Founders Everywhere: Doug Bemis
Doug Bemis is the co-founder of Syntracts, an AI-powered contract intelligence platform that delivers unparalleled accuracy, customization, and privacy by fine-tuning small language models.
Welcome to Founders Everywhere, where we highlight the incredible people behind the companies we’ve backed at Everywhere Ventures, a global pre-seed fund supported by a community of 500 founders and operators.
In the legal world, professionals face the relentless challenge of interpreting complex contracts with absolute precision and safeguarding sensitive client data. Traditional tools often rely on manual review or prompt-based AI that falls short in consistency, depth, and scalability. Prompting is often like squeezing a balloon: fixing one part of the output often causes new errors to pop up elsewhere, making reliable results elusive. Syntracts is an AI-powered contract intelligence platform that goes far beyond keyword searches or generic LLM outputs, delivering tailored insights and deep understanding of nuanced legal language. Built on a proprietary stack of small, open-source LLMs and synthetic data, Syntracts offers on-premises deployment, strict data segregation, and the ability to extract highly customized data points unique to each client’s needs. Nothing built for one customer is ever sold or shared with another, preserving confidentiality and competitive advantage.
Co-founders Doug Bemis and Christopher Martin have been friends since they were three years old. Doug, CEO, has a PhD in Neuroscience and is a serial entrepreneur who has dedicated his career to building systems that truly understand our language, rather than just generating text. He has experience implementing machine learning both at scale in Uber AI Labs and within startups, with a focus on bringing advanced ML techniques to production. Christopher is an experienced M&A attorney who led the Emerging Technology Team at Latham & Watkins, where he specialized in deploying and integrating legal technology across the firm. Together they created Syntracts, named after the core technical process of generating synthetic data in order to train their expert contract models. Their complementary expertise in advanced AI and practical legal operations drives Syntracts’ to deliver unparalleled accuracy, efficiency, and intelligence to legal professionals.
What inspired the creation of Syntracts?
I come from machine learning and language, not law. But I’ve always liked law because it’s an analytical framework built in words: it’s complex but also constrained. My background in language makes me believe that in order to correctly process language and have a true understanding, you need a more structured approach than current tech allows. My co-founder Christopher was a big law attorney who wanted to improve law with technology. When ChatGPT came out, he, like many, thought it might solve everything. But he quickly realized the gaps. He brought those problems to me, and we saw a perfect opportunity to develop the technology and launch Syntracts.
What is Syntracts’ North Star?
Our mission is to provide instant and complete understanding of any contract. We built a platform with custom machine learning models specifically designed for legal documents. These models learn to analyze contracts in a more structured and reliable way.
Tell us about some recent milestones that Syntracts crushed.
An AMLaw 25 Firm signed a multi-year partnership to deploy Syntracts across the entire firm, after they tested us against alternatives.
We were accepted into Allen & Overy’s Fuse incubator, which is highly selective.
What sets Syntracts apart?
A lot of companies are taking a “prompt and hope” approach, asking questions of large language models and hoping for accurate results. It looks good on the surface, but often fails under real demands and lacks precision and scalability. Our platform is built to iterate and improve over time using small, precise machine learning models that adapt to client documents with simple feedback, rather than constant prompt engineering. Our solution delivers both high accuracy and scalability, and we uniquely offer on-premise deployment with no third-party APIs involved, ensuring privacy and security. Each client effectively has their own dedicated model, tailored to their specific contracts and needs.
What has been the greatest challenge founding Syntracts?
There’s a lot of noise in legal tech right now because generative AI makes it easy for anyone to claim they can solve complex legal problems. Too many think it can happen simply by prompting large language models, which isn’t true. While optimism around GPT “solving everything” has faded somewhat, there’s still widespread confusion and hype that make it hard for real solutions like ours to stand out.
Any favorite podcasts or books?
I listen to the Equity podcast a lot. Recently I’ve been reading Founders at Work, an older book, but great short vignettes about how various companies got started. It’s interesting to see the different paths founders take.
Fun Facts:
I have twins; I completed an Ironman; I played the violin for 18 years, but transitioned to the guitar.
Listen to Jason Saltzman with Jenny Fielding on the Venture Everywhere podcast: What a Relief! Now on Apple & Spotify and check out all our past episodes here!
