How to Streamline the Patent Process with AI

For patent filing and prosecution strategies to support global business realities, patent practitioners must negotiate a complex landscape of geographical and cultural differences. Questel’s Benoit Chevalier explains how artificial intelligence (AI) tools for IP preparation and prosecution can deliver guidance and support through the patent process.   

As much as globalization has boosted the desire for a unified, interoperable, and harmonious ‘global’ patent process, there remain profound cultural differences between the different patent jurisdictions worldwide. From the US to Japan via the EU, patents remain territorial by nature, much like the legal systems that regulate them.   

From drafting robust patents to managing patent prosecution formalities, patent practitioners must employ great skill to navigate this protean landscape. Fortunately, new AI-assisted patent preparation and prosecution tools can offer them much-needed support through the patent filing and prosecution process.  

Why Use AI to Optimize the Patent Process?

The potential for AI in the IP sector has been discussed for some time, but it’s only in the last few years that we have seen the technology come into its own. From AI drafting tools to support the patent application process, such as our Patent Drafting Software with AI, to AI-assisted Patent Prosecution Workflow tools, such as Qthena, the technology now has the requisite expertise to research, draft, validate, and adapt its output to each legal jurisdiction.  

This includes harnessing the power of natural language processing (NLP) and deep learning to inform the content and approach to the nuances of each jurisdiction’s patent process. For example, by informing how patent claims should be drafted according to the case law or examination practices in each region, as well as how to process the vast volume of documents, reports, and systems that patent practitioners interact with every day.  

• Can AI draft patents?  

While we do not believe that AI can—or should—replace human expertise, we do believe that it can assist practitioners in navigating global patent complexity more effectively, while also streamlining their patent search, drafting, and prosecution workflows.  

We estimate our Patent Drafting Software with AI saves clients a mammoth 40% of their drafting time by producing a first draft in minutes, whether users upload a simple claim tree or a more detailed invention disclosure.  

Patent practitioners are in control throughout the process with each section employing its own designated AI model to generate editable titles, summaries, problem statements, and the initial technical problem, as well as potential variations.  

As importantly, the collaborative tool uses specific patent templates for each jurisdiction—including autocomplete, renumbering, diagram generation, and recommendation/suggestion features that consider relevant case and law and PTO examination guidelines—to deliver a compliant and robust patent application at the end of the process for submission to the relevant patent office.   

Along with information about relevant case law and examination guidelines (EPO and USPTO), our AI patent drafting software features a wide range of legal resources, insights, and data, helping users identify potential antecedent problems in the claims and anticipate requests by examiners for amendments. The full kit of useful tools includes the ability to check instantly for unclaimed matters, check antecedent basis issues, improve numerals management and indexing, and generate interactive generalizations—and we have even more features in the pipeline!  

• How can AI assist patent prosecution?  

Alongside support for the patent application process, assisting patent prosecution workflows is low-hanging fruit for a technology as advanced as AI, given its ability to reconcile data, documents, and systems “under the hood” to ease the user’s experience and enable them to focus instead on the task at hand.   

We know that patent teams expend considerable time, effort, and oversight simply jumping between tools and files. By integrating the Qthena Patent Prosecution Workflow tool into our solutions, including our Equinox IP Management Software, we can help clients overcome these time and information challenges.   

However, it’s not simply that Qthena centralizes all the information users might need for the patent process into a single interface—or that it automatically organizes, reconciles, and OCRs relevant data and documents, from invention disclosures to prior art, including non-patent literature. Crucially, it also employs large language models (LLMs) to analyze multiple documents simultaneously: office action, prior art, product presentation, and patent application.  

By using this “AI assistant” and instructing it to generate responses in line with the current jurisdiction, a patent attorney can spot lack of clarity, challenge examiner arguments, work on amendments, and define strategy in record time. The Qthena models have been meticulously trained to generate inventive step arguments perfectly aligned with EPO practices. Our solution extends its capabilities to accommodate standards of other patent offices, including the USPTO, ensuring global applicability.