The USPTO’s AI Search Pilot Is Closed—What Should Applicants Learn from It?

When the USPTO launched its Artificial Intelligence Search Automated Pilot Program—ASAP!—the headline was easy to read as “the examiner now has an AI search tool.” By August 2026, however, the operational fact is different: the pilot is closed, and the Office stopped accepting petitions after June 1, 2026. What remains valuable is what the pilot tells applicants about the direction of patent examination.

First, Get the Current Status Right

ASAP! tested an automated search before substantive examination of certain original, noncontinuing, nonprovisional utility applications. Participants could receive an Automated Search Results Notice identifying potential prior art before regular examination began.

The USPTO later extended the program through June 1, 2026, then closed it. The agency’s current page states that no further petitions are being accepted and will be updated if the program is renewed or replaced.

Why the Pilot Design Was Strategically Interesting

The important feature was timing. Earlier visibility into potentially relevant prior art gives an applicant options before the first substantive Office Action: reconsider claim scope, file a preliminary amendment, evaluate whether continuation strategy needs adjustment, or decide that prosecution no longer justifies the cost.

The Automated Search Results Notice did not itself require a response. That distinction matters. The pilot was an information mechanism, not a substitute Office Action.

What Chinese Applicants Should Take from It

Chinese-originated U.S. applications are often filed on a tight translation or PCT-entry timetable. If automated prior-art tools become a more routine part of examination, defects that previously surfaced later may become visible earlier.

claims translated too literally from a Chinese drafting structure
generic functional language that does not emphasize the actual technical distinction
weak fallback positions in the specification
priority-dependent features that cannot safely be added later
terminology that obscures correspondence between the claims and the disclosed embodiment

The practical response is not to “write for the AI.” It is to make the application technically and legally coherent before filing so that earlier search visibility becomes useful rather than merely exposing a brittle claim set sooner.

AI Search Does Not Remove the Need for Human Prior-Art Analysis

An automated top-ten list cannot decide whether a reference is legally prior art, whether it anticipates every claim limitation, whether a combination would be obvious, or whether a reference is materially different once claim construction is done correctly. Search and legal analysis remain separate tasks.

This distinction is especially important for FTO and validity work. Better retrieval can reduce the cost of finding candidates; it does not eliminate the expert work required to characterize the candidates correctly.

The Longer-Term Signal

Even though ASAP! is closed, the pilot reflects a broader institutional direction: AI is being used to reorganize information and surface candidate prior art earlier. Applicants should expect examination workflows to become more computationally assisted, while the legal standards for novelty, obviousness, written description, and eligibility remain human-interpreted rules.

Takeaway

Do not market ASAP! as a currently open opportunity. Treat it as an instructive pilot. Its lasting lesson is that earlier machine-assisted search increases the value of filing-ready claim architecture, disciplined disclosure, and expert interpretation of the search results.