AI Assistant for Contract Review under DORA Regulation
SUMMARY
- Client
- Berenberg
- Tech stack
- Angular, Azure, GPT4, Helm Charts, Python, Spring Boot, Typescript
Berenberg Bank was looking for a more efficient way to review ICT third-party contracts in line with DORA requirements. They needed a solution to automate reviews, reduce costs and errors, and speed up contract adjustments. Profinit delivered an AI assistant that streamlined the process and ensured regulatory compliance.
Results
Project Background
Berenberg, one of the oldest and one of Europe’s leading privately owned banks, sought a faster and more reliable way to review its ICT third-party service provider contracts against the requirements of the Digital Operational Resilience Act (DORA).
Challenge
An AI solution was needed to enable fast, automatic, and reliable verification of hundreds of ICT contracts. However, achieving this required overcoming several key challenges:
- Complex contracts: IT and outsourcing agreements often contain specialised or ambiguous language
- Time constraints: Legal teams were spending excessive hours on manual reviews, driving up costs and delaying decision-making
- Risk of non-compliance: Missing or misclassified clauses could lead to non-compliance with DORA, exposing the bank to regulatory actions
- Multiple addenda: Contracts often include separate addenda, further increasing complexity
Business Needs
With DORA compliance becoming mandatory by January 2025, Berenberg needed a solution that would:
- Automate contract reviews: Systematically identify discrepancies between contract terms and a DORA-compliant model
- Ensure DORA alignment: Minimise manual review time while improving accuracy
- Reduce costs and errors: Extract key parameters for building and maintaining the DORA-mandated Register of Information.
- Support the Register of Information (ROI)
- Accelerate contract renegotiations with ICT with ICT service providers by generating addenda proposals, ensuring faster agreement on standard and custom clauses
Solution & Results
By rapidly analysing Word, PDF, or scanned documents, the system completes reviews within minutes—driving a 70% improvement in processing speed. Moreover, 80% of manual effort has been eliminated, allowing legal and procurement teams to focus on higher-value tasks.
During validation, Berenberg’s team confirmed a 90–95% accuracy rate in identifying and comparing elevant clauses, ensuring consistent and reliable reviews.
To meet DORA’s requirements for transparency, relevant contract parameters are automatically extracted for the mandated Register of Information (ROI), streamlining ongoing compliance and reporting processes.
By automatically generating addenda, the system
reduced negotiation timelines with ICT contract
partners from 2–3 weeks to minutes.
The AI-driven contract review is supported by an intuitive user interface and a clear, systematic presentation of results, enabling both the legal and procurement departments to efficiently assess contracts.
HEAR FROM THE CLIENT
The AI assistant for contract review has significantly accelerated our approach to DORA compliance. Its automated clause-by-clause analysis quickly identifies discrepancies, suggests relevant amendments, automatically generates addenda for renegotiation with our ICT third-party service providers and extracts relevant contract data for reporting to the information register—all with remarkable accuracy. By reducing manual effort by 80%, we have significantly reduced both, our costs and resource bottlenecks, allowing our team to focus on higher value tasks.
