Term Scout’s AI Strategy: Why They Chose Depth Over Breadth in Legal Tech

As every tech company rushes to integrate ChatGPT and large language models, Term Scout took a different path. In a recent episode of Category Visionaries, CEO Otto Hanson revealed why focusing on depth over breadth in AI development became their competitive advantage in legal tech.

The Contrarian Approach

“We’re kind of more like a small language model. We want a much smaller data set, but much higher quality signal in that data set,” Otto explains. This decision wasn’t about technical limitations – it was about understanding what enterprise customers actually need.

After an early failed experiment with crowdsourced contract reviews, Term Scout learned a crucial lesson about the importance of expertise. “Literally inside of a day, we had five different attorneys review the same contract on like, ten key metrics… and the results were all over the place. Like, one attorney gave it a five on privacy, another attorney gave it a one on privacy.”

Building AI with Subject Matter Expertise

Instead of chasing broad AI capabilities, Term Scout focused on embedding deep legal expertise into their models. “We’ve custom built a training set that has been hand trained by subject matter experts and has been really effective at answering the types of substantive questions that we know an experienced attorney in the field would want to know.”

This specialized approach allows Term Scout to tackle specific, high-value problems. As Otto notes, they’re “answering the types of substantive questions that we know an experienced attorney in the field would need to know the answer to in order to determine is this contract better for you or for me?”

Maintaining Trust Through Transparency

Term Scout’s AI strategy is deeply tied to their trust-building approach. “Customers that get in there quickly find out, wow, they actually show their work. If we tell you that X clause is present within one click of the mouse, you can actually see the source language from the contract that proves the veracity of that statement.”

This focus on verification and transparency has led to impressive results. “We’re seeing more than 40% of customers are starting to sign that contract without negotiating,” a remarkable achievement in enterprise sales where contract negotiations typically cause significant delays.

Evolving with the AI Landscape

While Term Scout remains committed to their focused approach, they’re not ignoring broader AI developments. “We’re super excited to see this breakthrough, the conversational, chat based AI and the improvements underlying the GPT 3.5 updates there… we’re even experimenting as we speak with adding some elements from GPT into our product to enhance it even further.”

The key is maintaining their core differentiation. As Otto explains, “At the same time, we feel even more confident about our own direction and the AI that we’ve built and trained in house, which is really like a different type of model.”

The Future of Specialized AI

Looking ahead, Term Scout sees their specialized approach as key to expanding beyond enterprise software contracts. “Contracts touch people and businesses equally. They don’t discriminate,” Otto notes. This vision requires AI that can handle increasingly complex contract analysis while maintaining accuracy and trust.

For founders building AI-powered products, Term Scout’s journey offers valuable lessons about the power of focus. While large language models offer impressive general capabilities, there’s still immense value in building specialized AI that deeply understands specific domains.

The real innovation isn’t always about having the biggest model or the broadest capabilities. Sometimes, as Term Scout demonstrates, it’s about having the most relevant expertise encoded in your AI, delivered in ways that earn and maintain customer trust.