As in all other aspects of business, AI is dramatically changing how IP professionals work. Tools that felt experimental a year ago are now being used in client work, boardroom presentations, and licensing negotiations. But not everyone is adapting to this change at the same pace. A poll taken at the start of our recent webinar, “How AI Is Transforming IP Strategy Across the Patent Lifecycle,” revealed that most attendees were using AI in only some parts of their workflows, with a significant share not yet using it at all. The webinar panel featured Gene Quinn, patent attorney and founder of IPWatchdog; Matt Moffa, partner at Perkins Coie LLP, registered patent attorney, and IP litigator; and Dr. Tim Pohlmann, Managing Director Americas at LexisNexis Intellectual Property Solutions. The poll results from the webinar did not surprise Tim. He had heard similar responses at the Global Standards and Leadership Conference 2026, a Silicon Valley event that concluded a few days earlier. “I’m completely sold that AI got so much better,” he said, “but it takes time before we have integrated, tested, and brought this to market, and for people to learn how to use it.”The conversation during the webinar covered topics including where AI is proving its value today, why general-purpose tools fall short for specialist work, and what transparency actually means in practice.
For Matt, who is currently an active participant in the space, the clearest early payoff has been in prior art and patentability analysis. “Everybody I’ve talked to at the firm said we’re never going back,” he noted. The ability to synthesize large volumes of prior art, surface relevant references, and generate preliminary conclusions in a fraction of the time it once took has changed how his team works on invalidity analysis. Beyond search, three areas came up repeatedly as places where AI is delivering real value. Portfolio analysis. AI allows IP teams to assess which patents in a portfolio are most strategically relevant, identify coverage gaps, and model the financial implications of maintenance decisions without pulling engineers into the process. Tim described how the AI assistant, LexisNexis® Protégé™ in PatentSight+™, can map information from a company’s website and strategy documents against its patent portfolio to show where alignment exists and where it does not. “Sometimes press announcements don’t correlate at all with your patent filing behavior,” he said. “That’s quite impressive for business leaders to see.” Competitive intelligence. AI can now aggregate information across a competitor’s patent filings, product announcements, and trade publications and surface insights that would previously have required a dedicated research team working over weeks. Matt noted that teams can now ask a question and receive an analysis drawing on current product manuals, journal articles, and investment-grade information almost immediately.Communicating findings to leadership. One underappreciated capability is AI’s ability to translate complex patent data into language and visuals that non-specialist stakeholders can understand. “When the decision comes down to where we’re going to spend our money, you’ve got to present it to somebody who’s a business decision maker in a way they can appreciate,” said Matt. As one IP analyst put it in his description of working with Protégé in PatentSight+, turning dense IP analysis into clear, executive-ready output is increasingly a core function of patent analytics work.
A recurring theme in the webinar was the gap between what general-purpose AI models are designed to do and what patent professionals actually need.General models are optimized to answer the questions most people ask. In a specialist field like patent law, that creates real problems. Gene, who moderated the discussion, shared an example from his own work. He asked a leading general-purpose AI to explain how it is possible to adopt a technical standard without infringing a standard essential patent (SEP). The model pushed back, insisting that adopting the standard amounted to infringement. In most general contexts, that answer would pass. In IP practice, however, it reflects a misunderstanding that could impact million-dollar decisions in a licensing negotiation or a court proceeding.“If you get that wrong in our space, it’s catastrophic,” Gene said. “Carefully curated data from people who understand that precision matters in critical business functions is what you need.” Tim drew a clear distinction between what general AI tools offer and what a purpose-built tool like Protégé in PatentSight+ provides. Patent-related queries in Protégé are routed exclusively through LexisNexis’ curated, harmonized global patent database. The tool does not search the open internet for patent information. “Any IP and patent-related question is never going to go out of the system into the internet,” he said. “We know who owns what, we know legal status, and we have value metrics that our customers have been trusting for years.” Beyond data quality, there is the question of how an AI system processes a query. Purpose-built systems encode the institutional knowledge of how patent analysis is actually done. “We define the process in our system very specifically for each use case, because we know how customers typically do it,” said Tim. “We know the best practices.” That means the system reasons through problems the way a trained analyst would, following validated steps rather than improvising.
