Since Shine graduated from law school in 2013, document review and discovery mechanics evolved into electronic information as emails became the primary source of records in litigation and investigations.
Shine, now a partner at Selendy Gay PLLC, said the volume of records became unwieldy in recent years as organizations added chat-based systems such as Slack and Teams.
While law firms have turned to technology-assisted review and predictive coding for these electronic records, challenges persist. Rooms full of contract reviewers were hired on complex cases, which Shine said was costly and time-consuming for both law firms and clients.
"This was always one of the most painful aspects of civil litigation," Shine told Law360 Pulse.
Then came ChatGPT.
The arrival in late 2022 of generative artificial intelligence chatbots, which are capable of generating content, caught the attention of some law firms as a way to eliminate this e-discovery bottleneck. Selendy Gay was one of those firms.
While some law firms shied away from generative AI due to fears about false outputs from the technology, Selendy Gay embraced the tool to make discovery more efficient. The New York-based firm, which focuses on complex litigation and arbitration, has 80 lawyers, including 17 partners.
"The partners here were very progressive-thinking for lawyers," David Dinakar, the chief technology officer at Selendy Gay, told Law360 Pulse.
In late 2022, Selendy Gay began conversations with legal generative AI companies Harvey and Casetext. Both are well known and have been used extensively by the largest law firms, but Dinakar said that their products came with enormous price tags.
"It was ridiculously expensive," Dinakar said. "We got some pricing. It just didn't make sense."
Then a lesser-known legal AI startup named Fileread reached out to the firm in 2023. The company changed its name to Aurelogy in January. Unlike Harvey and Casetext, the startup lacked popularity, size and extensive use among large law firms. However, it was looking for partners to beta-test its new platform.
Selendy Gay's partners met with company representatives in January 2023 and negotiated on pricing. By March, the firm had signed on for a pilot.
Picking an Unknown Startup
Affordability was one reason for selecting Aurelogy.
A representative from Aurelogy told Law360 that the company's pricing is transparent and simple, depending on document volume and use cases. The company charges per document, which may be different from other providers that charge per user or page. While Aurelogy declined to give specific prices, the company said that the price depends on different factors such as document consumption volume, use cases and whether there is a subscription.
By contrast, Casetext prices its tool at $400 to $500 per month, the company told Law360 Pulse for a pricing analysis in October 2023. Harvey did not respond to a request for comment for that analysis.
Selendy Gay declined to comment about why Aurelogy's price was reportedly better than prices from the other vendors.
There were other reasons Selendy Gay selected this startup. The platform was fully in the cloud and integrated with Relativity, the e-discovery platform that Selendy Gay uses.
Shine said that the firm was comfortable with Aurelogy because Selendy Gay itself is a young firm, founded in 2018, and shares a "startup quality" with the company.
"We are relatively forward-thinking about technology, including our willingness to sort of put things in the cloud earlier than some of the law firms that have more legacy systems," Shine said.
The biggest reason for picking the startup was the customization, the firm said.
Dinakar said that other tools such as Harvey and Casetext were focused more on serving the needs of larger firms and offered out-of-the-box functionality. This helps with many use cases, but law firms sometimes don't want one-size-fits-all solutions.
By contrast, Aurelogy developed its platform in partnership with Selendy Gay and customized the tool to meet the needs of the firm.
"It was less attractive to us to just take something off the shelf and hope that it met our needs," Shine said. "The offer that Fileread was making, which was very attractive to me, was that their technical staff would come and sit down with our attorneys and they would work through product issues together."
Customizing the startup's platform with Selendy Gay gave the firm more input and control over the development of the technology, according to Shine.
"Working closely with partners like Selendy Gay is important to ensure that our product is developed in a way that attorneys can find tremendous value from," a representative from Aurelogy wrote to Law360 Pulse in an email.
It also provided more comfort than a generic product because the tool would be adapted and suitable for specific use cases. Selendy Gay declined to comment on specific use cases.
"It's great having a partner like that to work with you in your firm, talking to our attorneys in real time and developing the product to suit our needs," Dinakar said.
Benefits aside, risks loom large for law firms that work with a generative AI startup, especially one that is brand new to the market. The concerns about generative AI include data privacy and the production of false outputs.
Selendy Gay has been monitoring recent cases in which lawyers from other firms faced trouble for using generative AI to create briefs. Shine said that those cases are different because Selendy Gay double-checks the AI output on the back end instead of directly submitting the output to the courts.
After signing up for a pilot with the startup, Selendy Gay used the tool for two matters in 2023. While the firm declined to mention the nature of the two matters, Shine — who was involved with both cases — said that the matters were very large and technically complex commercial cases with a lot of money at stake and a lot of documents in discovery. Selendy Gay declined to disclose how many pages or bytes of documents had to be reviewed.
Shine admits that the generative AI tool could not produce the same level of quality as a room full of human document reviewers, but the results were still very good. The tool allowed the firm to conduct natural language searches in those documents in minutes.
"It was a huge force multiplier for us in terms of looking out over this huge universe of documents and pulling out the essential things that the lawyers needed to know relatively quickly," Shine said.
Shine added that the generative AI tool mimicked a natural conversation, which allowed the firm to ask follow-up questions about evidence in the documents. The tool also provided citations to its answers.
"It allowed us to get to the heart of our factual questions about our cases much faster," Shine said.
Selendy Gay is now rolling out the generative AI tool for other matters.
It's still early days for the use of generative AI in litigation and e-discovery, but Shine is excited about the technology's potential.
"I think it allows any individual lawyer to be more productive and more efficient," Shine said. "It's a great complement to what we do."
--Editing by Robert Rudinger.
For a reprint of this article, please contact reprints@law360.com.
