Two years Into the Generative AI boom: how patent work is evolving
Last week, Sakari Arvela spoke at IP Service World in Munich, sharing his insights on the impact of generative AI in the patent industry. Drawing from a recent survey of IPRally’s admin users, he explored the current state of AI adoption within IP teams, typical use cases, and future expectations surrounding AI's impact on patent processes. In this blog post, Sakari summarizes his presentation. We're moving past the initial generative AI hype and into practical implementation, he concludes in his key takeaways.
Greetings from Munich
Last week, I attended the IP Service World fair in Munich. AI was everywhere, as the trend has been. One fellow Finn, working in a corporation summarized blatantly: “Too much AI!”. I can relate, as the daily work, at least in high IP positions, is filled with problems that are not solvable with artificial intelligence, and sometimes not even the good old non-artificial intelligence.
But some problems are assistable, if not entirely solvable, with AI. Particularly the closer one gets to the daily execution of patent processes. The patent industry is experiencing a significant transformation as generative AI technologies mature and find their way into daily workflows.
But what is the actual rate of adoption? How does the industry feel about the emergence of Generative AI tools? I decided to find out and did a small survey among IPRally’s admin users, mostly consisting of Heads of IP and IP intelligence and senior patent researchers, managers and attorneys. The survey consisted of closed and open questions. I shared the results in my presentation in Munich and will summarize them below too.
The state of AI adoption in personal work
The first question of the survey related to AI-assisted time spend. While generative AI adoption is growing, it's happening at a measured pace.
Most (78%) professionals use AI tools for 0-30% of their work time, however with a significant segment (22%) utilizing AI already for 30-50% of their time. Generative AI usage is only slightly lower than “any AI” usage, indicating that Generative AI plays a key role in AI adoption.
Claude summarized the results and open answers in the survey like this: “The adoption pattern suggests a "careful optimism" approach where professionals want to leverage AI tools for efficiency but are proceeding cautiously with implementation, demanding robust security measures and clear proof of accuracy before fully integrating these tools into critical workflows, as evidenced by the current low adoption rates (mostly 0-30% usage) despite high future expectations.”
What IP teams are using Generative AI for
The most established applications for generative AI in patent work included the following:
Patent search and review: About 85% of professionals are finding particular value in using AI for the initial screening of large document sets while maintaining human oversight for critical analysis. The amount corresponds to the share of customers that have enabled Generative AI features on their accounts in IPRally. Taking into account that search and review are the main functionalities of IPRally, this question worked as a good control question in the survey.
Patent analytics: The second place went clearly to analytics, which 50% of respondent teams were using Generative AI for. To me, this was a surprisingly high number, but unfortunately the survey didn’t reveal any details on exactly how it is used.
Patent classification, Portfolio Analysis and Drafting: These tasks shared third place, with 17% of respondents leveraging Generative AI for each. They were followed by a smaller share citing other tasks.
Claude’s summary of the closed and open answers was: “Patent searching and analysis tasks emerge as both the most currently adopted and promising use cases for Generative AI, with professionals showing particular enthusiasm for tools that can automate the initial screening of large document sets while maintaining human oversight for critical analysis and decision-making.”
Future expectations
I’m not sure how successful I was with the question-setting, but I tried to understand how ready our customers are to welcome more “intrusive” end-to-end Generative AI solutions into their realm with two questions:
- Belief in full automation of some routine, but still complex tasks (like novelty search) in 3-year time span and
- Most useful jobs to solve, given the fact that AI will likely not be perfect in the first iteration.
Interestingly, no respondent thought that full automation of tasks is completely out of reach. Most (71%) of the respondents think that this will “likely” or “inevitably” happen!
The most useful applications (with the confidence levels I gave) were:
- Novelty search (45%)
- Invalidity search (29%)
- Evidence of Use search (12%)
- Freedom-to-Operate search (10%)
The results and open answers highlight that there generally is high confidence and desire in substance-level work automation in the coming years. Notably, there is strong correlation between early adoption of Generative AI and belief in its industry-transformative potential.
All types of automated searches were found extremely valuable when the
- confidence level is proportional to the task and
- AI’s reasoning is made visible and verifiable, which builds trust in the tooling.
What was also evident from the open answers is that the word “automation” is understood and perceived in different ways: Positive (= an efficient helper doing the tedious tasks and allowing human oversight), Mixed (= an efficient competitor for humans taking away the fun tasks and increasing risks), and Doubtful ( = “Yeah … but no” / “Show me the proof first”).
Final thoughts
The patent industry might not be the fastest-moving in terms of AI adoption, but the change is steady and meaningful. Generative AI is playing a significant role therein. As AI capabilities continue to improve and costs decrease, we're likely to see accelerated adoption and more sophisticated applications.
By automating routine tasks with trustworthy and transparent solutions, AI will enable patent professionals to focus on higher-value activities that require human judgment and expertise.
As an example in my presentation, I explained how I used IPRally’s powerful AI search and Generative AI review features to conduct a real-life patent invalidity search and analysis case in minutes instead of hours and days and high costs.
For those in the patent field, the message is clear: while cautious implementation is wise, staying informed and beginning to integrate AI tools into your workflow is becoming increasingly important for maintaining competitive advantage in the evolving landscape of patent work.
I concluded my presentation at IPSW with these key takeaways:
- We're moving past the initial Generative AI hype cycle into practical implementation
- Early adopters see the most value and have the strongest belief in AI's transformative potential
- Both general-purpose and patent-specific tools for meaningful Generative AI adoption in IP teams are already available
- Substance-level patent task automation is actively developing
