How Patent Prosecution Strategies and Behaviors Are Evolving Amidst a Changing USPTO

The USPTO is not the same office it was two years ago. Since January 2025, shifts in leadership, policy emphasis, and internal guidance have changed how applications are examined and how prosecution outcomes are determined. Some of these changes have generated significant attention in the patent community, while others have been more subtle. All of them have real consequences for practitioners and applicants trying to build and protect patent portfolios today.

I recently attended IPWatchdog LIVE, one of my favorite events in the IP calendar and a place where I always come away with new perspectives from practitioners, in-house teams, and thought leaders across the industry. I had the pleasure of moderating a panel on how prosecution strategies are shifting in response to the many changes underway at the USPTO, and I wanted to summarize my key takeaways here, along with the data graphics we pulled for the discussion (as promised to the audience). 

This post examines five of the most significant developments at the USPTO and how practitioners are adapting their strategies in response. Where possible, we have examined data from Juristat to distinguish between what has been announced and what is actually happening in practice.  

1. Examiner interviews: A shorter runway

Two policy changes have significantly narrowed the role of examiner interviews in prosecution: 

• First, the USPTO retired the AFCP 2.0 program in December 2024, eliminating a widely used mechanism for securing examiner consideration after a final rejection without the cost and delay of an RCE.

• Second, a new Performance Appraisal Plan launched in October 2025 now credits examiners for only one interview per application, reducing their incentive to engage in follow-on discussions.

The data reflects this shift clearly. The share of second-plus interviews dropped from roughly 33% before the policy change to a post-policy steady state of 28 to 29 percent, a structural decline of 4 to 5 percentage points. The sharpest single-month drop occurred between September and October 2025, and the rate has remained suppressed ever since.

Practitioners are reacting by front-loading their preparation. Because the opportunity for a follow-on conversation is less likely to materialize, the first interview carries more weight. Practitioners are investing more heavily in pre-interview preparation, using examiner analytics and art unit data to sharpen their positions before walking in the door. There is also a growing tendency to evaluate, early in prosecution, whether pursuing an interview is worth the effort at all, or whether the examiner data suggests it is better to move directly toward appeal.

For resource-constrained applicants in particular, the loss of AFCP 2.0 and the reduced incentive for repeat interviews has shortened the runway considerably and made appeals a more viable strategic option at an earlier stage of prosecution.

2. Section 101: Policy signals vs. Examiner reality

The USPTO has taken a notably more applicant-friendly stance on subject matter eligibility over the past year. An August 2025 memorandum directed examiners to apply a higher evidentiary standard before issuing eligibility rejections, with a reminder that close calls should resolve in the applicant's favor. The November 2025 precedential decision in Ex Parte Desjardins tightened constraints on overbroad abstract idea rejections, particularly for AI and software claims. The MPEP was updated in December 2025 to incorporate that decision directly. The USPTO also introduced Subject Matter Eligibility Declarations, or SMEDs, as a formal tool for applicants to submit evidence of eligibility.

The question practitioners are asking is whether any of this is actually changing examiner behavior. The honest answer is: it depends on where you are practicing.

Art Unit 2120, which covers AI, machine learning, natural language processing, and knowledge representation, shows the clearest post-policy response. Rejection rates there ran between 68 and 74 percent from January through July 2025, then fell to roughly 60 to 63 percent from December 2025 onward, a decline of 10 to 13 percentage points that tracks closely with the policy timeline. This is the most credible evidence that the USPTO's guidance has had a real examiner-level effect.

The picture looks very different in business method art units. AU 3690, which covers data processing for financial, business, and e-commerce applications, ran between 82 and 87 percent throughout virtually the entire period, showing almost no response to the policy changes until a modest dip to around 79 percent in early 2026.

Practitioners in technology areas outside of software and AI are generally not seeing any meaningful change. In mechanical engineering, manufacturing, and medical device art units, overall rates remain essentially flat, though the baseline there was never as acute.

The strategic implications are significant. Desjardins-based arguments and SMEDs are most likely to gain traction in AI and software cases. Practitioners in business method and financial technology art units should plan for continued resistance regardless of the policy signals, and those with strong cases are increasingly taking a harder look at appeal as a first-line strategy rather than a last resort.

3. Discretionary IPR denials: The new gatekeeping regime

In 2025, the USPTO significantly tightened its gatekeeping over post-grant challenges. Director Squires took personal control over whether to institute IPRs and PGRs, replacing detailed panel opinions with concise summary notices. Discretionary denial rates climbed from roughly 35 percent to over 57 percent in just a few months.

More recently, Director Squires issued a new memo adding three additional factors to the discretionary denial analysis: whether the patent owner manufactures in the United States, whether the challenged patent covers a domestic product, and whether the petitioner is a small business. The memo applies immediately to pending proceedings.

