I’m often asked: “What can I patent?” It’s a great question – but AI founders actually need a different question:
“What SHOULD I patent?”
Having the right patent can attract acquisition interest, foster collaborations, and build an innovative image. But not all patents will achieve these aims – in fact, most won’t.
That’s because what you patent matters! Everyone assumes they should patent their coolest, most advanced ideas. But that’s usually wrong – especially in AI.
To find the right stuff to patent, consider two things:
1. The threshold for patenting is lower than most realize.
anything new and technical can potentially be patented.
2. There are no patent police. You have to detect if someone is infringing your patent. Consider this example:
Recently, an infringement lawsuit against Mailchimp, about a patent for an internal server process, was thrown out of court because the plaintiff actually had no idea whether Mailchimp was infringing the patent or not. If you can’t detect infringement, the patent probably can’t help you!
Since most of AI’s magic happens deep under the hood, your best opportunities are probably not your most advanced AIs.
For effective patents, focus on what’s externally “visible”, like:
🔹 Technical aspects of the business model, i.e. around how/when to use the AI, new use cases or verticals
🔹 User interfaces, or anywhere there is human/machine interaction (HMI)
🔹 “Visible” design elements that are not commonly found in other products
🔹 Hardware (processors, sensors, etc.) used with the AI, or
🔹 Any other way the model(s) show up in your product, e.g. how inputs or outputs reveal what’s underneath, etc.
If any of these are also important differentiators, they will be your top contenders to consider for patenting.
A well-planned patent starts with finding the best subject.
It’s not complicated, but your natural instinct may deceive you!
Follow this guidance, and you’ll be way ahead of the pack.
Postscript: Yes, big companies patent undetectable stuff. Because their patent goals are different than yours. As entrepreneur Steve Blank says: “A startup is not a smaller version of a big company. Startups need different strategies.”