How to create a killer patenting strategy using AI: Rouse case study
Rouse, a management consulting firm specializing in intellectual property, collaborated with GetFocus to develop a comprehensive patenting strategy for its clients. Facing challenges in effectively gathering IP intelligence across diverse domains, Rouse utilized GetFocus' custom AI classifiers. This approach allowed Rouse to take control over AI-assisted patent searches, bypassing the limitations of traditional Boolean and unsupervised machine learning methods.
Erik Oskarsson
Rouse Sweden, Principal
The challenge
Rouse wanted to create a patenting strategy for one of its clients but had no reliable way to gather the insights it needed to advise its client well.
The outcome
Rouse investigated 4 technology areas using GetFocus and got crystal clear insights on which direction to go in to create a killer patenting strategy for its client.
Rouse uses GetFocus’ “Strategic intelligence” to help its clients take better decisions
Rouse is a management consulting firm that specialized in IP matters. It was established in the UK in 1990 and has 18 offices in 13 countries.
Introduction
Rouse is driven by delivering excellent IP related advice to its clients and has built a great reputation over the years. Consequently, Rouse attracted a wide range of clients, operating across many different industries and technological domains.
Due to the diversity of its clientele, Rouse faces significant challenges. In order to advise its clients well, Rouse needs the ability to quickly gather IP related intelligence across many domains. As its analysts cannot be experts in everything, they need a tool that helps them get expert level results regardless.
Traditionally, patent search tools allow you to either perform manual Boolean searches, which require you to specify every single search term, or they provide an unsupervised machine learning algorithm that categorized patents for you. Neither of these solutions work very well.
With Boolean search, you must be extremely skilled in a topic to know all relevant search terms, and even then, you are likely to get noisy results. On the opposite end of the spectrum, unsupervised machine learning provides the issue that the machine chooses the categorization of patents for you, without you having any control over the process. Both options lead to poor intelligence gathering, which harms the quality of the advice you can give to your clients.
"We have tried all sorts of tools but it is always a problem to get reliable results."
- Erik Oskarsson, Principal Rouse Sweden
Creating patent strategies is one of Rouse’s specialties. In this case study, we investigate how Rouse used GetFocus to gather all the intelligence it needed for creating a killer patent strategy for its client.
The challenge: Mapping out the external landscape
Creating a winning patent strategy involves two key steps. First, you figure out which proprietary technology elements give you a competitive advantage. Second, you learn how these elements fit into the external patent landscape, so you can prioritize your patent filing in the least densely patented area.
Rouse’s client identified 4 technology areas that gave it a competitive advantage. Because of the sensitive nature of the information, we use generic representations here. The areas were: A: production process 1, B: production process 2, C: device detail, and D: use of certain materials. Rouse’s challenge was to get a reliable representation of the external patent landscape for these technology areas.
"We needed to know exactly what the external patent landscape looked like."
- Erik Oskarsson, Principal Rouse Sweden
So, how did GetFocus help Rouse get reliable results?
The solution: Custom AI classifiers
GetFocus’ patent search technology is different, it neither relies on Boolean search nor on unsupervised machine learning. GetFocus’ AI classifiers are fully supervised, meaning that you are in total control of the AI. By making the process supervised, Focus combines the best of human intelligence with the best of AI. You are in control, but the AI does all the hard work.
With GetFocus, Rouse could train an AI classifier for each of the technology areas it wanted to investigate. With AI classifiers, Rouse taught GetFocus’ algorithms which patents were relevant, without having to do complex Boolean searches. By executing simple searches and simply clicking the relevant patents, Focus automatically learned which unseen patents were relevant.
"Having the ability to create our own technology categories allowed us to find every single relevant patent. Usually, you are stuck with whatever categories a tool forces onto you and they never match how the client sees the world."
- Erik Oskarsson, Principal Rouse Sweden
After a short training period, Rouse let GetFocus’ patent finding algorithm loose on global patent data to retrieve every relevant patent.
The results
After running the classifiers, Rouse got crystal clear insight into how densely each technology area was patented. There were over 4000 patent families in area D, over 600 families in area C, about 200 families in area B, and only 8 families in area A.
Because area A was hardly patented, it was the clear winner. In this area, Rouse’s client was the most likely to obtain broad protection for its technological competitive advantage.
"The results were crystal clear and helped us create a great patent strategy for our client."
– Erik Oskarsson, Principal Rouse Sweden
Patent landscapes made with GetFocus AI classifiers are flexible, easy to generate, and of excellent quality. A similar undertaking performed with traditional search tools would have resulted in a frustrating experience with poor results.
- Rouse was able to train 4 AI classifiers with ease.
- They got clear insights into how densely each of the technology areas were patented.
- Without GetFocus, this exercise would have been more frustrating, more time consuming, and would have yielded lower quality results.
- The insights created by GetFocus were used by Rouse to create a killer patenting strategy for its client.
Interested in our other use cases?
We strategically orchestrate AI to give you an edge.
%20(1).png)
%20(5).png)
.png)
%20(3).png)
Reach out to us!
Start a conversation. We are happy to help.