Menu Close
Contact
Information Technology United Kingdom

Driving client's revenue growth with a new product: AI tool that cuts legal documents comparison from months to six weeks

Executive Summary

Challenge: Our client sought to determine if their idea of a new product, an AI tool for automating information scrutiny, could revolutionise the way legal professionals’ handle sensitive disclosure documents.

Approach: With our understanding of the client’s domain, in just three months, we delivered a fully functional demo that could be presented to potential investors. After the MVP’s success, we’ve started working on the full version of the product.

Result: The final software version is soon going to be tested by a group of end users. With the help of our ML solution multi-page, complex legal documents can be compared in just 6 weeks, saving months of manual work of legal aides.

Table of Contents

About the client

Our client is a UK-based startup, founded by a group of disruptors with a proven track record of making global companies’ teams even more productive with end-to-end tech solutions.

Business challenge

Our client had an idea for a new product they wanted to propose to top-tier law firms: an AI-based tool for scrutinising and comparing complex legal documents.

The organisation sought the best ways to prove that their product would work flawlessly in highly regulated and demanding environment. They were also looking for opportunities to showcase the product’s value to future clients and investors at an early development stage.

For that they needed a Proof of Concept (PoC), and they needed to have it delivered fast.

Proof of Concept

At the beginning of our collaboration, we met our client for discovery workshops. We worked together to select the most important functionalities, and ensure we addressed only the key needs of legal professionals right from the start.

Our partnership approach, however, didn’t stop with the workshop. During development, the client contributed additional ideas we were always ready to consult and smoothly incorporate into the PoC, without compromising the agreed deadline.

As a result, in just three months, we developed a fully functional prototype of the solution.

It proves that our Machine Learning solution can streamline the legal aides’ work in comparing complex documents and searching for content that supports claims within them. All this is done with focus on cybersecurity, ensuring that highly sensitive data is fully protected.

Delivering results

After the PoC’s success, our client decided to work with us on the full version of the product.

We’ve significantly expanded the scope of the project, so that the app can handle even more types of extensive legal documents.

We’re working on making the verification process even more efficient and precise. To achieve this, we’re utilising GPT4 and existing algorithms, even from distant fields such as DNA analysis, which we have proven to be highly effective in text structure analysis.

As a result, our model compares content of multi-page, complex legal documents in as little as 15 minutes, reducing time needed to prepare full documentation form months to even six weeks, and making the work of legal aides even more accurate.

The solution is soon going to be tested by a group of end users.

Main benefits of our partnership

  • Accelerated Product Development: we developed a fully-functional demo in just three months, allowing the client to showcase their product at an early stage, accelerating their go-to-market strategy.
  • Market Validation: with successful Proof of Concept and working towards product launch, our client can demonstrate viability and effectiveness of their AI tool.
  • Competitive advantage: the app we developed helps our client build credibility and attract interest from top-tier law firms and investors.
  • Enhanced Product Scope and Capabilities: utilising advanced algorithms and language models to enhance the product’s efficiency, saving months of manual work of legal aides.

Technologies used in the project