Initial Situation
With the entry into force of the Digital Operational Resilience Act (DORA) , financial institutions and their IT service providers faced an enormous challenge: thousands of existing and new contracts had to be reviewed for compliance with strict new EU regulations. A leading consulting firm recognized early on that manually reviewing these extensive contract portfolios would be not only extremely time-consuming but also prone to error.
Instead of investing hundreds of hours in manual reconciliation processes, the company saw a strategic opportunity to scale this process using modern AI technology. The goal was to provide consultants with a tool that automates the tedious search for relevant clauses and delivers a rapid initial compliance assessment. The solution needed to process complex contract packages and precisely locate specific DORA requirements within the documents.
Our Solution
AMAI developed a specialized Retrieval-Augmented Generation (RAG) solution that uses Large Language Models to intelligently audit contract documents against a predefined DORA checklist. The system acts as an intelligent assistant for compliance experts by automatically identifying and pre-evaluating relevant text passages.
The technical implementation includes the following core components:
- Semantic contract analysis using RAG: The system breaks down extensive contract portfolios into semantic units and stores them in a vector database. Through advanced retrieval algorithms, the solution can identify the most semantically relevant contract clauses for every point on the DORA checklist, even when the wording differs significantly from the legal text. This enables a context-aware review that goes far beyond simple keyword searches.
- Automated compliance pre-assessment: A powerful Large Language Model analyzes the identified text passages and evaluates whether the contractual provisions meet specific DORA requirements. The system provides not only a yes/no assessment but also links directly to the supporting text passage in the original document. This creates transparency and allows human experts to quickly validate the results.
- Scalable cloud architecture on Azure: To process sensitive contract data securely and efficiently, the solution was implemented as a cloud-native application on Microsoft Azure. The architecture ensures data security and scalability, allowing large volumes of contract documents to be processed in parallel. Integration into existing workflows is achieved via an intuitive user interface for consultants.
A particular challenge was the high variance in the legal language of the contracts. Through targeted prompt engineering and the optimization of the retrieval strategy, the system was fine-tuned to reliably recognize and correctly categorize legal nuances.
Results & Business Impact
The introduction of the AI-powered audit assistant led to a massive acceleration of compliance processes. Where an expert previously needed several hours for the detailed review of an extensive contract portfolio, the system now delivers a qualified analysis within 1-2 minutes.
The solution achieves high accuracy in identifying relevant clauses, which drastically reduces the manual search effort. Experts no longer need to read hundreds of pages; they simply validate the matches suggested by the system. This allows the consulting firm to offer DORA audits much more efficiently and cost-effectively, while simultaneously increasing the consistency of the reviews. The proof of concept successfully demonstrated how generative AI can resolve regulatory bottlenecks and relieve experts of repetitive tasks.