These changes are reshaping how practitioners think about post-grant strategy from both sides of the table.

For patent owners, the environment has become considerably more favorable. Challenged patents are surviving institution at a much higher rate, and the shift toward summary denials makes it harder for petitioners to understand or challenge the reasoning. This environment is particularly beneficial for smaller patent owners and, it should be noted, for non-practicing entities as well.

For challengers, the calculus has shifted dramatically. Practitioners are reporting that ex parte reexamination has become a more common fallback when IPR petitions are denied, even though reexamination is a less efficient and less predictable vehicle for challenging validity. The data supports this trend.

Reexamination filings that ran between 20 and 55 per month throughout 2024 accelerated sharply in late 2025, reaching roughly 95 to 115 per month by November and December 2025. That trajectory suggests practitioners are treating reexamination not as a preferred tool, but as a necessary substitute when IPR is no longer a realistic option.

The new "Made in America" factors add another layer of complexity. Practitioners advising clients on whether to file a petition now have to assess not just the prior art case, but also whether domestic manufacturing considerations or entity size will factor into a discretionary denial decision. Building a strong petition on the merits may no longer be sufficient.

4. Fee increases: Real pressure on prosecution budgets 

Effective January 19, 2025, the USPTO implemented its most sweeping fee restructuring in years. Front-end fees increased roughly 10 percent across the board, but several specific fee increases have had an outsized effect on prosecution strategy:

• The fee for a first RCE increased from $1,360 to $1,500, a 10 percent jump. For second and subsequent RCEs, fees climbed from $2,000 to $2,860, a 43 percent increase.

• New IDS size fees were introduced.

• PTAB petition fees for IPR and PGR rose approximately 25 percent.

• New surcharges were added for continuations filed more than 6 or 9 years after the earliest benefit date.

• Excess claim fees doubled for large entities.

Taken together, these changes are creating real pressure on prosecution budgets, particularly for companies managing large or global portfolios.

Looking at the data, the RCE fee increase has had a measurable two-phase effect on applicant behavior. In the first three quarters of 2025, applicants reduced prosecution rounds but kept filing RCEs on pipeline cases, temporarily pushing the RCE-to-office action ratio up. By the fourth quarter of 2025 and into 2026, both absolute RCE counts and the RCE-per-OA ratio fell to their lowest levels in the dataset. Applicants are increasingly choosing to abandon, appeal, or resolve cases rather than pay higher fees to continue prosecution.

Practitioners are responding to the overall fee increases in several ways. Some say that second RCEs have become rare in their practice. Continuation strategy is being actively managed around the 6 and 9 year surcharge thresholds. Teams are pruning portfolios more aggressively, making harder calls earlier about which cases are worth continuing to invest in. International filing programs are being scaled back at some organizations, with a harder look at where foreign protection is truly necessary.

The fee increases have also accelerated investment in data-driven decision-making. Practitioners are using examiner analytics, art unit benchmarks, and allowance rate data to make more targeted decisions about where to invest prosecution resources and where to cut losses.

5. The technology advantage: Data and AI as strategic tools

Perhaps the most consistent theme emerging from conversations with practitioners navigating this environment is the growing role of data and AI in prosecution strategy.

With fewer interview opportunities, higher RCE costs, and a more complex post-grant landscape, practitioners are relying more heavily on examiner-level analytics to guide decisions. Which examiners are responsive to certain types of arguments? What is the interview lift in a given art unit? What is the likelihood of success on appeal versus amendment? Practitioners are increasingly answering them with data.

AI tools are also changing the economics of patent prosecution in ways that are starting to offset some of the fee pressure. Prior art searching, competitive landscaping, claim drafting assistance, and potential infringement analysis are all areas where AI is extending what practitioners can accomplish within a given budget. The tools are not perfect, but the consensus among practitioners actively using them is that the advantage is real and growing.

For in-house teams under budget pressure, the combination of data-driven decision-making and AI-assisted workflow has become less of a competitive differentiator and more of a baseline requirement for running an efficient patent program.

Looking ahead

The USPTO is still mid-reset. The policy changes described here are real, their effects are showing up in the data, and the strategic implications are still unfolding.

The fee structure has changed the economics of prosecution in ways that will take time to fully absorb. The IPR landscape is shifting in real time with new discretionary denial factors that have yet to be fully interpreted. And the 101 reforms, while meaningful in some technology areas, have not resolved the underlying uncertainty that has existed for years.

The practitioners navigating this environment most effectively are the ones who have stopped waiting for the dust to settle and started treating prosecution as a data-driven discipline, where examiner behavior, policy trends, and portfolio economics are analyzed together to inform strategy.

The USPTO may keep evolving, but the advantage will belong to the teams that can turn data into strategy. Juristat helps IP teams do exactly that—see how with a Juristat demo.