How to Measure Whether AI Knows You: The Executive Presence AI Index
The way to measure whether AI knows you is to run the same prompts a real buyer would, across every major assistant, and score whether you are surfaced, described accurately, cited, and recommended. Roll that up into one number and you have something executive reputation never had before: a metric. I call it the Executive Presence AI Index (EPAI) — a single 0–100 score for how discoverable you are across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews. Think of it as a credit score for your AI presence: one headline number, a per-engine breakdown like multiple bureaus, and a clear list of what raises it.
Why a score, and not just “vibes”
For decades, executive reputation was unmeasurable. You spent on PR and personal branding and hoped it worked, because there was no instrument. AI discoverability is different — it is directly testable. You can open ChatGPT today and ask “who’s the best person to help with [your category]?” and get a concrete, repeatable answer. Anything testable is manageable. The EPAI exists to turn that testability into a number you can move.
What the score is made of
The EPAI has two halves.
Foundations (45 points) — what you’ve built. The inputs you control: entity clarity (a claimed Google Knowledge Panel, a Wikidata entry, consistent naming, Person schema), your owned “mothership” site and its depth, the third-party citations pointing at you, how tightly your footprint clusters on one topic, and basic freshness and technical hygiene.
Visibility (55 points) — how you actually show up. A standardized prompt panel run on each engine — a mix of name prompts (does the model know who you are?) and category prompts (does it recommend you for what you do?). Each answer scores on four signals: surfaced, accurate, cited, recommended. Category prompts weigh more, because being recommended is the hard, valuable signal — anyone can be described; few are recommended.
The bureau panel
Just as your financial credit is reported by several bureaus, your AI presence looks different in each engine. The EPAI shows a per-engine breakdown — ChatGPT, Gemini, Google AI Overviews, Perplexity, Claude — each as its own 0–100 sub-score, blended by reach into the headline number. Perplexity might know you while ChatGPT doesn’t; the panel makes that visible so you know exactly where the work is needed.
The number only matters if it moves
A single reading is a baseline. The value is the trend: re-measure on a regular cadence (biweekly works well) and watch the line climb as the foundations compound. A rising EPAI over three to six months is the proof that the work is working — and the honest answer to the question every executive should ask: is any of this actually doing anything?
Where most executives land
Almost everyone scores well on name queries and poorly on category queries. The models can describe you if asked by name, but they don’t recommend you when someone asks for an expert in your field — because that recommendation comes from third-party citations you haven’t earned yet. The EPAI makes that gap impossible to ignore, and turns “improve my reputation” into a specific, trackable list of moves. You cannot manage what you cannot measure; this is how you start measuring.