Following the USPTO’s high-profile actions in Desjardins, the next question was how that guidance would play out in ordinary PTAB appeals. Ex parte Carmody provides an early answer.

The decision itself is short. The analysis is sparse. And that is precisely why it matters.

The Holding, in Brief

In Ex parte Carmody, the PTAB reversed a §101 rejection of claims directed to an AI-based system for marketing and sales orchestration. The Board agreed that the claims implicated abstract ideas at Step 2A, Prong One, including organizing human activity and mental processes.

But at Step 2A, Prong Two, the Board concluded that the claims integrated those abstract ideas into a practical application. The reason was straightforward: the claims recited an improvement in training of models for use by a recommendation engine to generate useful orchestrations, and not just the use of AI to achieve a business outcome.

On that basis, the Board reversed the §101 rejection.

Why the Analysis Is So Short

What stands out most in Carmody is not what the Board says, but how little it feels the need to say.

There is no extended Alice analysis. No deep dive into mental steps. No Step 2B discussion. Instead, the Board identifies the modular ML architecture, cites Desjardins, and moves on. This brevity appears intentional.

After Desjardins was made precedential and integrated into the MPEP, the Board seems to no longer treats AI eligibility as an unsettled eligibility question when claims clearly recite technical improvements to machine-learning systems themselves. In that context, Carmody reads less like a hard-fought appeal and more like a routine application of settled policy.

What Carmody Tells Us About §101 Today

Carmody reinforces several practical points that applicants and practitioners should internalize:

  • The PTAB will readily acknowledge abstract ideas at Step 2A, Prong One.
  • The decisive question is whether the claim clearly recites how the AI system itself is technically improved, with support in the specification.
  • Improvements to ML training, architecture, modularity, or deployability are strong candidates for eligibility under Step 2A, Prong Two.
  • When those features are present, the Board may resolve §101 quickly and without resort to Step 2B.

Practical Takeaways

For AI-related patent claims:

  • Focus on how the machine-learning system works, not just what it decides
  • Claim specific training signals, model structures, and interactions
  • Highlight architectural choices that improve scalability, updateability, or performance
  • Ensure the specification clearly explains why those choices matter technically

Carmody does not expand §101 doctrine. Instead, it confirms that Desjardins is already shaping day-to-day outcomes at the PTAB.

****

Subscribe to get updates to this post or to receive future posts from PatentNext. Start a discussion or reach out to the author, Matt Carey, at mcarey@marshallip.com or 312.474.9581. Connect with or follow Matt on LinkedIn.