Every public affairs firm in Washington is talking about AI. The question isn’t whether your firm has added AI to its pitch deck – virtually all of them have one – rather the question is whether they know what they’re doing. Here’s a framework for cutting past the marketing.
How do they balance experience with technology?
AI can surface opposition arguments and map the information environment at a scale no human team could match. But it can’t replace the judgment that comes from decades in Washington. AI can spot a narrative threat; only a veteran can decide whether to respond aggressively, reframe quietly, or let it die for lack of oxygen. Ask directly: who has final strategic authority: the technologist or the public affairs veteran? If they can’t answer clearly, they’ve added a tool, not solved integration.
Do they have the analytical backbone?
Washington has long been a world of English majors, who are skilled communicators and relationship builders. That talent still matters, but AI-informed work needs a different rigor. Ask to meet the people who can interpret what platforms surface and translate findings into strategy and not just those presenting results.
What’s their actual track record with innovation?
Firms genuinely equipped for this moment were experimenting with data-driven approaches long before AI became a Washington buzzword. Ask for their demonstrated record, not their vision for the future.
How are they using AI for clients today?
Ask them to walk through concrete examples: which platforms, how they query them, what an AI-informed deliverable looks like. And ask for their written ethical guidelines. AI can produce confident misinformation; a firm without guardrails is a liability. If they don’t have them, that tells you something.
Are they building tools or reselling them?
Consumer tools adapted for public affairs beat nothing, but they weren’t built for congressional research, regulatory proceedings, or diplomatic communications. Ask whether the tools are their own or repackaged.
Have they developed agentic AI capabilities?
Systems that autonomously research, monitor, and generate intelligence are the next frontier. Firms already building them are investing where the work is heading.
Can they measure what matters?
Impressions and sentiment scores borrowed from consumer marketing don’t capture the policy information environment. Ask how they define and measure success and be wary of activity reports dressed up as outcomes.
Do they understand AI’s limits — or are they selling a miracle?
A firm that can articulate AI’s advantages, limits, and failure modes understands its tools. The honest answer is nuanced. Be skeptical of anyone who says otherwise.
Is your data safe?
When a firm uses AI on your behalf, your proprietary positions and confidential communications may be entering those systems. Ask which platforms receive client information, and under what terms.
The bottom line
The best-equipped firms aren’t the largest or the ones with the most impressive AI vocabulary. They’re the ones that have genuinely integrated AI with experienced judgment and can show you, specifically, how it makes their advice better.