During our journey together with many industry-leading corporations, we have encountered countless of questions, stances and challenges relating to adoption of AI at IP departments. As we saw many of these frequently repeating, we decided to dig a little deeper into them and offer some guidance for overcoming the most common hurdles.
This is an introductory article of a series of postings where we’ll dig into topics like why and where to consider AI in the first place, transparency of the solutions, how to build trust in the tools, and how you can evaluate if a tool fits your organization, people and processes. We hope that this series will help you with getting started faster and easier in your journey towards AI-powered patent research and management.
The fundamental question – why?
“If it ain't broke, don't fix it”, goes the saying. Is the IP system or some parts of it broken? Or at least inefficient? To put it more mildly – is there room for improvement in some of the repeating tasks and processes that IP people are carrying out? In our experience, many people inside the scene will answer yes to at least one of these questions.
The AI market size is expected to grow 40% (CAGR) annually during the 2020s, and we’ve already seen a profound transformation in many industries. But is AI the likely future of intellectual property management? It is very brainwork intensive after all.
We at IPRally are dealing with patent research, which is considered one of most the data-intensive and intellectually challenging sub-domains of IP. And believe us – yes, it is. Humans need and want assistance, and today we have the technology that can provide that. EPO and USPTO, among many others, are already relying on AI to help their examiners search in the immense and constantly growing body of prior art. In a report published by World Patent Information last year, 70% of EPO examiners believed that artificial intelligence and machine learning could provide a valuable support for patent search.
Besides that there is a problem to be solved and means for solving it, there are also practical reasons: in many organizations, the reality is that resources are getting smaller, or fail to keep up the pace with increasing data volumes and workload. One trend here seems to be a pressure to derive insights from IP data quicker and more frequently – both to the decision-makers (management) and to the R&D departments (engineers). If the IP department wants to respond to that demand, they need to be prepared for a change – or even lead the change.
Adopting a patent AI solution – where, who and how?
We’ve put together a 5-step guide going from deciding on the area to be improved to actual adoption of a patent AI solution in your organization. The key factors are:
- The purpose of the AI solution: What are your needs and how to identify them? Which practical problems do you wish to solve?
- Transparency and trust: How to choose the candidates to evaluate in practice? What factors are most important when building trust in an AI solution?
- The process and the users: Are you looking for quick wins that fit current personnel roles and processes or are you prepared to change them to maximize the value? There are no rights and wrongs, but it is good to know how far you are willing to go.
- Success and ROI: A true solution needs a combination of acceptance from users and long-term benefits. How can you evaluate these factors amidst all the other work that needs to be done?
- Commitment: If you decide to invest in it, use it. How to maximize the adoption and value and how to continue improving.
We’ll cover each of these topics in detail in the following posts in the coming weeks.
Until then, stay tuned!
.webp)
.webp)
