For Roxanne Russell, holding a patent is a family tradition.
Her grandfather held a patent for air conditioning infrastructure and her father held one for stadium bleachers.
At the time we decided to go for the patent, I had seen a report that only 21% of patents included a female inventor. It was important to me to contribute to efforts to change that.
Owning patents was a point of pride because it qualified them as inventors, and the patents later yielded handsome profits when sold. So when she developed Read Ahead, a tool to transform any digital text into a guided reading activity in seconds, her thoughts went to patenting it.
“I found the idea of a patent to be of value both personally and professionally,” said Russell, who is currently the director of online education for Columbia University’s Mailman School of Public Health. She explained, “At the time we decided to go for the patent, I had seen a report that only 21% of patents included a female inventor. It was important to me to contribute to efforts to change that.”
While creators in the edtech space often feel they’ve made something patent-worthy, two problems usually stand in the way. The first hurdle is the cost, as attaining a patent is often pricey. The second challenge is time. From filing a provisional application to securing approval, it took Russell five years to get Read Ahead patented.
She opened up to EdSurge about the process she went through to give would-be edtech inventors a glimpse into the trials and tribulations involved in getting a patent.
Edsurge: When did you first consider patenting Read Ahead?
Russell: While I was a graduate student at Georgia State University, I had the idea to make PowerPoint presentations that chunk material into smaller bits as a way to help keep students’ eyes on the page. A research colleague and I conducted a study using these PowerPoints and found that our method worked. Not only did it increase reading motivation, but scores on reading comprehension tests improved as well. But we were spending hours making each presentation and teachers just don’t have time to do that. That was when I recruited a team to build a tool to do it automatically.
It turned out our examiner was considered extra strict and had a record of granting only 10 percent of patents over eight years.
We finished a working prototype and could tell that we had created something unique. No other tool existed to allow teachers to quickly create unique reading presentations using AI to identify and illuminate key concepts, a machine learner to benefit from user input and a player-style display to chunk reading assignments into digestible segments. We knew it was novel, nonobvious and useful, which checks all three boxes required for a patent.
What was your first step to getting your patent?
We hired a really good lawyer. That’s the most important piece of the puzzle after you have something that can clear the hurdles of what can and can’t be patented. Our lawyer was very honest with us about how volatile intellectual property is, especially in the software space. She was always honest about how unlikely we were to get the patent, but we pressed ahead.
The first thing she did for us was conduct a patentability search, which verified that nothing similar existed prior. The next step was making sure we protected ourselves by clarifying who owned the IP we created before we jumped into the process. Our lawyer guided us to complete a patent assignment process that secured ownership rights while acknowledging the contributions of all inventors. When we made our first provisional patent application, there was no question about who owned the patent rights for Read Ahead. (Russell and her team spent about $8,000 on these early steps.)
Were there any tricks to filing the patent application?
One trick our legal team knew was to call Read Ahead an apparatus instead of a process. Things get messy when you try to patent a process itself. We made sure to clarify, describe and define what our tool actually did. With the application submitted, we were able to say Read Ahead was patent pending. (This application process cost Russell’s team about $4,000.)
What difficulties did you encounter?
It took a while but we eventually found out that we had been rejected. This is when we really learned how wonderful our lawyer was. She brought in an appeals expert who looked up the patent examiner we were assigned to. It turned out our examiner was considered extra strict and had a record of granting only 10 percent of patents over eight years. This examiner had an inordinate amount of control over the decision.
Fortunately, our legal team was allowed three examiner interviews. They took advantage of each opportunity to speak to the examiner and listen to his reasoning and logic behind the rejection and submitted targeted responses in order to appeal. Our lawyer pointed out a policy change described in a U.S. Patent Office memo that invalidated one of the examiner’s reasons for rejection. It took a really great lawyer to fight for us, and having those productive conversations with the examiner ultimately led to us winning the appeal. Our costs were easily greatest during the appeal period, totaling about $15,000.
Was it worth spending about five years and $27,000 to secure the patent?
Ultimately, it was worthwhile because Read Ahead really is groundbreaking—we wouldn’t have gotten the patent otherwise—and we’re now looking to partner with content providers to make it more flexible for teachers. Read Ahead might be further along with more content partners at this point if we had been able to have a more single-minded focus, but my partner and I have both been both working other full time jobs. In spite of this, we are impressed by how much Read Ahead has grown organically because of user enthusiasm. Thousands of teachers are consistently plugging in open content from CommonLit and Wikipedia into Read Ahead.
Most content in schools is still proprietary. Right now, teachers and students deal with a central textbook production unit that tells them how a text should be interpreted and what it means to have comprehended a text. We flip that and allow students and teachers to identify what they think is important in the text, make choices, and teach our AI engine how actual readers are reading the material. All of that information is fed back into the reading activities that are created next time. We are excited about what is possible when teachers and students are invited into this process.
I’m glad we did it. I feel okay about the cost, especially over how many years it took. All I had to do was forego buying a new car for five years to afford it. With everything that is happening right now with natural language processing in the edtech space, it is strategic to have our apparatus protected.