At IPBC this year, the most common thread was IP teams reconsidering long-standing decisions as AI reshapes their day to day. We spent three days in San Diego with senior IP leaders, and the conversations that stayed with us were not about any single product or panel. They were about choices these teams had treated as settled for years, and no longer do.

The clearest example is build versus buy. Several in-house teams shared the same top-down pressure: take advantage of AI by vibe coding their own tools in-house rather than signing a new vendor. It sounds reasonable in a leadership meeting. The question it skips is who actually builds and maintains those tools. Usually that lands on the same IP team that is already underwater, and most of the counsel we talked to had no interest in becoming part-time software maintainers. They would rather bring in something built for the job and get back to their actual work. That gap, between a mandate to build and a practitioner who wants to buy, came up again and again, and it was rarely resolved.

Cost came up just as often, and in a sharper form than we expected. A patent program's budget used to be something you defended once a year and then left alone, and that is changing fast. Teams are reopening questions they had long treated as closed, from what the program should really cost to which patents they are renewing out of habit rather than strategy. These were not idle complaints about expensive lawyers. They were people rethinking how the whole function should run, and willing to make uncomfortable changes to do it.

What connects both themes is AI. These decisions are reopening now because AI has made it reasonable to ask questions that did not have good answers a few years ago. When a tool can prune a portfolio, draft a response to an office action, or surface which assets are worth keeping, "we have always done it this way" stops being a good enough reason to keep doing it. Most of the teams we spoke with are somewhere in the middle of working that out, and the ones furthest along had simply stopped treating their old setup as permanent.

We built ArcPrime for this kind of team. It is an AI-native IP management system built by patent counsel for in-house counsel, made for the department that wants what AI can do without becoming the team that has to build and maintain it.

We came away from San Diego more sure than before that this shift is real, and moving faster than most expected. If your team is working through the same questions about what to build, what to buy, and what your program should cost, we would be glad to compare notes.