100 Questions Every Marketing Agency Asks About AI Search Visibility

The client asks why their visibility score dropped this week when nothing changed on the website.

AI answers change even when a client's site does not, because the model, the training data, or a competitor's new content can shift. Show the trend line instead of one week's number, and separate site side changes from model side changes in the report so the client learns to read both.

How do you explain share of model to a client who has never heard the term before?

Compare it to share of voice in traditional media. Share of model is the percentage of times an AI engine mentions the client's brand versus competitors when answering the same set of buyer questions, tracked over time instead of in a single snapshot.

The client wants one single number instead of six platform scores. What do you show them?

Build a simple weighted average across the platforms that matter most for that client's audience, then keep the platform breakdown available underneath for anyone who wants detail. Executives want one number on the cover slide and the full picture on the next page.

The client thinks the tool is making up citation data because they cannot see the same answer when they check it themselves.

AI answers are not fixed. The same prompt can return different results minutes apart because of personalization, location, and normal model variance. Explain this early, before the client goes looking for proof on their own phone, and share raw response logs when you can.

What do you say when a competitor gets cited in every ChatGPT answer but the client's site is technically better built?

Technical quality is necessary but it is not what earns a citation on its own. Models cite sources that already carry authority signals: third party mentions, structured comparisons, and consistent factual presence across the web. Show the client exactly which sources the model is pulling from instead.

The client wants monthly reports, but AI answers change daily. How do you set the right cadence?

Track daily or weekly internally so you catch real shifts early, but report monthly trends to the client unless something urgent happens, like a sudden drop tied to a model update. Frequent internal checks plus a calmer external cadence keeps everyone informed without causing alarm over normal noise.

How do you show a skeptical CFO that AI visibility work is not just a vanity metric?

Tie citation frequency to a downstream number the CFO already tracks, like branded search volume, demo requests, or self reported "how did you hear about us" data. A rising citation rate that lines up with a rising lead number is far more convincing than the citation rate alone.

The client keeps comparing AI visibility numbers to their old Google Analytics traffic numbers. How do you reset that expectation?

Explain plainly that most AI answers do not produce a click, so traffic is the wrong yardstick. The new yardstick is whether the brand gets mentioned and recommended at all, which is a visibility and trust metric, not a traffic metric, and needs its own baseline.

What do you tell a client whose brand gets cited, but with the wrong price or outdated information?

This usually means an old page, an outdated third party listing, or a stale forum thread is still the model's preferred source. Find that source, get it updated or replaced with fresh content, and track whether the correction actually shows up in later answers.

The client wants to see competitor citation data too, but the contract only covers their own brand. How do you handle the scope creep?

Competitor benchmarking is genuinely useful, so it is worth offering, just not for free inside a brand only retainer. Package it as a defined add on with its own line item, so the client understands it is extra value rather than something you forgot to include.

Two different AI visibility tools gave two different citation counts for the same brand and the same week. Which one is right?

Both can be right and still disagree, because each tool likely tracks a different prompt set, sample size, and query timing. Compare methodology, not just the number: how many prompts, how often, and on which platforms, before deciding which tool to trust for reporting.

Why does the same prompt in ChatGPT sometimes mention the client brand and sometimes not?

Generative answers are probabilistic, not deterministic, so the same input can produce slightly different output each time. This is exactly why single spot checks are unreliable and why tracking needs repeated runs over time to reveal a real trend instead of noise.

The tool shows a mention, but you cannot reproduce it when you type the same prompt into the actual app. What is going on?

Your logged in account, browsing history, and location all shape what you personally see, while the tracking tool usually queries from a neutral, logged out state. Neither view is wrong, they are simply two different vantage points on the same model.

A client's location changes the ChatGPT answer they get. Whose result actually counts for tracking?

Track from the locations your client's actual customers are concentrated in, not from your own office. If the client serves multiple regions, track each major region separately rather than averaging them into one misleading blended number.

The AI visibility dashboard just stopped updating overnight with no warning. What is the first thing to check?

Check the vendor's status page or changelog first, since AI platform API changes cause these outages more often than the tracking tool itself breaking. If nothing is posted, contact support directly rather than waiting, since a client facing report deadline may be close.

