According to Bryon Bratcher, the managing director of Gravity Stack, a standalone legal tech and consulting firm connected with Reed Smith, the firm is fully adopting generative AI. The group was formed in 2018 and developed various legal tech tools for clients, which it still maintains. Alongside these projects, the firm’s consulting work has consistently grown. Now, the primary focus is on offering tech advisory services related to AI.
Bratcher explained to Artificial Lawyer: “We consider this a significant shift in approach. AI is now accomplishing what we had envisioned in the past.”
In other words, generative AI has provided (while facing some ongoing challenges with accuracy and hallucinations) the level of productivity and flexibility that many lawyers had hoped the initial wave of NLP / ML legal AI tools would be capable of right from the start.
With this technology now available, Gravity Stack is seizing the opportunity to assist clients with all their legal genAI needs.
‘GenAI is a game-changer; we must incorporate it into everything we do. For instance, our team uses genAI tools, we use Perplexity AI instead of Google, and we’re also experimenting with OpenAI’s new SearchGPT capability,’ he states.
They are utilising Vellum AI, a tool for evaluating prompts and genAI models for their client work.
Bratcher mentions that this evaluation work is part of what they refer to as their ‘AI Lab’, where they will assess a client’s requirements, explore the use cases, examine the most suitable tools and solutions, and then provide advice on implementation.
‘We handle the entire process for them. Bratcher explains that we carry out tests in our labs, assess open source and commercial LLMs, and analyse costs and security—everything goes into the mix.
Clients can also use Gravity Stack’s AI Lab to test applications independently.
‘We can provide them with a test area or conduct tests together,’ he adds. ‘And there is dummy data they can test if they prefer to use that.’
Moreover, Reed Smith’s tech team also uses the AI Lab. (While discussing the main law firm, Bratcher notes that Reed Smith is currently using Harvey. When asked about its impact on the firm, he responds that ‘it’s early days’.)
Bratcher explains that they recommend both legal tech companies with their genAI product and advise clients to use an LLM suited to their needs, with some level of modification for their specific use case.
In some cases, clients may want to retain an existing platform. Still, Gravity Stack can assist in determining how it can be enhanced, such as developing a customised LLM solution—or, as Bratcher describes it, ‘ a layer to drop on top of their current system.
The genAI companies they use most frequently are OpenAI and Anthropic, with occasional use of Gemini. In the open-source realm, it’s Llama.
This site has observed that this selection seems to be the most common among law firms and tech vendors. It also suggests that although the LLM market is rapidly evolving, OpenAI and Anthropic are still clearly the preferred providers… for now.
All good. What about other aspects, such as accuracy?
Bratcher does not emphasise this as some legal tech teams and states: ‘It’s important to consider, but we do not view it as crucial. A lawyer’s work product is the primary factor and will be reviewed.’
In other words, AI will guide you to a certain point in a legal task, regardless of its accuracy, and then the lawyer’s overall work will be reviewed by more senior lawyers, so the immediate accuracy of the genAI tool is not as crucial.
Of course, one could argue that not spending additional time reviewing the lawyer’s work product would be beneficial because the genAI outputs have guided it so effectively. Or even better, I hardly spent any time reviewing it… because the genAI that contributed significantly to the work product was so accurate.
But moving on, does he foresee a change in business models for law firms? In other words, will efficiency lead to faster delivery, rendering selling time obsolete and necessitating a new business model?
Bratcher notes that speed does play a role in certain areas where clients specifically inquire about efficiency, such as with M&A due diligence. However, he does not perceive any significant changes in business models across law firms due to genAI.
However, one could then argue that if the business model remains unchanged, what impact does genAI have? As Bratcher mentioned, the response could be that it’s still early.
Returning to the initial topic, Bratcher emphasises that where they truly observe change is in the heightened demand for assisting clients in implementing genAI applications. He adds that this comes with its own challenges, such as sometimes clients are simply unprepared for it.
‘You can’t go from 0 to 100 simultaneously; you need other layers before integrating genAI. For instance, some clients want to apply genAI to their knowledge base, but they have not organised their data; maybe there isn’t even an index for it,’ he observed
Source Artificial Lawyer
Leave a Reply