
Meet Percy: the legal-eagle AI platform giving Simmons & Simmons the edge
Simmons & Simmons, a global law firm with 2,600 professionals across 21 offices worldwide, has built its own AI powered platform – Percy AI – that is not only improving productivity and work quality, but also market competitiveness.
The legal profession is awash with documents – long, complex documents – and lots of them, from commercial contracts to witness statements, court filings to client correspondence.
They all need to be analysed, understood, compared and contrasted.
Law firms’ hard-won reputations are built on executing this work to the highest possible standard, to tight deadlines; their clients expect no less.
So Generative AI (GenAI) is proving to be a valuable assistant for those legal firms with the foresight and ambition to make the most of it.
Simmons & Simmons is one such firm. It has built an AI-powered platform called Percy AI that is empowering its employees and enabling the firm to differentiate itself from the competition.
Catalina Perdomo, a senior legal engineer at the firm, gave Microsoft a demonstration of what the latest version of Percy AI can do.
She fed it a chunky 550-page transcript of a legal case and prompted it to summarise the key themes and arguments each party was making, presented in tabular format, quoting references for each section – a crucial facility for lawyers who live or die by facts and evidence.
A process that would have usually taken a lawyer many hours, took just a matter of seconds.

She then showed how the user could interrogate such documents in any number of ways that would not only save time but also reveal insights that could previously only be unearthed through a laborious manual process.
She demonstrated how Percy AI can compare and contrast witness statements and identify inconsistencies in their testimony, for example, and can also compare how different businesses within an industry are reporting the risks they face.
And it can handle all forms of correspondence and documents, whether PDFs, images, emails – even hand-written notes.
“Although Percy is undoubtedly saving us time and improving productivity, it is primarily benefiting us by improving the quality of our work and our competitiveness in the marketplace,” explains Drew Winlaw, Partner and Global Lead of Gen AI & LLM at Simmons & Simmons.
Percy’s genesis
The firm built the first iteration of Percy AI in 2023 on Microsoft Azure OpenAI, initially as a secure environment for experimenting with Gen AI, says Winlaw.
“We wanted everyone in the firm to be able to use and benefit from it, not just lawyers,” he says.
Factual accuracy is mission-critical for many tasks within the legal profession, so the team built Percy with a sophisticated Retrieval-Augmented Generation (RAG) capability.
Percy deploys an enhanced ‘inference engine’ [a core component of AI systems that applies logical rules to data] over specified files that are uploaded by each user, depending on their use case. The use of RAG helps Percy output answers that are contextually relevant, thereby improving accuracy and user confidence.
The firm also appointed David Huston, an LLM [Large Language Model] Programme Lead, to focus their efforts around Generative AI, and 71 AI Champions and Coordinators who consulted employees on how they felt Generative AI could be used most effectively.
They came up with more than 270 use cases, and now have the capability to deliver more than 50% of these with their current AI tools.

“The AI Champions network was designed to help with education and adoption, and to create a global network of experts who would aid with adoption and help build the next generation of products and services,” says Huston.
“This led to us leading on a major re-development of Percy into a platform for GenAI.”
Usage statistics and continual additions to a dedicated “prompt library” demonstrate increases in both the application and the complexity of use cases, and these both saw a big increase with the global launch of Percy 2.0 in February 2025, says Winlaw.
This version included document upload and inference capabilities, and enhanced the underlying ‘engine’ to enable everyone to use it for general productivity tasks as well as for more specialised uses.
Most employees at Simmons are now turning to Percy for significant tasks each week, and many are using it every day, Winlaw adds.
Percy AI’s ability to accelerate the completion of routine tasks, such as document drafting and data analysis, is freeing up lawyers to focus on more complex and strategic aspects of their work, he says.
”There are huge opportunities to streamline our operations and improve the quality of our work,” Winlaw says. “We now have the tools to deal much more effectively with the unstructured data that, as lawyers, we need to bring to support the task at hand.”
Pro bono Percy
Percy AI is not only driving significant improvements in productivity and innovation but is also having a positive social impact, the firm says.
Free ‘pro bono’ legal services are a unique professional contribution lawyers can make to promote access to justice, and Simmons dedicates substantial time and resources to the cause, supporting individuals, charities, NGOs [Non Governmental Organisations], and social enterprises.
Percy has already begun to supercharge the firm’s pro bono activities, for instance, playing a crucial role in projects requiring translation services. This is helping to speed up the progression of cases. By expediting routine tasks, it has allowed its teams to concentrate on delivering an even better service to pro bono clients.
Percy’s future
Looking forward to Percy 3.0, the firm is planning to develop the solution into a “fully-featured Generative AI platform, including a space where lawyers can collaborate, and to be one of the places where people can ask agents to perform complex tasks,” says Winlaw.
Huston agrees: “We’re collaborating with Microsoft to create a legal inference engine for much more complicated work for specialist divisions. The LLM-driven engine that is unique to Percy produces significantly higher outputs than a standard LLM and it is being constantly upgraded.
“We look forward to version 3 coming out soon!”
From Winlaw’s perspective, the firm began by enhancing AI literacy across its workforce and is now working on developing innovative services to generate new income streams, anticipating changes to their business as their clients adopt AI.
“This truly is a watershed moment for AI and automation in the legal sector,” he concludes.