KISSPatent Case Study – Building A Strong AI Patent
In our last post, we talked about how to win the AI patent process. Beyond winning, we want to show you how to build a strong AI patent—and we believe that the best way for you to learn is through a case study, based on work with real clients and their patents. We can’t reveal too many details because these patents aren’t published yet. Remember: patents only publish 18 months after filing , so there are a lot of AI patents that are waiting to be published – including a great number from our clients!
We’ll take you through the steps you need to get your AI patent. And at the end, you’ll have the tools that you need to get started. Let’s go!
Step 1 –Reflect: What solution are you selling?
Remember the old saying, “people don’t buy drills – they buy holes”? Tons of startup advice emphasizes the importance of selling a solution. After all, your customers aren’t interested in your cool back office technology – they want a solution to their problem.
The very same question applies to patents. Before you begin preparing your patent, you need to figure out the solution that you’re selling. Without this vital first step, you’ll have a great patent on your technology – but it might not protect your business.
Patents are, first and foremost, a business tool, so think about your business model first, not your technology. And then you’re guaranteed to get a valuable patent.
In the AI world, we’ve personally heard from lots of innovators with great ideas – for technology. But when it comes to the business solution, these innovators don’t always know what they want to sell, or who they’ll sell to.
One of the best articles on why you should pick a specific niche with a specific problem to solve using machine learning is by Tak Lo of Xeroth.ai. Basically, the more specific the niche, the less competition you’ll have from BigCos – and the greater your chance to dominate that niche before competitors move into it.
Once you’ve figured out your solution, are you ready to talk technology for your patent? Not quite.
Step 2 – Ask yourself: What’s your business model?
The next step is to understand how you’re going to sell your solution – that is, your business model. Will you sell a full stack solution? Integrate into a solution offered by a partner? Provide a bolt-on technology that will work with other platforms?
The answer to this question will affect how your patent is structured. Each of these different business models will have different patent needs. Filing a generic patent that covers only your technology isn’t helpful. Remember what we said in step 1 – patents are a business tool.
For deep neural networks and machine learning, this means moving away from only focusing on the exact structure of your AI solution. Instead, think about how your AI will be packaged as a product and sold. A vague description of your customer solution won’t help here – you need to be concrete about exactly what your customer will get when they buy your product.
And avoid “AI-washing” your products – where you talk about your great AI features, but not your customer experience. Yes, AI is cool but you don’t want to drift away from your core customer mission – and the business model that will enable you to deliver it.
Once you’ve nailed your business model for your AI idea, you can be certain that your patent will protect your business – not just your technology.
Step 3 – Whiteboard your patent
After you’ve reflected and asked yourself about your business model, we recommend whiteboarding your patent. What does this mean?
When you describe your idea to someone for the first time, it often helps to sketch on a whiteboard. Drawings can help others understand your great idea. This is particularly helpful for AI, as machine learning is very new and not always understood.
Whiteboarding your patent has the same benefits. You can sketch out the answers to Steps 1 and 2, while simultaneously describing your cool technology (finally!) You have to integrate all three in order to get the best patent for your AI idea. By putting all of your ideas on a whiteboard, you can be certain that they are a) clear and b) complete.
After all, you don’t want to leave anything out of your patent! Or have it fail because you didn’t describe your ideas clearly enough.
Step 4 – Get help
The time has come for you to write and file for your patent—but can you do this on your own?
The short answer is that you can file your own provisional application. But patents are complicated. It’s not easy to get all of the information that you need (although we’re doing our best to provide you with a steady stream of information). And perhaps even more difficult-- you have to find the time to do it.
If you do decide to get help, beware of budget-busting patent firms that charge by the hour. Check out our post on common AI patent pitfalls.
In our next post, we’ll offer an affordable solution so that you can get your AI patent, and not go over budget.
Wondering if your idea is patentable? Have a question about this article? We can answer all of your questions — just hit "contact us" down below!