Your Next ICP Is Going To Be An AI Agent

For years, senior communications leaders could treat the website as the front door: win search, earn the click, tell the story, capture the lead. That model is starting to break. The next high-intent reader may not be a person browsing your site. It may be an AI agent researching on behalf of a buyer.

The Economist is already trialling that split. It's testing agent-readable versions of content that has been restructured for AI answer engines. The human version can remain rich, visual and persuasive. The agent-readable version is structured around clear text, questions and answers, and explicit proof points that AI answer engines seek.

AI optimisation isn't just SEO renamed. The old goal of featuring in those 10 blue links on Google ended with a measurable click around 40-50% of the time. For comparison, AI answers only result in a click-through around 1% of the time. Compounding the issue is the opaqueness of AI answers — you may never know if your brand is missing until it hits the bottom line.

Comms teams may already be lamenting that with AI answer engines they need more tooling and it's yet another channel to manage. It's bigger than this. It's another entire audience to manage.

Comms in the era of AI

How should comms teams approach their second audience? Well the good news is we've got a head start versus other marketing disciplines. LLMs have been trained to learn what good content looks like. It values everything that marks out good PR — consistent language, clear positioning, credible evidence, verifiable claims. Serving this new audience is largely additive, and often complementary.

But what of the buyer? Their behaviour has also changed. If the buyer has got a first cut of information from the LLM, they are looking to back up this knowledge from other sources — so comms teams need to consider authenticity around their brand. This means building the profile of your leaders — through social media, with influencers, or through earned media. PR has always shaped those signals. Now those signals feed both machine summaries and human confidence.

That makes earned authority a go-to-market asset, not a vanity metric. A competitor with clearer messaging, fresher coverage and stronger third-party validation can shape the shortlist before anyone from the buying committee visits your owned channels. In an agent-mediated buying journey, reputation gaps become discoverability gaps.

Comms leaders need a more operational discipline around this. They should know how AI tools describe the business, which sources those tools cite, which competitors appear in recommendations, where the evidence is weak, and which external voices can credibly close the gap. The PR plan should not only ask what story the market needs to hear. It should ask what evidence an agent will find when it goes looking.

This is not a case for writing robotic copy for robots. The Economist's own framing is useful: AI as infrastructure, not authorship. The comms mandate is similar. Give machines the structured knowledge they need to fill in the gaps, and give people the credible sources they need to trust the answer.

The first audience is still human. Trust, judgement and emotional confidence still decide deals. But the first reader may increasingly be an agent. Senior comms leaders who equip that reader, while also building authentic third-party validation around the brand, will have a material advantage in the buying journeys AI now shapes.