When I founded Tangibly, I strongly believed in the idea of being purely a trade secret play. There were plenty of SaaS tools for other types of IP, especially for patents and trademarks. Wading into those areas would undoubtedly be a distraction. For the first year, if anyone mentioned an idea around patents, trademarks, or copyrights, I quickly dismissed it. Tangibly is all about trade secrets, or so I believed.
Then our customers started telling us what they thought…
Our first point of enlightenment came when a patent agent from our partner (and now investor) DLA Piper said that she liked to analyze patents to try to find areas where there might be underlying trade secrets. The light bulb went off! I knew immediately that Liat could create an AI model to do the same thing. Fast forward six months and we launched Patent X-Ray a tool that ingests issued patents and published patent applications and predicts the underlying trade secrets. Patent lawyers and agents loved it but wanted that same capability for draft applications so they can decide in real time what gets disclosed and what remains a trade secret.
Our second point came when a Fortune 100 customer wanted to run a pilot but were most interested if Tangibly’s functionality could be easily adapted for managing copyrights. Sure it can, but my trade secret myopathy made me skeptical that this was a real application. Fast forward about nine months to today, and copyrights and their influence on Large Language Models is a global headline. It’s coming from every direction: Sarah Silverman suing OpenAi, EU regulation requiring training data to be reported, and of course Harry Potter. Luckily for us much of the workflow and process we architected to manage trade secrets also applies to copyrights.