industry
Financial Services

Implementing generative AI chatbot for contract analytics

Key results:
  • Saved a potential ~$5M by reducing time spent on manual data interpretation
  • Achieved ~95% accuracy for contract-related queries
Value levers pulled:
  • Cost savings
  • Efficiency improvement
  • Data security

Picture this...

You’re a US-based PE firm with over $100B in AUM. Your portfolio companies deal with a large variety of customer contracts and price variations, and your sales, finance, and legal teams spend an excessive amount of time reviewing and understanding key themes and risks while deriving insights from the contracts.

You turn to Accordion.

We partner with your team to develop a contract analytics tool with OpenAI GPT models that is capable of processing and analyzing the information from customer contracts—enabling users to have interactive conversations based on contract information. We do this by:

  • Developing a user-friendly interface for querying insights from ~1500 complex contracts.
  • Enabling high-precision contract data extraction with Azure Document Intelligence OCR.
  • Securely hosting on Azure with strict data integrity and confidentiality measures.
  • Improving response accuracy by optimizing context window, chunking size, and Dynamic Top-K Tuning based on RAG methodology.
  • Adding user engagement features such as searchable chats, guides, feedback, and review capabilities.

Your value is enhanced.

With your new contract analysis tool, you save a potential ~$5M by reducing time spent on manual data interpretation and scrutiny of intricate clauses, while also bolstering data security through the implementation of ring-facing measures. Additionally, the tool helps you achieve a ~95% accuracy for contract-related queries, streamlining decision-making by providing reliable revenue and churn forecasting.

Going forward, you can now safely harness the power of generative AI in your daily workflows without security and data privacy concerns.

Enhanced value:

You reap multiple benefits, including:

  • Saved a potential ~$5M by reducing time spent on manual data interpretation
  • Achieved ~95% accuracy for contract-related queries