How many prompts do you actually need to track before the data becomes reliable?

For most mid sized clients, 20 to 50 well chosen prompts covering real buyer questions is enough to see a meaningful trend. Fewer than that and normal variance can look like a real shift; far more than that mostly adds cost without adding insight.

Do you need to track the exact same prompt wording every time, or can you vary it slightly?

Keep a core set of prompts worded identically over time so the trend line stays comparable, then add a second rotating set with varied phrasing to catch how differently worded questions behave. Mixing both without labeling them clearly is what causes confusing reports.

What happens to historical tracking data if the agency switches AI visibility vendors mid contract?

Export everything before you cancel, including raw prompt level history, not just summary charts. Most vendors do not guarantee data access after a subscription ends, and rebuilding a full year of baseline history from scratch is expensive and slow.

Why does Perplexity cite completely different sources than ChatGPT for the exact same question?

Each engine retrieves and ranks sources with its own method, so overlap in citations is common but not guaranteed. Perplexity leans more heavily on live web search results, while ChatGPT blends training data with search, so the two rarely behave identically.

A client asked for raw screenshots as proof instead of trusting the dashboard. How do you handle that request?

Provide them, since a handful of real screenshots alongside the dashboard trend builds trust faster than any chart alone. Just set the expectation that a screenshot is one moment in time and the dashboard trend is what actually matters for decisions.

Should AEO work be billed as a new line item or folded into the existing SEO retainer?

A separate line item, at least at first. Folding it in quietly makes it invisible when you need to justify the value later, and it signals to the client that AEO is a real, distinct discipline rather than a repackaged existing service.

A client wants AI visibility tracking added, but refuses to increase the monthly budget. What do you do?

Offer a smaller starter scope, a limited prompt set on the two or three platforms that matter most for that client, instead of full coverage. A modest paid addition beats giving away the full service for free and normalizing that price point.

How many hours per month should you budget internally for one client's AEO program?

Most mid sized programs run on 8 to 15 hours a month once the initial setup is done: reviewing tracking data, briefing content updates, and preparing client reports. Setup months and citation outreach pushes take considerably more time than steady state months.

What is a fair rate to quote for a one time AI visibility audit?

Most agencies price a focused one time audit, covering a defined prompt set across the major platforms plus a written findings report, between $1,500 and $5,000 depending on client size. Price it as a discovery product that naturally leads into a retainer conversation.

A client wants a guarantee that ChatGPT will mention them within 90 days. How do you handle that request?

Do not promise a specific citation outcome on a fixed timeline, since no agency controls a third party model's output. Commit instead to a defined set of actions and a measurable visibility trend, and put that distinction in writing before signing anything.

How do you price AEO work for a client with five different product lines and five different audiences?

Treat it closer to five smaller engagements than one big one, since each product line needs its own prompt set, content review, and reporting. Price per product line with a small bundled discount for running them together under one account team.

Should citation outreach be charged separately, or bundled into existing digital PR work?

If the client already pays for digital PR, extend that scope rather than duplicating it, since pitching journalists and pitching for AI citations often target overlapping publications. Charge separately only if the client has no existing PR retainer at all.

A prospect wants AEO service but has almost no existing content. Do you quote it differently?

Yes. AEO cannot fix a content gap that does not exist yet to optimize. Quote a content foundation phase first, then move to tracking and refinement once there is enough real content on the site for a model to actually retrieve and cite.

How do you justify a retainer increase when AI visibility numbers improved, but organic traffic stayed flat?

Reframe the conversation around brand presence and buyer trust rather than traffic, since improved citations mean the client's name is reaching people at the exact moment they compare options, even without a click. Pair that story with any lead or branded search movement you can show.

What should be in the contract if a client asks the agency to guarantee AI Overviews inclusion?

State clearly that inclusion in Google AI Overviews is controlled entirely by Google's own systems and cannot be guaranteed by any agency. Define the scope as the specific optimization actions taken, not the outcome, so the contract protects both sides from an impossible promise.

How do you stop prompt data from one client's workspace leaking into another client's report?

