Customer Stories
How Groq's legal team does more research in-house
Research, drafting, due diligence, litigation support
When Michelle Park joined Groq as Senior Legal Director, she inherited a fast-moving company, a lean legal team, and a research workload that had no obvious ceiling.
Every question that arrives in a lean legal department carries a hidden cost: the time it takes to find the answer before you can even begin to advise on it. At Groq, one of the fastest-moving AI infrastructure companies in the industry, those questions don't slow down for headcount.
Michelle Park and Chief Legal Officer Claire Hart cover IP, litigation support, contract review, regulatory questions, and executive communications for a company that expects legal to move at the same speed as the business. Outside counsel handles the work that genuinely requires it. Everything else stays in-house. The question is what "everything else" actually costs in time, and what it means to route a research question to outside counsel when the answer might already be in your own documents.
Before Ruli, that cost was often invisible. After Ruli, it became measurable.
A team covering the scope of a much larger one
Groq builds inference infrastructure at a pace few competitors match, which means the legal team fields questions across jurisdictions, product areas, and deal types that would stretch a department twice its size. The challenge was not expertise. It was the research layer underneath every judgment call: the document searches, the first drafts, the hours spent reconstructing context before substantive analysis could begin.Searching Large Volumes of Documents for Due Diligence: Ruli excels at locating documents across large Google Drive folders, including patent applications and invention disclosures. One standout example: the AI quickly identified relevant patent filings that mapped to a new invention disclosure, saving time and increasing efficiency in the decision making process.
Claire Hart describes the shift directly. "We've reduced hours of manual work with the help of Ruli. It accelerates research, helps us with drafting, assists in managing a complex IP portfolio, and gives us the confidence to make strategic decisions faster."
The change was not about replacing legal judgment. It was about how much ground the team could cover before that judgment was required.
Research grounded in your own institutional knowledge
Michelle describes how she approaches Ruli in terms that explain why it works for her specific situation. "I treat it like almost a first year associate," she says. "I say, hey, there's that document, it's called this, please take a look at that, they've responded to it, and I try to give it the history."
The difference between that and a general AI tool is what Ruli is working from. Groq's knowledge base holds prior matters, playbooks, correspondence, contracts, and internal documents, all synced from Google Drive. When a research question comes in, the output is grounded in Groq's actual institutional knowledge rather than generic legal reasoning. The team's existing positions, language, and context are already there.
The search that showed what the tool could do
The clearest proof of how this works in practice came from a due diligence search across Groq's patent portfolio. A new invention disclosure arrived alongside an engineering design document. Michelle uploaded both and asked Ruli to find prior filings in the existing portfolio that mapped to the disclosure's claims. The results came back in minutes.
"It identified relevant patent filings in our portfolio against a new invention disclosure in minutes," she said. "Something that would have taken longer manual review. It's proven useful as a preliminary research assistant for early-stage discovery."
She took the findings to outside patent counsel, who confirmed the cases Ruli identified were the right ones. A search that would have required significant manual review across hundreds of documents, or a call to external counsel before the team even knew what they had, was complete before the external engagement began.
The same pattern holds across other research tasks. A regulatory memo that would have taken hours to brief now arrives on a business leader's desk as a one-page summary of risks, action items, and decisions. A discovery request with a broad scope gets a preliminary document pass before attorneys narrow what actually needs review. A demand letter that used to start from nothing begins with a draft grounded in prior correspondence already in the knowledge base.
The attorney still reviews and decides
Michelle is direct about what this means for the role of the lawyer. In-house counsel do not want their work to look like they simply asked the AI. The professional judgment and the accountability stay with the attorney. What Ruli removes is the work that precedes that judgment: the search, the synthesis, the first draft. The lawyer still reviews, decides, and signs off. The difference is how much ground has already been covered before that moment arrives.
"Ruli feels like a trusted assistant that never gets tired," says Claire Hart. "It gives us the confidence to make strategic decisions faster."
For a team covering the scope of Groq's legal work, that confidence has a practical effect. Less time on the research layer has meant more time on the strategic partnership work that in-house counsel is supposed to be doing.
The legal team at Groq covers the kind of scope that, at most companies, requires significantly more headcount or significantly more outside counsel spend. What changed was not the team's size. It was the cost of getting from a question to an answer, and what happens when that cost drops far enough that more research stays in-house.
"
"Ruli feels like a trusted assistant that never gets tired. It accelerates research, helps us with drafting, assists in managing a complex IP portfolio, and gives us the confidence to make strategic decisions faster. We've reduced hours of manual work with the help of Ruli."
Claire Hart
Chief Legal Officer
Customer Stories
