“Intellectual property is the oil of the 21st century,” Mark Getty once said, and businesses have caught on. In 2000, 182,218 patents were issued by the U.S. Patent and Trademark Office. In 2020, that number more than doubled to 399,055. And in China? It’s over a million new patents each year.
While creating IP may not require any physical drilling, it does require a lot of prospecting—and mining through the vast amounts of patent data out there can be incredibly intimidating.
To determine the patentability of a product or idea, or defend one against challenges, proper patent research and analytics require technical experts in particular areas to think through every possible synonym, word combination, Boolean search term, and source to ensure they’re not missing anything. Doing so could stop a patent application from being granted, wasting time and money; or worse, lead to even bigger losses in the courtroom.
It’s an expensive task, and despite everyone’s best efforts, this kind of searching might still miss an important piece of prior art if a word or nuance isn’t thought of along the way. There is just so much data to get through—and so much risk of missing something critical.
The use of Boolean logic has long been the methodology of choice in patent searches in the digital era. But that’s on the cusp of change. Here’s a high-level look at what’s true today—and what’s on the horizon for IP counsel.
1. IP feeds into many different business strategies. They all require data wrangling.
“For some companies, patents are a really important property right to fence off space from competitors or provide an important competitive advantage; for other companies, it’s different,” explained David Donoghue, head of Holland & Knight’s IP group, during a patents-focused session at Relativity Fest 2021. “I had one client say ‘My patents are playing cards—I need a certain number in my deck, and don’t care about much beyond having that critical mass.’”
Securing patents for new ideas and projects gives organizations a competitive advantage—by preventing others from mimicking unique offerings—and a portfolio of valuable IP assets. Feeding both the power of innovation and the power of ownership, intellectual property is a key feature in the modern business’s strategy for not just surviving, but thriving.
And the strategizing doesn’t stop with “simply” applying for and securing those patents. Assuming these organizations—and their IP counsel—plan to assert ownership, challenge threats, and defend their portfolios, the ability to quickly and thoroughly search through patent data for the information needed to do so can make or break their ability to operate strategically in the market.
Whether you’re defending a client and need to invalidate competitive claims, performing due diligence on a merger or acquisition, or something else entirely, cutting edge patent searching skills will help.
2. Patent search data stores are becoming so vast that human effort alone can’t manage them anymore.
“We used to literally fly to the Patent Office and look through paper copies to find prior art,” Kristopher K. Hulliberger of Howard & Howard Attorneys recalled to Relativity Fest attendees. “You’d see things that would look or sound like your invention, and you’d have a sense for where yours would fit if it was there too.”
But not anymore.
“In electronic archives, you only see what comes to the top of your search pile. You don’t know if somewhere buried down in the middle are key references,” he continued. “But what you can do—with the right tool—is get comfort from seeing many similar things, using analytics to see more prior art that looks the same. As you work through that onion and dig through each layer, it can be a challenge without the right tools and experience.”
Put simply, the dataset is too large—and it’s growing too fast—for human analysis alone to suffice. Patent search teams need to use machine learning to cut through the noise and discover what matters more quickly and accurately.
Fortunately, human effort alone isn’t the only option anymore. Software like Relativity Patents can help by bringing industry-leading technology to patent search and analytics.
When it comes down to the work itself, human judgment is still more important than what a machine can do. Thus, IP experts should rally around a search tool that leverages machine learning capabilities in ways that make it easy to use on queries large and small, but feel intuitive to the expert minds who are doing the most important work.
3. Augmenting human effort with machine learning will yield better patent search results. But it’s going to take practice to popularize the option.
With the right combination of technological and human thinking, IP teams can be more confident that they’ve turned every stone that matters in their efforts to secure, defend, or challenge a patent. Still, despite the exceptional results, it isn’t always easy to make a paradigm shift away from using Boolean search alone.
“I’ve seen many people feel resistant to the tech, saying, ‘Oh, there’s no way AI can be as good as Jake,’ who is an actual partner of mine and is a phenomenal tech sleuth,” David explained at Relativity Fest. “But the reality is that Jake has limitations that technology doesn’t, and he can have a bad day—whereas AI can’t have a bad day.”
For David’s team and others, the proof is in the pudding.
“What I’ve found internally, using Relativity Patents, is that you have to use the ‘John Henry races the train’ fable as inspiration: do that race, compare the results, and after the AI ties our brilliant colleague Jake—because he’s really that good—a couple of times, people start to appreciate the upside and that the consistency is really valuable.”
Giving the AI-powered options a try and comparing them to human effort, at least at first, can be a highly effective way of showcasing real-world benefits to skeptical IP teams with minimal cost.
The bottom line? In today’s landscape, if you're not using machine learning, you may be missing something crucial—and doing a disservice to your clients (internal or external).