Four ways AI can help you find more relevant results
Navigating the AI landscape of patent intelligence.
There is an artificial intelligence (AI) arms race in many industries right now and patent intelligence is no exception. As various providers work to ensure that they are not being left behind, the user community is left sorting through all the noise to determine what kind of solution makes sense for their specific business or legal practice. This can be a difficult task, even with a reasonable understanding of AI capabilities.
Finding the right fit for your needs.
The critical question to guide your evaluation is a simple one: how do you define a “good” outcome for your patent search? Based on the answer to that question you may be able to find a vendor that is deploying the AI technology that best suits your specific needs. After all, there is no wrong answer to AI in patent intelligence, just different possibilities, applications and outcomes.
Our perspective is that the best solution solves productivity problems and satisfies the needs of businesses to scale by making it easy to combine information from many sources – not just from semi-structured documents, such as patents. Only by allowing users to seamlessly search across multiple sources with one search strategy can you truly provide context and deep understanding in a meaningful way. Those sources might include patents to be sure, but also news, research papers, third party data and importantly, internal documents.
It’s time to look beyond Boolean.
The old way of searching by using key words or Boolean strings falls short. Using these methods, your search must be unrealistically precise and the opportunity to miss relevant patents is high. This reflects the fact that it is up to the searcher to know and list all the semantically relevant search terms in their search strings when building their strategy. Searchers may compensate for this by creating a broad search strategy, but this often yields so many results that help from an outside expert is required to sort through it all. For these reasons, many platforms require substantial training or multiple steps to get searches right, and outside experts are often needed to complete even basic searches.
As a result, one search project requires many individual searches to gather all the relevant information which then needs to be combined and summarized. Faced with a cottage industry of sorts, many companies end up limiting what initiatives “merit” the investment of a patent search to control costs, rather than maintaining active searches across their portfolio. Having to pick and choose can create strategic blind spots, leaving executive decision-makers with rough educated guesses in place of confident data-driven decisions.
It’s plain to see that AI-driven search is preferable to manually building, curating, and maintaining search queries for most users. But not all AI models are created equal when it comes to patent searching. Some key features that we think are important include:
Search concepts not words – A search process that focuses on semantic searching will yield more relevant results. This means that you can search based on meaning as opposed to every possible word and/or synonym for your invention. This can increase the efficacy of your search by making it easier and faster to find relevant data by matching your claims to other patent claims based on context and meaning. This can help ensure your search returns a complete result set, including those results that may not be obvious, but are similar.
Apply a single search across different document types – Because intelligence data comes in many formats, you need a platform that excels at dealing with searches across unstructured data at scale. This allows users to leverage a single search strategy to pull information from many document types, both internal and external. This enables efficient similarity searching across multiple document types based on meaning rather than a series of keywords. This is a more flexible system that does not require documentation of explicit relationships and structure.
Leveraging scoring algorithms to measure relevance – While any system should give you control over depth and breadth of your search, some platforms allow you to calculate the semantic similarity between concepts across patent claims. In other words, a system that implicitly forms relationships between patents. Similarity scores like this can make it much easier and faster for a researcher or attorney to identify the most relevant information for review.
Extend and preserve your intelligence – Anyone responsible for an IP portfolio understands that a limitation of conventional patent searches is their “snapshot” nature. The intelligence produced by the search represents a single point in time, contrary to infringement risk. A robust mechanism for maintaining active searches that identify changes in risk over time is paramount. This includes maintaining a search as “active” with the ability to notify you of any relevant changes in the market or technical landscape. It also allows you to preserve relevant information on a research project to support ongoing collaboration within your team and across other disciplines. Importantly, by preserving your intelligence within the application, you significantly reduce departure risk associated with individual expert searchers.
The benefits of a one-stop AI shop
The beauty of AI for patent searching is that when leveraged appropriately, you can solve all core pain points with one powerful yet intuitive application. You should be able to find all the semantically relevant data you seek with one search strategy. That search strategy should be simple to construct, and you should be able to leverage that search across multiple types of data. Importantly, results should be organized and presented in such a fashion as to enable collaboration and help you make better business decisions faster. Any patent searching application that leverages AI should be able to do all those things.
Ready to unlock the power of advanced patent searching?
Interested in streamlining your patent search and uncovering valuable insights? Whether you're an inventor, researcher, or IP professional, let’s discuss how FluidityIQ can give you a competitive edge.