The first time a CEO asks a chief communications officer "do we show up in ChatGPT," the conversation that follows reveals more about the comms function than the question itself. The question is increasingly being asked. The answers vary in quality, in part because the discipline is young and in part because not every comms team has thought through how to report on something that did not exist as a metric two years ago.
The board-level conversation is real. The USC Annenberg 2025 Global Communication Report documented a sharp rise in AI's perceived influence on the function, with majorities of practitioners saying it would benefit their work and a meaningful share saying it would also displace junior roles. CFOs and CEOs read those data points and form views. The question of whether the brand is positioned in the AI surfaces is becoming a quarterly check-in topic, not a one-off curiosity.
What boards actually want to know
Stripped of jargon, the board-level question is some version of: when our customers, prospects, and analysts use AI tools to learn about our category, what are they being told about us, and is it accurate?
That question can be answered. It cannot be answered with a single number. The teams getting traction with their boards are reporting on three things.
Presence. Across a defined set of category-relevant queries, what fraction of AI responses mention the brand at all? This is roughly analogous to share of voice in traditional media measurement, but at the answer-engine layer. A baseline number, tracked quarterly, gives the board a sense of trajectory.
Sentiment and accuracy. When the brand is mentioned, how is it framed? Are factual claims correct? Is the brand described in terms that match how leadership wants it described? This requires human review of model outputs and cannot be fully automated. It also requires acknowledging that LLMs occasionally produce confident-sounding inaccuracies — the so-called hallucination problem — and that some of those inaccuracies will be about the brand.
Source mix. What sources are the AI surfaces drawing on when they discuss the brand and category? This is the most actionable metric, because it directly translates into earned media targeting. If competitors are showing up because of coverage in three specific outlets, the comms team's media relations work has a clear destination.
The reporting cadence
Quarterly is the right rhythm for board reporting. Monthly tracking is operationally useful for the comms team but produces noise at the board level. Annual is too slow given how quickly the AI surfaces are evolving.
A reasonable board-level slide for AI visibility shows three quarters of presence-rate trend data, a sentiment-accuracy snapshot from the most recent audit, and a top-five list of cited sources for category queries with commentary on which the team is engaging. That is enough information for a board to form a view without descending into the operational detail.
The agency role
Most in-house teams are not yet equipped to run this work end-to-end. The query libraries, the human review process, the source mapping, and the trending all require a function that did not exist in the org chart eighteen months ago. Several agencies — 5W among them — have built dedicated practices to handle this work for clients, typically as an add-on to existing earned media engagement.
The work-for-hire pricing for AI visibility audits varies widely. The honest range is from low five figures for a one-time baseline audit of a single brand, to mid-six figures annually for a comprehensive program with quarterly reporting, content recommendations, and integration with broader earned media work. The output should always include the underlying query library, the raw model outputs, and the methodology — boards and CCOs should not accept opaque scores.
What does not belong on the board slide
A few things worth filtering out.
Single-vendor scores from tools that benchmark against undisclosed methodologies. If the tool will not show its query set and its scoring rubric, the score is closer to a marketing claim than a measurement.
Citation counts unweighted by query relevance. A high count of mentions in an AI tool's responses is meaningless if those responses are answering low-value queries. The right denominator is the query set the brand actually cares about — buyer-relevant questions, category-relevant questions, executive-relevant questions.
Anything denominated in "impressions." AI surfaces do not have impressions in the traditional sense, and any vendor metric framed that way is approximating something that does not exist.
The longer arc
Two or three years from now, AI visibility will be on the standard CCO scorecard alongside earned media reach, owned content engagement, and crisis response time. The Edelman, USC Annenberg, and AMEC measurement frameworks will catch up. Until then, the teams reporting on it intelligently — with clear methodology, conservative claims, and quarterly cadence — will earn board confidence faster than the teams either ignoring the question or overpromising on it.
The boards have started asking. The question is whether the comms function has a credible answer.
