
Participation on the Client's Side
- End users
Form of Supplier Involvement
Joint implementation with the client
Operation and Maintenance
Operational Model
Customer Success Team
Needed Competencies on the Client's Side
An understanding of Resistant AI indicators
Other Resources or Infrastructure
None
Impact and Results
Qualitative Benefits
The underwriting team’s efficiency doubled and they were able to focus on genuine risk evaluation rather than time-consuming document authentication.
Quantitative Results
£800,000 saved in 8 months 2x faster review times 22x return on investment
Client Feedback
“We’re talking significant savings and, more importantly, it’s got people talking about fraud. People are finding ways to put better controls in place.”
Lessons Learned and Recommendations
Key Success Factors
“The aftercare support from Resistant AI has been unbelievable. We just haven’t had that level of support from others previously.”
Recommendation for Others
Document fraud detection is becoming more and more sophisticated, so companies should resist the temptation to get a document automation solution that doesn’t put fraud at the center.
Promotion
- CompanyResistant AI
- ContactJoe Lemonnier
- Emailjoe.lemonnier@resistant.ai
- Websitehttps://www.resistant.ai
- AddressLazarská 13/8, 120 00 Praha