There’s a lot of debate around whether AI can one-shot patent tasks or whether human-in-the-loop will always be required. Based on what our team has achieved, I believe it will be closer to one-shot than not.

We’ve built an extensive AI pipeline for responding to office actions. We’re not the first to release an OA response tool, but after seeing a few, I believe ours is among the best.

The reason is the pipeline.

High-quality office action responses require relentlessly analyzing the weakest parts of an examiner’s argument. Doing that well, without hallucinations, requires giving the AI clean, tightly scoped data at exactly the right place and time.

We’ve organized a network of law firms comprising Big Law or former Big Law partners, who report 30–80% time savings while achieving the same or higher quality using our tool. And those savings should only increase as the tool improves.

Because of those time savings, firms in our network have reduced their fixed fees, thereby passing meaningful cost reductions to their clients.

For those of us who have been in this industry for a while, a law firm proactively reducing their rates is practically unheard of.

At a recent conference, in-house practitioners were asked about their budgets. Most were not expecting those budgets to shrink. My expectation is that spending shifts rather than disappears.

If routine prosecution becomes significantly less expensive, companies can invest more in work that has historically been difficult to justify: acquisitions, licensing, competitive analysis, portfolio optimization, and understanding which patents actually support products and business strategy.

Patent filing costs may fall, but the rate of patentable innovation is not guaranteed to increase. Instead, organizations can spend more time extracting value from the portfolios they already own and measuring IP the way engineering and sales teams increasingly measure themselves: by business impact and ROI.

Some of that work will naturally flow back to outside counsel.

Rather than spending most of their time drafting office action responses and applications, prosecutors may become prosecution powerhouses—handling substantially more work in fewer hours—or more strategic advisors focused on readiness, competitive intelligence, portfolio analysis, and better IP decision-making.

I’m not convinced AI reduces the amount companies spend on intellectual property. I think it changes where they spend it.

I like the way Jensen of NVIDIA frames it: separate the task from the purpose of the job.

In patents, writing a patent application or responding to an office action is a task.

The purpose of the job is protecting a company’s intellectual property and helping the business create economic value from its innovations.

AI will continue to automate more of the tasks. But if the purpose remains valuable, the job does not disappear. The tasks evolve with the job. And practitioners get to spend more time on different tasks that better accomplish the job.