Knowing what not to file

Gladstone Institutes is an independent, nonprofit biomedical research organization in San Francisco's Mission Bay, founded in 1979 and academically affiliated with UCSF. Its roughly 30 faculty and 450 staff work across five institutes on cardiovascular disease, virology, neurological disease, data science and biotechnology, and genomic immunology, with a record that includes Shinya Yamanaka's development of induced pluripotent stem cells and two Nobel laureates among its senior investigators. Gladstone licenses its intellectual property to startups, established companies and other research institutions through its Corporate Liaison and Ventures team.

Girard Courteau is Patent Counsel at Gladstone Institutes, where he handles the institute's invention disclosures, patent prosecution and the licensing and marketing work that follows.

Every record of invention that comes out of Gladstone's labs has to be triaged: is there something patentable here, is there a market for it, and is it worth the money to file? For a small office, the triage itself was the bottleneck. Assembling the prior art, the commercial picture and a recommendation for each disclosure took real time, and the inventions that would eventually be dropped consumed as much of it as the ones that went forward.

Girard's team had been weighing AI tools for disclosure review and marketing for some time. The question at the outset was not whether the technology would help but whether a nonprofit research institute could justify the cost.

Girard puts the time saved on initial triage at somewhere between 30 and 50 percent. The bigger effect is downstream. Because Tradespace's assessment surfaces early which disclosures are not worth pursuing, the office no longer spends attorney time or filing fees on them, and the effort goes to the projects that have a real chance of being licensed.

If you're not going to pursue something based on the information Tradespace provides, you don't waste your time filing a patent on it. You don't waste the money filing the patents on it. The whole process is conserved enough that you can then spend your efforts, energies and money on other projects.

That reduction in time and effort on inventions that will not go forward has, by his account, been enough to cover the cost of the system. His advice to colleagues in technology transfer who are still deciding about AI is to start now.

• 30–50% less time on initial triage
• Savings on abandoned disclosures cover the cost of the system