Intro
Most AI visibility pitches are written for marketing directors. Owners and general managers get a slide deck and a renewal request twelve months later, with no clear way to tell whether the money worked.
That gap is solvable. AI visibility is a commercial line item, not a creative one. It has a defined scope, a fixed deliverable list, a measurable proof set, and clear boundaries with the other vendors on the property.
This page is for the person who signs the contract. It explains what is in scope, what is not, how to read the work, what to ask any vendor before signing, and what proof to demand every quarter.
Book an owner-level audit walk-through
What you need to know as an owner or GM
AI visibility is the discipline of measuring and improving how AI answer engines — ChatGPT, Perplexity, Google AI Overviews, Gemini — describe and recommend your property when a future guest asks.
Three things matter at the owner level.
It is a commercial risk channel, not a content channel. When an AI engine routes a high-intent guest to an OTA aggregator or names a competitor first, that is direct revenue leakage. The work belongs alongside distribution and revenue management, not buried in social media.
It cannot be solved by writing more copy. The fix is structural: which facts the engines reuse, which sources they trust, which prompts surface you, which competitors substitute for you. Content is one input among several.
It is auditable. Every answer the engines produce can be captured, dated, scored, and compared quarter over quarter. If a vendor cannot show you the evidence layer, they are selling impressions, not visibility.
Owners do not need to learn the technical layer. They need to know what good looks like — and what to refuse.
What’s in scope (and what’s not)
A clean AI visibility engagement has hard edges. Confusion here is where budgets get diluted.
In scope:
- Prompt library design across discovery, comparison, trust, and conversion intents
- Multi-engine answer capture, dated and stored with model version metadata
- Eight-dimension scoring: presence, position, citation share, source quality, competitor dominance, fact accuracy, sentiment, commercial risk
- Fact accuracy audit and correction priorities on owned properties (brand site, key directories, structured data)
- Quarterly retest with side-by-side comparison and risk register update
- A short, sequenced action list — the next moves, in priority order, with effort estimates
Out of scope — and this is the part most decks blur:
- Content production at volume. Writing twenty articles a month is editorial work. It can be a partner deliverable, but it is not the visibility method itself.
- PR retainer. Earning new third-party citations is publicist work. The method tells you which sources matter; it does not place them.
- Paid media buying. Meta, Google Ads, OTA co-op spend — separate vendor, separate budget line.
- Photoshoots, video, creative. Asset production is creative agency scope.
- Reputation management replies. Responding to reviews is operations.
The method tells you what to fix and proves whether the fixes worked. The fixes themselves may be done by your existing partners. Owners who keep this separation get cleaner reporting and a fairer read on every vendor on the property.
For the full breakdown of how the method is structured, see the Capston Core methodology.
How to read the deliverables
There are four documents an owner should expect to receive, and four only.
- The prompt library. A locked list of 40 to 80 prompts that reflect how real guests ask AI engines about the property’s category, market, and competitive set. Once locked, it does not change between retests — that is what makes the score comparable.
- The baseline scorecard. The eight-dimension score with engine-by-engine detail. Read it like a P&L: dimensions are line items, the composite is a summary, the dimensions tell you where the work is.
- The risk register. Two or three concentrated commercial risks — OTA capture on conversion prompts, a competitor consistently named first, factual errors repeated across engines — each with a recommended response.
- The quarterly retest. Same prompts, same engines, same competitors. The delta against baseline is the work product.
If a vendor’s monthly report is a slide of impressions and a sentiment chart, that is not visibility work. Ask for the four documents above.
Six questions to ask any AI visibility vendor
Use these before signing. Vague answers are a signal.
- What is your prompt library and how was it built? A serious vendor co-designs the library with the property and locks it. If the prompts are generic or change every month, the score cannot be compared over time.
- Which engines do you capture, and how do you store the answers? Look for dated captures with model version metadata, retained for audit. “We use a dashboard” is not an answer.
- What is in scope and what is out? Demand a one-page boundary document. If content, PR, and paid all blur together, the engagement will be impossible to evaluate.
- How do you separate your work from my other vendors? The method should sit next to your existing partners, not absorb them. Ask what they hand off and to whom.
- What does the quarterly retest look like? Ask to see a real retest, anonymized if needed. Side-by-side, prompt by prompt, with the delta called out.
- What is your QA standard? Different analysts on different brands should produce comparable scores. Ask how that consistency is enforced. See Capston Core QA standards for what good looks like.
If a vendor cannot answer five of these six clearly, the engagement is a marketing retainer with AI vocabulary attached.
What proof to demand quarterly
Every quarter, the report should answer four questions in plain language.
- Did presence improve on the prompts that drive bookings? Conversion-intent prompts matter more than discovery-intent prompts. The retest should call this out separately.
- Did factual errors get corrected, and across how many engines? Fact accuracy is the cheapest, fastest dimension to move. If it is not improving, the engagement is not operating.
- Did commercial risk concentration go down? OTA capture share, aggregator substitution, intermediary routing on branded prompts — these are the lines an owner should track quarter over quarter.
- Where is the next quarter’s work? A short, sequenced list. No more than five priorities. Ranked.
Refuse anything that cannot be expressed in those four answers. The work is either commercial or it is decoration.
For properties operating to a published standard, the Capston certification framework formalizes this proof discipline.
How this fits into Capston Core
Capston Core is the method. The owner-level engagement uses the Capston Core methodology for the five-stage process, the Capston QA standards for consistency across analysts and quarters, and the certification framework when a property wants the proof discipline formalized.
Hospitality applications — boutique, resort, group, destination — are documented in the hospitality vertical.
→ Back to Capston Core
FAQ
Do I need a marketing director to manage this?
No. The method is designed so an owner or GM can sign off on scope and read the quarterly retest directly. A marketing lead is useful for executing the fixes, not for evaluating the work.
What budget tier should I expect?
Qualitative ranges only. Independent boutique properties sit at the lower tier with a focused prompt library and quarterly retests. Resort groups and multi-property operators sit higher because the prompt set, competitive set, and risk register all scale. Avoid vendors who quote without first scoping the prompt library.
How long before I see results?
Fact accuracy and structured data fixes move within one retest cycle. Competitor dominance and citation share shift over two to three quarters. Anyone promising faster is overselling.
Can I run this alongside my existing SEO and PR vendors?
Yes — and you should. The method identifies what needs to change; your existing partners can execute the fixes within their scope. Clean separation makes every vendor easier to evaluate.
Final CTA block
Get an owner-level read on your property’s AI visibility.
Book the audit walk-through
Read the methodology