After a first look at ChatGPT in November 2023, Bob Hansen and his partner began experimenting with AI, hosting open-source models on their own secure systems to protect client confidentiality. The quality wasn’t quite there.
Then they heard that a large law firm was using Solve Intelligence. It was a wake-up call: if a large firm could combine high billing rates with AI efficiency, Marbury’s boutique advantage could disappear.
The firm signed up for Solve, impressed by its security standards and by a team that handled model optimization so attorneys could focus on using the tool. Bob tracked his own drafting time before and after, and once he had hard numbers, the business case became undeniable.
Day to day, the chat inside Solve replaces copying and pasting between tools: it can rewrite any part of a document in tracked changes and explain unfamiliar technology and terminology. Every document loaded into a project becomes part of its context, so the AI works from the right material without complex prompts.
To bring others on board, Bob relied on demonstrations: Solve’s data security information for clients who asked, a side-by-side cost comparison, and a session at the firm’s AI Summit where he drafted a 110-paragraph application live with Solve.
“After I drove home, I dropped the information into Solve to generate the drafts. The next morning, we filed three provisional applications. That was not physically possible before Solve.”
A complex application used to take Bob 10, 15, maybe 20 hours. With Solve he does it in two or three hours, and covers more embodiments, more alternatives and more background technology, for better enablement. Where he once prepared applications of 70 or 80 paragraphs, he now turns around 120-paragraph applications.
Clients responded: once he showed them the security assurances, the cost benefit and the turnaround time, they gave him more work. And when a client called about three new inventions, he filed three provisional applications the next morning.
