Legal technology startup Theo Ai Inc. today announced it has raised $3.4 million in new funding led by Run Ventures, alongside plans to expand its artificial intelligence-powered litigation prediction platform.

The raise brings the company’s total funding to more than $10 million, following its $4.2 million seed round in May.

Founded in 2024, Theo Ai provides access to AI-driven litigation intelligence and predictive analytics for legal settlements. The platform is designed to give law firms and corporate legal teams guidance based on historical data, with a long-term goal of “pricing every single lawsuit.” By using its AI systems and accumulated knowledge, Theo helps individuals and businesses quickly understand the true cost and risk inherent in lawsuits.

“When people watch legal dramas, they may think that most cases settle in front of a judge, while only 3% of cases are settled with the judge. Nearly 97% are settled out of court,” Theo Ai Chief Executive Patrick Ip told SiliconANGLE in an interview.

Ip conceived of Theo Ai while earning his law degree at UCLA School of Law, during the release of OpenAI Group PBC’s GPT-3.5 model, the technology that powered the launch of ChatGPT. The timing inspired him to explore how AI could improve legal prediction and transparency in settlements.

Unlike many other legal AI tools, Theo lives almost entirely in email. The platform sidesteps traditional dashboards and web portals, integrating directly with the primary communication channel of most law firms. It analyzes attachments and correspondence, then generates predictions directly within a client’s inbox.

“We’ve built something that doesn’t make you change your behavior to get the fruits of the technology,” Ip said.

The system can automatically identify case type, statute of limitations, case insurance coverage and expected time to resolution all without manual input.

Theo relies on multiple large language models on the backend, including a mix of OpenAI’s GPT, Google LLC’s Gemini and Anthropic PBC’s Claude. Each model is assigned tasks suited to its strengths. For instance, a model trained on medical records assists with injury cases, while others specialize in legal text processing. “We live in this multimodal, multi-agent type of world,” added Ip.

As with any use of AI, Theo provides companies with everything they need to verify reports. The platform does this through deep attribution. Whenever a prediction report is generated, the system cites and attributes the documents it used to reach its conclusion, making trust and traceability central to its design.

“No company can guarantee no hallucinations, but the way that we try to do it [is through] traceability,” Ip said.

Mustang Litigation Funding serves as a flagship client in personal injury cases for Theo. Ip said the firm originally started with a $51,000 contract but quickly doubled their commitment within four weeks.

When benchmarked against Mustang’s cases, Theo’s platform achieved an 85% accuracy rate, compared with an estimated 60% to 65% accuracy of human judgment. The increased precision and faster turnaround, Ip said, make Theo an appealing proposition for legal teams navigating the high costs and uncertainties of litigation.

The company said the new capital will accelerate the expansion of its AI capabilities, support hiring additional engineering talent and fund the continued development of its predictive analytics platform for the legal industry.

Image: SiliconANGLE/Microsoft Designer