The legal tech CEO spoke to CL Talk about AI’s limitations but how it is now solving key challenges
Tali Green is the CEO of legal tech platform Goodfact, which builds case chronologies for litigators. She spoke to the CL Talk podcast about why she initially did not build her product with artificial intelligence but has recently integrated it. Green also spoke about the shift in the legal profession's acceptance of AI and plans for her product, including medical record optimization and expanding into global markets.
Below is a summary of the conversation:
Tali Green says her team initially avoided artificial intelligence when designing the product that builds case chronologies. “Some of the chronology tools out there… are basically AI-powered, and ours is not,” she explains. Her decision was rooted in concerns about accuracy. “When you put the documents we wanted to work with into an [AI model] like ChatGPT… the output was not accurate enough.”
Green describes the challenges her company aimed to address early on, such as parsing large batches of email threads and chat conversations into precise chronological narratives. “Our goal is to… put every individual message, whether it’s from an email thread or a chat conversation, where it belongs in the chronology,” she says. This commitment to precision meant building a product that would reliably sort and organize messages by date and time without relying on AI. “It had to be 100 percent accurate.”
After establishing the platform’s core features, Green and her team began incorporating AI to tackle specific challenges, such as handling PDFs. “We knew the power of AI, but we didn’t want it to… [risk inaccuracies],” she says. Goodfact’s first use of AI focused on tasks traditional programming struggled with, like identifying correct dates and naming files in PDFs.
The integration has since expanded to include summarizing case data, a feature that Green says enhances the way users interact with the tool. “Imagine you have just a chronology of documents… now AI summarizes every document or fact,” she says. This update allows users to view high-level overviews of cases while drilling down into specific details. According to Green, the tool now provides a “layered exploration” of facts, breaking down cases by year, month, or week and linking summaries to their source materials.
Green says text message evidence is becoming increasingly important in litigation, particularly in family law and estate cases. She notes the inefficiencies of relying on screenshots for this type of evidence and explains how Goodfact addresses the issue. “We have the entire chat conversation, which will be analogous to an entire email thread,” she says. Her company struck a partnership with the company Hearsay to extract these chat histories, which allows users to seamlessly include messages, emails, and other documents in a single chronology.
Green says the perception of AI in the legal profession has shifted over the past few years. “Two years ago… [lawyers] were skeptical or nervous about legal tech and AI in particular,” she says. Now, she notes, there is a growing demand for tools that can process and summarize vast amounts of data. However, she remains cautious about adopting AI too broadly. “We took a bet with not being AI-first… and that was the right call,” she says, explaining that users have reported inaccuracies with AI-first tools.
She emphasizes the importance of reliability in legal tech, particularly for litigators. “Litigators… are definitely not going to rely on a tool that’s even remotely inaccurate,” she says.
Green shares plans to further expand Goodfact’s capabilities, including features for optimizing medical records. She explains that this development could benefit not only lawyers but also clients in other sectors that rely on document-heavy workflows. With users across multiple countries and the platform’s ability to work in various languages, Goodfact’s potential applications are broad.
On the future role of AI, Green highlights the possibility of integrating drafting tools that create affidavits and other legal documents from chronologies. While this could save time, she reflects on the implications for the practice of law. “The soul of litigation… [is] applying the facts to the law,” she says, expressing concern about automating tasks central to the profession.
“We wanted to avoid [the risks of] AI… and only use it where we absolutely needed it.”
This conversation can also be found here:
The episode can also be found on our CL Talk podcast homepage, which includes links to follow CL Talk on all the major podcast providers.
