A thought experiment. And a checklist.
AI Communications, defined
AI Communications is the discipline of building authority across AI answer engines — ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — alongside earned media, digital, and influencer channels. It combines public relations, Generative Engine Optimization (GEO), and AI-visibility research to influence the answers anyone asking now begins with.
It is a Tuesday morning. A board member sends a screenshot.
She asked ChatGPT to recommend the top brands in the category. The brand is not on the list. Two direct competitors are. A third name is a brand nobody on the team has ever heard of. The fifth name is wrong — the engine confused the brand with a similarly-named company in a different country.
The board member's email is six words long: "Is this something we need to discuss?"
There is one hour before the meeting.
This is the seventh piece in the AI Communications series, and a thought experiment for the leadership team of every brand, every founder, every institution that has not yet built the infrastructure described in the Stack. The premise: this day is coming, the only question is whether the response capacity is built before it arrives.
What just happened
Three things happened simultaneously.
One. The AI engine assembled an answer using the sources it considers most authoritative for the query. Those sources did not include the brand.
Two. The board member ran the query because someone — a friend, a journalist, an analyst — told her she should. By the end of the week, three more board members will have run it.
Three. A competitor the team has been watching for years just landed the most valuable placement in the new media graph: a name in the ChatGPT answer for a category-defining query, retrieved by everyone who asks the question for the foreseeable future.
The brand did not lose to a press cycle. It lost to a structural absence the press cycle never noticed.
What can be done in the next hour
The gap cannot be fixed in an hour. But it can be addressed with the right answer at the meeting.
The wrong answer: "We are getting a lot of press, the answer must be wrong." The press the brand is getting may not be in the sources the engine pulls from. Press alone does not move Citation Share.
Also wrong: "It's just one engine." The pattern repeats across Claude, Perplexity, Gemini, and Google AI Overviews. If the brand is absent from one, it is usually absent from all five. The board member will check.
The right answer: "We have an AI visibility gap. We are not measuring it yet. The gap is closeable in twelve to eighteen months with disciplined work across the seven layers of the AI Communications Stack. Here is what we need to start, and here is what the investment looks like."
That is the answer that turns a panic moment into a strategy moment.
What should have been done six months ago
Five things.
1. Baseline Citation Share. A defined query set, scored across the five engines, compared against a named competitive set. Without this, the team has no idea where it stands.
2. Map the sources. For every query that matters in the category, log which publications, websites, and authors the engines cite. That list is the new media plan.
3. Audit Layer 1 (Identity). Website, founder LinkedIn, old press releases, Wikipedia entry, Crunchbase profile — do they say the same thing? If they say different things, the engine averages, and the average is rarely flattering.
4. Audit Layer 2 (Authority). Wikipedia, Wikidata, schema markup, glossary inclusion, structured authorship on bylines. The signals that tell the engines the brand is real and worth retrieving.
5. Establish monitoring. Weekly or biweekly Citation Share runs. Same query set, same engines. So when the board member's screenshot arrives, it is not the first time the team has seen the data.
The team does not have to do all five at once. It does have to start before the screenshot arrives.
What to do now
Three moves, in this order.
Move 1: Acknowledge the gap. Do not minimize the screenshot. Do not deflect to press coverage. Acknowledge that the AI answer is the new front page, that the brand is currently absent, and that closing the gap requires deliberate work over twelve to twenty-four months.
Move 2: Commission the audit. Authorize a baseline AI Communications audit — Citation Share, source mapping, Identity audit, Authority audit. This produces the document the board needs to understand the scope of the gap and the path to closing it.
Move 3: Sequence the layers. The audit names which layers are broken. Most brands find Layer 1 (Identity) and Layer 2 (Authority) are the heaviest debt. Start there.
The brand that takes these three moves on the day the screenshot arrives is sixty days ahead. Six months in, the gap is closing. Twelve months in, the brand is back in the answer.
What stays true
The screenshot will come. For every brand. The only question is whether it triggers a strategy or a crisis.
Three things determine which one.
Whether the brand has measurement infrastructure. Without it, the screenshot is the first time anyone sees the data, and panic is the default response.
Whether the brand has framework discipline. Without the Stack, every team member has a different theory of what to do, and resources scatter across uncoordinated tactics.
Whether the brand has a defined response process. Without one, the response is improvised, the timeline blows out, and the gap widens before it closes.
Build the infrastructure before the crisis — not during it.
The harder truth
Most brands are not yet at this stage. The board member's screenshot will arrive sometime in the next eighteen months for nearly every brand of any consequence. Most leadership teams will not have built the response capacity in advance.
The brands that built it will pull ahead. The brands that did not will spend two years trying to catch up.
The brands writing checks for the infrastructure now are writing them at the bottom of the market. The brands waiting until the screenshot arrives will write them at the top.
Either way, the checks get written. The question is what they buy.
