The Senate Got Claude; Mapping the AI terrAIn is More Urgent than Ever for Public Affairs

The Senate (officially) got Claude this week. Earlier this year I argued that it was safe to assume that plenty of congressional staffers were already using AI tools to help with their work. But now, with the Senate formally procuring Anthropic’s Claude, its ubiquitous use must be expected.

And now, even in the US Senate, AI increasingly sits between the information you put into the world and the people you’re trying to reach.

The Senate’s rollout specifically identifies tasks like research, information synthesis, and drafting as potential applications for Claude. These are fundamental aspects of staff work; evaluating competing claims, and translating complicated information into recommendations for policymakers. Even if the decisionmaking is still left for humans to muddle through, the choices around the framing and prioritization of information can really stack the deck, which is why having insightful staffers whose thinking aligns with their lawmaker is so essential (shout out to my many former staffer colleagues). 

And the Senate is hardly alone. I saw this evolution firsthand during my time at the State Department, where I sat on the Enterprise Data Council during the adoption of StateChat and led the team responsible for developing Northstar, an AI-powered media and information analysis platform. Those experiences showed me how effectively deployed AI tools can dramatically expand an organization’s capacity to collect, organize, and make sense of information.

But they also showed me how dramatic this shift is, and the opportunity that exists for those who understand it. For years, Public Affairs work has focused on getting the right information in front of the right audiences, whether through earned media, direct engagement, digital communications, or other channels. The underlying assumption has often been that getting information in front of an audience is the critical step. Once someone encounters your perspective, they can evaluate it alongside other available information. But increasingly, there’s another step between exposure and understanding.

That means public affairs professionals need to think not only about information exposure, but about information interpretation.

It’s a shift that distinguishes generative engine optimization (GEO) from traditional search engine optimization (SEO). Ranking prominently in search results is one thing. Understanding how AI systems interpret, synthesize, and present information about your issue is another.

You need to ask: When a congressional staffer asks an AI tool to explain your issue, what does it say? Which sources does it draw from? Which arguments does it elevate? What context does it leave out? And does your perspective make it into the answer at all?

This is exactly the challenge we anticipated when we developed terrAIn, Precision’s approach to GEO.

Our work isn’t simply about winning visibility for a particular topic, or even a particular search. It’s about understanding the specific questions audiences are asking, examining the information AI tools use to answer those questions, and identifying opportunities to make the resulting answers more complete, accurate, and representative of our clients’ perspectives.

Because the difference between being part of the conversation and being part of the answer is becoming increasingly important.

And while the Senate’s adoption of Claude is a particularly visible milestone, the broader trend is already underway across government, businesses, media, and other institutions. AI isn’t replacing every existing information channel. But it is increasingly becoming the interface through which people navigate them.

For public affairs professionals, that means GEO can’t be treated as an interesting experiment on the margins of a communications strategy. It needs to be part of how we understand and engage the information environment.

Of course, the Senate’s new tools may also give Hill watchers a fresh opportunity to spot (and mercilessly mock) the occasional obviously AI-generated flourish in congressional press releases. But beyond the inevitable AI-generated prose, there’s a more consequential change taking place.

The people you’re trying to reach are increasingly asking AI what they should know. It’s worth knowing what AI is telling them.