PR is doing the easy three-quarters of GEO
This article originally appeared in PRWeek
Photo of Courtney Glymph, founder and CEO of YourStoryPR - By PRWeek
A PR strategy for AI that leaves out the bigger picture an LLM sees is missing the point.
PR needs to prioritise if the information available about a brand across different sources is credible and consistent enough to build a reliable picture.
Four things matter when an AI system decides what to say about a brand: how clearly the information is presented, where it appears, how recent it is, and if the claims can be supported by other sources. With various tools at the industry’s disposal, PR has gotten very good at the first three.
Appearing in the FT does not automatically make a claim authoritative to an AI system. The system considers the wider body of information it can find, including other sources. When there are conflicting claims such as differing statements from the media outlet, the company website and the CEO, the inconsistency becomes part of the picture.
Deciding whether a claim holds requires a conversation with the brand about what it can say specifically and how it can support that statement. Years of following a house style might be a problem here, because vague sentences that can pass legal review may say very little. Phrases like “leading provider”, “committed to excellence” and “uniquely positioned in a fast-moving market” don’t give clear information. A journalist can read those words and call for clarification but a language model has nobody to call (at least not yet). It looks elsewhere for information that supports the claim, and unsupported claims are less likely to be repeated when an AI system has other sources to compare them against.
Perfect structure will not save that sentence, nor will putting it on LinkedIn. They help a machine find and understand information but they cannot give the information substance that was never there. That is the part of GEO PR needs to pay more attention to. Does the information surrounding a brand add up to a credible picture? Are the claims specific? Do the website, press coverage, expert commentary and other sources reinforce one another?
If they do not, better optimisation will not solve the problem. A brand that gives AI little solid information to work with leaves space for a competitor to provide it instead.