Use a platform with dedicated multi brand workspaces rather than one shared account with folders, since folders rely on people remembering to file things correctly. A true workspace separation removes the human error that causes cross client mistakes in the first place.

What is the best way to organize prompt groups when one client has three separate brands under one company?

Give each brand its own workspace and its own prompt set, even though they share a billing relationship, since mixing brand specific prompts into one pool makes trend data meaningless for any single brand. Report each brand separately, then summarize at the parent company level if needed.

How many client accounts can one AEO specialist realistically manage at once?

Most specialists can handle 6 to 10 accounts of moderate complexity if reporting and content review are the main tasks. Fewer than that if the specialist is also doing citation outreach and content writing directly instead of briefing other team members.

A big client wants a dedicated Slack channel for daily visibility alerts. Is that normal, and should you agree?

It is becoming normal for top tier accounts, and it can work well if the alerts are automated rather than manually posted, so the request does not quietly turn into a full time monitoring job for one person on your team.

How do you onboard a new client's AI visibility tracking without waiting a full month for baseline data?

Run a manual spot check across the core prompts on day one so you have something concrete to discuss immediately, then let the automated tool build a proper statistical baseline over the following two to three weeks before drawing any trend conclusions.

What is the right way to offboard a client and hand back their historical tracking data?

Export the full prompt history, citation logs, and reports into a client owned format before access is revoked, and confirm receipt in writing. This protects the agency from disputes later and is simply good practice regardless of why the relationship ended.

How do you split reporting access between a client's in house marketing team and their outside PR agency?

Give both view access to the same dashboard rather than sending two separate reports, so the two teams work from identical numbers. Reserve editing and prompt management access for your own team to avoid uncoordinated changes to the tracked prompt set.

A client asked to add 40 more prompts mid contract. How do you handle the workload without breaking the retainer?

Treat a large prompt increase as a scope change, not a favor, since more prompts mean more analysis and reporting time even if the tool itself absorbs the extra tracking automatically. Quote the incremental cost clearly rather than absorbing it quietly.

Should junior team members have access to every client's raw prompt data, or just their assigned accounts?

Restrict access to assigned accounts by default. Broad access increases the risk of an accidental cross client mistake, and most junior team members do not need visibility into competitor client data that has nothing to do with their own workload.

How do you keep prompt lists consistent across clients in the same industry without copying one client's strategy onto another?

Build a shared library of generic, industry level question templates that any client in that category could use, then customize a portion of the list for each client's specific products and positioning. This keeps efficiency high without handing one client's exact playbook to a competitor.

Why does a client's brand show up in Google AI Overviews but never in AI Mode for the same query?

AI Overviews and AI Mode use different retrieval and ranking logic even inside the same company, so strong presence in one does not guarantee presence in the other. Track them as two separate surfaces in your reporting rather than one combined Google score.

Does Claude cite sources the same way ChatGPT does, or does it need a separate tracking approach?

Claude's citation behavior and web access differ from ChatGPT's, so treat it as its own tracked platform rather than assuming similar results carry over. What earns a citation on one does not automatically earn one on the other.

Why does Perplexity keep citing a five year old blog post instead of the client's updated page?

The old post likely still has more backlinks, more third party mentions, or a stronger existing authority signal than the new page has managed to build yet. Getting the new page cited instead usually takes active outreach, not just publishing an update and waiting.

How often does Gemini actually change its cited sources for the same prompt?

More often than most people assume, especially for competitive commercial queries where several sources carry similar authority. Track Gemini over multiple runs across a week before drawing any conclusion about which sources it truly favors for a given prompt.

Do you need to track AI Overviews separately from regular Google search rankings?

Yes. Ranking in the traditional blue links and appearing inside the AI Overview box are two different outcomes that can move independently of each other. A page can hold position one and still be absent from the AI Overview summary above it.

Why did the client's brand disappear from ChatGPT answers right after a major model update?

Model updates can shift which sources get retrieved and trusted overnight, sometimes with no relation to anything the client did. Check whether competitors experienced a similar shift at the same time before assuming the drop is specific to the client.

Is there a reliable way to tell if a citation came from the model's training data or a live web search?

