Patent research: Why we added Generative AI and Boolean to IPRally

While AI has been the topic of conversation for almost everyone everywhere these past few months, it has been our sole focus for the past 5 years. And while everyone has discussed and tested different forms of AI these past few months, we have (also) been focused on a patent search methodology that has been around since the 19th century, namely Boolean.

What's the point with Boolean in IPRally? 

You may not be aware that the Boolean methodology – which is now an integrated part of IPRally – has been around for more than 100 years. It has, and it has been the dominant way to search for information in databases for as long as most of us can remember. Hence, it may seem counterintuitive for a company like IPRally, which has developed cutting-edge patent search AI, to spend time and effort on Boolean. Nevertheless, that is what we have done and Boolean searching is now available alongside our proprietary graph search in IPRally! 

So why have we focused on this? As always, it’s because it will help our users do more, better. First, classical search is still the right tool for basic bibliographic data searches. Second, having the possibility to use Boolean for searching, next to or combined with our Graph AI, opens up possibilities that previously required multiple tools. Now users can also, in IPRally:

• Search and find patents using traditional patent search criteria, like assignees and keywords, using standalone Boolean just like you would expect it to work,
• Filter graph AI-based search results more effectively using Boolean filters to combine the search methodologies,
• Search inside or filter collections of patents, like earlier AI or Boolean-based search results

In short, you will now be able to do 95% of your searching and monitoring in IPRally, and make 1+1 equal more than 3 by combining cutting-edge AI and traditional proven methods. We are very excited about this!

Ask AI – AI-assisted patent review 

The Graph AI in IPRally is now very effectively understanding technical information, describing the technical information and finding or classifying patents using these descriptions. Until now, once the information has been found, the patents have had to be reviewed and assessed through processes that are fairly manual. Yes, we have been ranking results by relevancy, providing contextual highlighting, automating classification etc to make this easier, and we will continue to do this. But with our latest development, we are really making a difference and will change the way that this is done forever. 

We have now launched Ask AI, our first iteration of generative AI, which allows you to simply interact with IPRally as you would with a human patent professional, to help you narrow in on what’s relevant much faster. The difference is that the AI patent professional has been trained in every field of technology and is capable of internalizing the whole content of every patent in detail in the blink of an eye. 

In IPRally you're now able to, in a chat-like prompt:

• Ask detailed questions about your search results (e.g. is this technical solution or material described in the patent)
• Summarize patents in a custom way (e.g. tell me the advantage of the invention in 3 bullets)

The feature has been built in a flexible way that allows you to be creative in how to take advantage of it. The initial version allows the scrutiny of one patent at a time, but we are already building a version that scales to hundreds of patents in real-time. For example, analysis of large patent portfolios in a completely new way is just around the corner. 

In the near future, you will also be able to: 

• Filter your results (e.g. show only patents that talk about this material)
• Compare patents (e.g. what is the difference between patent x and y) etc.

The possibilities are (almost) endless. Many will have tried similar things in, for example, Chat GPT, already, and have a good idea of how this works. But by now most of us also know that Chat GPT is no place to search for information. IPRally will provide these possibilities fully integrated with the search tool you already know, using the curated data generated in IPRally, enjoying the security provided in IPRally. 

The Ask AI feature has been described in more detail in our Help Center.

Hence, every aspect of the search and review process – from learning a concept and defining a search to the review and assessment of relevance – will now be greatly enhanced by AI, and our users will have to find other things to do with the time they free up. A lot more will follow about these new elements in IPRally, but if you are curious to know what the new features would do for you, let us know by contacting your regular point of contact or signing up for a free trial.

So, the past and the present are converging now, and it’s happening in IPRally. Join us on the journey and see what the future looks like!