One concern that came up throughout the conversation was the black-box problem: AI systems that produce outputs without explaining how they arrived at them. For IP work, this is more than just an inconvenience. Protégé was designed specifically to solve this challenge. Every step the system takes, and every assumption it makes, is documented in its responses, so users can trace exactly how a result was reached, modify the approach if needed, and share the methodology with others who need to understand or validate it. “Results are reproducible and transparent; it’s not a black box,” said Tim. Of course, being an AI system that tries multiple approaches, there may be slight parameter-dependent variations in its responses. Matt reinforced why this matters for litigation. When AI-generated analysis needs to withstand legal scrutiny and be presented by an expert witness, the expert must be able to explain what they did in a reproducible way. An AI system like Protégé in PatentSight+ with documented methodology and a track record against known outcomes is far more defensible than one that cannot show its work, regardless of how accurate it may be.
Standard essential patents represent one of the most technically demanding areas of patent practice. Determining whether a patent is truly essential to a standard, and what that means for licensing negotiations, requires handling large volumes of dense technical and legal material. It is also an area where AI has made striking progress.Tim shared results from his team’s AI-powered SEP mapping work: approximately 98% alignment with court-determined essential patents, 95% consistency with patent pool data, and 94% similarity with manually created claim charts. “A year ago, everyone was convinced this space was too complex,” he said. “That has changed.”But he was careful to frame what that progress does and does not mean. Essentiality is one question. Whether a patent is essential to the specific implementation of a standard a particular company uses is a different, harder question. And even where essentiality is established, questions of portfolio value, comparable licensing, and what weight to assign different patents in a cross-license negotiation remain deeply complex. “You have to be an expert in the space,” said Tim. “It doesn’t mean you can’t use AI, but you just have to understand what it can give you and what the limits are today.”Matt’s position was one of calibrated confidence. He regularly uses AI-generated SEP analysis to advise clients and build negotiating strategies, but applies additional validation before presenting it in a courtroom context. “Relying on it in the sense of taking it to court, I’m still very cautious about that,” he said. He noted that as the methodology becomes more established and reproducibility is demonstrated through real cases, the bar will become easier to clear. But for now, secondary validation remains important.
The webinar touched on a change that goes beyond individual tools or workflows. AI is beginning to alter how IP teams are structured, what skills are valued, and how patent strategy connects to broader business objectives. Tim described clients using Protégé in PatentSight+ to compare their filing behavior against their stated strategy and product roadmap and to model cross-licensing positions with far more precision than a traditional approach based on claim count alone would allow. “We will see a lot more or different transactions,” he predicted. “Some business models will be improving because there’s more information on potential infringement or evidence of use.” Matt echoed that, noting AI’s capacity to surface portfolio value that analysts might otherwise overlook. “It might be something you didn’t realize was so good about your portfolio, and you might have let those patents go, or let them go for cheap.”Gene closed by asking Matt for one piece of advice for individual practitioners who have not yet integrated AI into their practice. His answer was direct: “AI is table stakes now, at least because the person on the other side of the negotiation is going to be using it too. If you’re not and your competitor is, you’re going to fall behind.” Tim put it as plainly as possible: “Your job is going to be replaced by someone who knows how to use AI better than you do.” The human expert remains essential for validating outputs, applying judgment, and asking the right questions. But expertise alone, without AI as part of the workflow, is becoming harder to sustain as a competitive position. For those just starting out, the practical advice was simple: pick one workflow you know well, apply AI to it, and pay attention to what you learn.
Your AI assistant for fast, decision-ready patent insights that drive strategic business decisions.