A visible source link or footnote usually points to a live search result, while an answer with no link and slightly outdated facts often comes from training data instead. It is not a perfect test, but it is a useful first signal.

Why does the client's Wikipedia page get pulled into answers more often than their own website?

Wikipedia carries an unusually strong trust signal for most models because of its editorial history and citation standards. A well maintained, neutral, and properly sourced Wikipedia entry is one of the highest leverage assets a brand can have for AI visibility.

Do voice assistants like Alexa and Siri need a separate tracking strategy from chat based AI engines?

Yes, since voice assistants often rely on different underlying data sources and typically return a single short answer instead of a full response with multiple citations. Most agencies treat voice as a smaller, secondary priority behind the major chat based platforms for now.

How much does personalization or login status change what a user actually sees compared to the tracked result?

It can change quite a bit, especially for logged in users with chat history the model can reference. Explain to clients that tracking tools measure a neutral baseline on purpose, since a personalized result cannot be fairly compared week over week.

Should you rewrite an old blog post for AEO, or just add a new FAQ section to it?

If the post's core facts are outdated, rewrite it fully, since adding an FAQ on top of stale information does not fix the underlying trust problem. If the facts still hold up, a targeted FAQ addition is usually enough to improve retrieval.

How do you know if a page's structure is the reason it never gets cited, versus the content itself being wrong?

Check whether competitor pages with similar structure but different content get cited for the same prompts. If they do, the problem is likely your content's substance, not its format. If nobody with that structure gets cited, structure is worth testing first.

The client's writers keep producing generic content. How do you convince them to add original data?

Show them a side by side comparison of a cited competitor page that includes original statistics or a survey, against the client's generic version. A visible gap in specificity usually convinces writers faster than a general instruction to be more original.

Do FAQ schema markups still help AI engines the same way they helped Google's snippet feature?

FAQ schema still helps, mainly because it gives the model a clean, unambiguous question and answer pair to extract, especially useful when a crawler cannot render JavaScript to see that same content rendered visually on the page.

How long does it typically take for a newly optimized page to start showing up in AI answers?

Anywhere from a few weeks to a few months, depending on how quickly the page gets crawled, indexed, and picked up by other sources that the model already trusts. Set client expectations toward the longer end rather than promising a fast result.

Should product pages be optimized differently from blog posts for AI citation purposes?

Yes. Product pages benefit most from clear specifications, pricing, and structured attribute data, while blog posts benefit from depth, original analysis, and direct answers to comparison style questions. Treat them as two different optimization playbooks.

What is the minimum content structure a page needs before it is even worth tracking?

A clear heading structure, a direct answer near the top of the page, and enough substance to actually answer the question fully. A thin page with no real answer is unlikely to get cited regardless of how well it is tracked.

How do you prioritize which of 200 client blog posts to update first for AEO?

Start with the posts already ranking well in traditional search but missing from AI answers on related prompts, since those already have earned authority and just need an AEO focused update rather than a full rebuild from nothing.

Do case studies actually get cited more often than regular blog content?

Case studies with real, specific numbers tend to perform well because they give the model a concrete fact to reference, something a generic blog post rarely provides. Vague case studies without numbers do not carry the same advantage.

Should you build separate pages just for comparison style prompts like brand A versus brand B?

Yes, for the comparisons buyers actually search for most. A dedicated, fair, well structured comparison page gives the model a clean single source to cite instead of leaving it to piece together an answer from scattered, less reliable third party sources.

How do you prove that a lead actually came from an AI platform and not organic search?

Check referral traffic for known AI domains in your analytics, and add a simple "how did you hear about us" field to lead forms that includes an AI assistant option. Neither method is perfect alone, but together they give a workable signal.

GA4 shows almost no referral traffic from ChatGPT, even though the tracker shows citations. Why the mismatch?

Most AI answers do not include a clickable link, and the ones that do often get filtered or misclassified by analytics tools not built for this traffic type. Citation without a click is still real value, it just does not show up as a session.

The client wants pipeline revenue tied directly to AI visibility work. Is that even measurable yet?

Direct, dollar for dollar attribution is still immature across the whole industry, not just for your agency. Be honest about that limitation, and offer the closest available proxy instead, like branded search volume growth or self reported source data from sales calls.

