In a very creative approach to using genAI for contract review, Syntracts – co-founded by Uber and Latham & Watkins alums – doesn’t need prompts, and instead uses orchestrated small language models (SLMs) trained on synthetic data, each focused on different aspects of a contract to provide lawyers with very high accuracy – and without hallucinations.
Syntracts is also different in another way: it operates on-prem and thus doesn’t need to get into the cloud at all, thus providing law firms and inhouse teams with, in theory, much better security.
In short, this is a real departure from how many genAI tools are deployed, and especially for contract review.
As the US-based company explained: ‘Unlike legal AI tools that rely on third-party APIs, cloud-hosted LLMs, or brittle prompt engineering, Syntracts takes a fundamentally different approach.
‘Its proprietary, open-source models are trained using synthetic legal data, which is generated from each customer’s own legal documents. This results in highly accurate, structured outputs built for real-world legal workflows, all while keeping sensitive information entirely on-prem.’
Doug Bemis, co-founder of Syntracts and former CTO at Uber AI Labs, told Artificial Lawyer: ‘You can’t turn a generalist LLM into a legal expert just by prompting. LLMs should be perfect for law, but they haven’t delivered. Legal work demands clarity, consistency, and control – none of which prompting can guarantee. Syntracts is for firms that need trustworthy performance from day one, with models trained on firm-specific synthetic data and deployed fully on-prem to produce structured, database-ready outputs.’
He added: ‘LLMs are trained to sound right, not be right. People try and prompt LLMs into submission. But, we use a different approach.
‘We create a foundational data layer that can handle legal language. We divide contracts into small pieces, then use synthetic data to fine tune small language models. We use multiple SLMs and then orchestrate – but we don’t prompt and don’t need lawyers to handle prompts.’
Meanwhile, fellow co-founder Christopher Martin, who was a senior lawyer at Latham & Watkins and served as Manager of Emerging Technology, told this site: ‘I have been disappointed in LLMs’ ability to handle real legal workflows. You need a data foundation to do interesting things and in Big Law even a small mistake can have a big impact.’
Martin explained that Bemis was an old friend and they had wanted to work together in the past, but now the time was right and together they have developed an approach to achieve what they wanted.
What they have built can deliver:
‘Reduced contract profiling time by at least 80%, shrinking multi-hour reviews to under 30 minutes, and, in API-integrated workflows, near zero.
Outputs are structured to match firm-defined schemas, requiring no post-processing or human-in-the-loop review.
Total Privacy: All models run on-prem; no data leaves the client’s systems
Structured Intelligence: Outputs are formatted for direct integration into legal workflows.
Scale Without Headcount: Automates extraction at scale without prompting or human review.
Custom-Trained for Each Client:Synthetic legal data models are built from each firm’s own documents.’
As noted, this is a long way from just tapping one of OpenAI’s models to give a contract a once-over.
Martin added: ‘Prompt-based systems don’t scale – you shouldn’t need more people to make AI work. We built Syntracts to meet the strictest data and compliance standards while delivering real results: structured, high-accuracy outputs based on how lawyers actually read contracts.’
The company also noted they’ve just signed a deal with a Top 25 AmLaw firm and are also in the A&O Shearman’s Fuse Incubator.
You can find more about Syntracts here.
Discover more from Artificial Lawyer
Subscribe to get the latest posts sent to your email.