By Ian Schick, PhD, Esq

The patent AI market has been flooded with copilots. That was a predictable first step. Copilots fit neatly into the old model: the human still runs the workflow, and the software helps the human move faster. In patent work, that usually means quicker drafting, quicker edits, and a somewhat more efficient path through familiar steps.

We do not think that is where the market ends.

By Ian Schick, PhD, Esq

Generative AI is making patent text cheaper to produce. That will not flatten the field. It will make quality filters, reputation, and review discipline more important.

That is the patent-world lesson from Thibault Schrepel’s recent article, The Matthew Effect at Scale: Attention Scarcity and the AI Output Explosion. His point is simple: when AI causes an explosion in output, the bottleneck does not disappear. It moves. The scarce resource becomes attention, and the institutions that allocate attention become more powerful.

By Ian Schick, PhD, Esq

Patent law firms do not need another reason to compete on drafting cost.

They need a better answer to the question sophisticated clients are increasingly asking: how does this filing strengthen the portfolio we are building?

Why copilots create prompt debt—and why agentic systems win

By Ian Schick, PhD, Esq

Copilots promise speed, but they often turn serious work into an endless loop of prompting, re-prompting, and cleanup—where you spend more time managing the model than moving the job forward. The hidden cost is “prompt debt”: fragile instructions that don’t scale, don’t repeat reliably, and quietly increase review and rework. This article argues that the solution isn’t better prompting—it’s a different architecture entirely: agentic systems that own the workflow end-to-end, verify outputs against standards, and pull you in only at the moments where human judgment actually matters.