Thank you for Subscribing to CIO Applications Weekly Brief
IP Intelligence and Valuation in the era of Industry 4.0
Andreas Iwerbäck, Director of Group IP Intelligence, Innovation & Technology - Husqvarna GroupNadel Phelan, Inc.
Industries converge; petrol driven engines are being replaced by electric, and battery powered engines and everything gets connected. This puts new challenges on all companies, regardless of where they come. One example is chainsaws, very traditional petrol driven heavy forest machine. Now consider developing a new sustainable chainsaw being battery powered instead. Traditional chainsaw makers would then need to learn, understand, and get access to battery technology, while battery-oriented companies need to learn sawing technology if they want to compete in this space. Who is better at what? Who will succeed first? And who’s IP will be the most valuable?
Another example is smart watering in your garden. Building a connected smart irrigation system needs connectivity. Either you are a traditional watering company and needs to learn and get access to connectivity, or you are a connectivity company needing to learn watering and irrigation. Who will succeed in that race? And is watering or connectivity IP most valuable?
IoT opens up both opportunities and challenges. From one perspective, it’s an excellent opportunity to add value to an already existing business. Connected devices and sensors are the fundamentals for AI, as you need the data collection for training the AI. It will also give you better insights into customer service and in the long run, better customer value.
The challenge is patent licensing for IoT since telecom players having developed the standards, now requests to license its patents to all new implementers. Many of these new implementers usually are not exposed to patent licensing before, hence the challenge on both ends. The licensor needs to educate and be transparent on how and why these companies need to take a license. In turn, the potential licensees need to have an open mind and understand the fundamentals for how these standards are developed, including the so-called FRAND obligations.
In the area of IP information and intelligence, we have gone from paper files, via electronic files and searchable files including metadata, and now ending up in an AI-flavored IP environment.
IP Valuation is another area where machines never will beat the human mind, since its more complex than pure mathematics or statistics
AI stands typically for Artificial Intelligence, but maybe we should consider calling it Augmented Intelligence.
For that reason, if we look at how AI (no matter the abbreviation) will be used for IP Intelligence, there are several ways it can help. First of all, the vast volume of all patent documents, now counting to well over 100 million, is an inhuman task to browse. Here AI can help to filter and to arrange documents to surface the most relevant.
Language is another area since over 60% of all yearly patent documents are published in Chinese, Japanese or Korean. Machine learning and translation have latest years taken a big step and are becoming much more reliable, and a must today.
Still, regardless of how much AI we are using, a human (mind) has, in the end, look at the outcome and make decisions.
AI will not replace us; instead, it will put a higher requirement on us. If machines, for example, can do the routine tasks of legal research, more legal analysis will be done, and even more, human attorneys will be needed to apply general intelligence to decide how to use the results. Hence, we need more qualified attorneys, not less.
Therefore, AI should spell out Augmented Intelligence, since it will augment us, humans.
IP Valuation is another area where machines never will beat the human mind, since its more complicated than pure mathematics or statistics.
Valuation of IP, and in particular patents is something being discussed for many years, with many different approaches and ideas. The problem is that most of them (if not all) are based on theoretical economic or mathematical models.
However, extracting value from a patent will never be theoretical or mathematical, nor statistical, but instead a practical exercise. At best, you can find a new patent with automatic searches, but it will not be a sign of value until you do your homework. This is because value depends on so many different parameters;
Claim analysis – No automatic models consider the actual claim scope. All automatic and statistical models are based on citations, which in turn never consider claim scope. Those few considering claims, only looks at word length of claims, etc., which doesn’t tell you anything about the actual claim scope, and furthermore nothing about the value.
Full Family analysis – To fully assess the value of a (single) patent, you must look at all corresponding family members, if these exist. There is always a risk that a corresponding family member was rejected by prior-art or other cited document that could affect the validity of all other family members.
Context & Purpose – What do you intend to do with the patent? Patents can be used for many purposes. Selling a patent will yield one value (actual transaction price) and licensing the patent can give value over time. During an IPO, patents are attributed to some value in the investors' view. Using patents as collateral for getting a loan is another way of assigning value to your portfolio.
Owner – Value also strongly depends on who owns or controls the patent. One ownership factor is encumbrances. Having big cross-license agreements takes away large parts of value (i.e. addressable market). Another factor is knowledge how to commercialize the asset. Do you have the technical and market knowledge to assess the patent and argue for its value?
We are entering into an interesting and IP massive future with AI and autonomous vehicles IP filings skyrocketing but must keep in mind that quality always beats quantity when it comes to IP and the human spirit still has a competitive edge against computers.