By Michael Borella –

The U.S. Patent and Trademark Office has extended its Artificial Intelligence Search Automated Pilot Program (ASAP!) until June 1, 2026.  Originally slated to close to new petitions on April 20, 2026 (see “USPTO Announces That It Has Turned the Corner on Unexamined Application Backlog“), the Office has expanded both the duration and the scope of the program “in order to obtain more information to evaluate the effectiveness of the program.”  At the same time, the USPTO has waived the program’s fee. 

By way of background, ASAP! is a pilot program under which an original, noncontinuing, nonprovisional utility application filed under 35 U.S.C. § 111(a), and including Track One applications, is subjected to an automated, AI-driven prior art search prior to examination.  The results are delivered to the applicant as an Automated Search Results Notice (ASRN), which lists up to ten documents ranked by relevance, along with a search string that can be entered into the Patent Public Search tool to retrieve the cited U.S. patents and publications.  The ASRN does not require a reply.  Nonetheless, an applicant who receives an ASRN may elect to file a preliminary amendment, request deferral of examination, or expressly abandon the application to recover the applicant’s search fee and any excess claims fees.  The examiner eventually assigned to the application is free to use or disregard the ASRN when conducting their own search.

Three aspects of the extension are worth highlighting.  First, the USPTO has increased the program’s target from approximately 1,600 applications (200 per Technology Center) to at least 3,200 applications (400 per Technology Center).  Second, the USPTO will continue to accept petitions until the earlier of June 1, 2026 or the date on which each Technology Center has been docketed with at least 400 applications accepted into the program.  Third, and perhaps most consequentially for cost-sensitive applicants, the $450 petition fee has been waived for any petition filed on or after March 23, 2026.  As before, the petition must be filed electronically in Patent Center on the filing date of the application, and the application must be filed in DOCX format.

The fact that the USPTO is extending the program and eliminating the petition fee is telling.  It suggests that uptake until now was uneven across Technology Centers or otherwise insufficient to support any evaluation the USPTO would presumably like to perform on the program’s efficacy.  Indeed, as of the time of this writing, only 169 petitions have been filed across all the Technology Centers, and just 76 of these have been granted.

The fee, while modest in absolute terms, is a real deterrent for applicants who see only a questionable benefit in return.  Removing that fee is an acknowledgment that the initial cost-benefit balance was tilted in the wrong direction.  Further, the applicant would presumably want to review any references cited by ASAP! for relevance, which would incur its own administrative costs in terms of time and/or attorney’s fees.  Thus, even without the fee, the program is not truly “free.”

 On the balance, the fee waiver and the expanded intake should make ASAP! a more attractive proposition for applicants who were on the fence.  The upside is a free automated search that may (or may not) surface relevant prior art of which the applicant’s was unaware.  For applicants filing in technology areas where the USPTO’s AI search tool has access to a deep and well-classified corpus (e.g., software and electrical arts), the ASRN may prove genuinely useful. 

One further development is worth flagging.  At last week’s University of San Diego Patent Law Conference, it was reported that Director Squires announced a version 2 of the AI search tool that will be released in the near future.  The principal improvement, as described by Squires, is that the updated tool will actually search the claims of the submitted application.  Apparently, the version currently in use under the pilot has been searching based on the specification alone.  That is a surprising admission, and a meaningful one.

Any prior art search should include and focus on the claims, because the claims define what is actually being examined for novelty and non-obviousness.  A search drawn only to the specification will tend to surface prior art relevant to the disclosed embodiments as a whole, rather than art that maps to the specific limitations an applicant has chosen to claim.  Thus, the usefulness of the current program is questionable on its face.  But if version 2 delivers, the quality of the ASRN should improve materially.