How do you separate the impact of your AEO work from the client's own PR team getting them mentioned in the press?

Coordinate with the PR team rather than competing with them for credit. Log major press hits alongside your own content and outreach work in the same timeline, so any visibility jump can be discussed honestly as a shared outcome.

Is there a reliable way to track AI referral traffic separately from direct traffic in analytics?

Set up custom channel groupings in GA4 for known AI assistant domains so they stop falling into the generic direct traffic bucket. It is not a perfect solution, since many AI clicks still arrive with no referrer at all, but it captures more signal than the default setup.

What should you tell a client who says AI visibility improved but sales did not move?

Ask honestly whether sales expectations were realistic for the timeline and whether other factors, like pricing changes or a slow season, might explain the gap. Visibility is one input to a sale, not the entire sales process on its own.

How long should a client wait before expecting to see any business result from AEO work?

Set the expectation at three to six months for meaningful movement, similar to traditional SEO timelines. Anyone promising fast, dramatic business results from AEO alone within a few weeks is setting up a conversation they cannot win later.

Do citations without a clickable link count as real value, or just brand awareness?

They count as real value, similar to how a strong brand mention on a podcast counts even without a link. The buyer still heard the recommendation at a moment of active consideration, which shapes their decision even if analytics cannot capture it directly.

How do you set a realistic baseline before showing month over month improvement?

Run at least two to three weeks of tracking before calling anything a baseline, since a single day's snapshot is too volatile to serve as a fair starting point. Present the baseline as a range, not one fixed number, to reflect that natural variance honestly.

What metric should replace traffic as the primary KPI when AI answers rarely produce a click?

Citation frequency and share of model against named competitors work best as the primary KPI, since they measure whether the brand is being recommended at all, which is the actual outcome AEO work is designed to move.

What does a prospect usually want to see in the first five minutes of an AEO focused pitch?

A live comparison showing whether their brand and their top competitor get mentioned for the same buyer question. Seeing a real gap on the spot is far more persuasive than any slide explaining what AEO means in theory.

How do you run a free audit for a prospect without giving away all your findings for free?

Share the problem clearly, like which prompts show the biggest visibility gap, without handing over the full remediation plan. The audit should prove you found something real; the paid engagement is what fixes it.

Should you lead a pitch with AI visibility data, or with traditional SEO data first?

Lead with whichever data tells the more urgent story for that specific prospect. For a brand already ranking well but invisible in AI answers, lead with the AI gap, since it is the newer, less expected problem and gets more attention in the room.

A prospect's competitor already works with an AEO agency. How do you position against that?

Show the prospect exactly what the competitor is currently getting right, using real citation data, and frame the pitch around catching up quickly rather than starting from theory. A visible, specific gap creates urgency competitors' generic pitches usually lack.

What is a fair scope for a paid discovery phase before a client commits to a full retainer?

A two to four week engagement covering a defined audit, a prioritized action list, and a proposed prompt tracking plan works well. It should be affordable enough that saying yes is easy, and substantial enough that it feels like real, paid work.

How do you handle a prospect who thinks AEO is just a rebrand of SEO with no real difference?

Show them a real query where their SEO rankings look strong but their brand still never appears in the AI generated answer above those rankings. A visible, undeniable gap does more to change minds than any explanation of terminology.

What proof point convinces a skeptical CMO that AI search visibility actually matters for their category?

Industry specific usage data works better than generic statistics, so pull the closest available figure for how often people in that CMO's category are already using AI assistants for research, then connect it directly to their own buying journey.

Should you offer a performance based pricing model for AEO work, and does that ever work well?

It rarely works well, since citation outcomes depend partly on factors an agency cannot control, unlike traditional pay per lead models. If a client insists, tie any performance bonus to a visibility metric your work directly influences, not to sales the agency cannot fully attribute.

How do you qualify whether a prospect is actually ready for AEO, or still needs basic SEO first?

If the site has thin content, broken technical fundamentals, or almost no existing organic presence, fix those first, since AEO amplifies an existing foundation rather than replacing the need for one. Selling AEO before the foundation exists sets up an unwinnable engagement.

What is the best way to show a prospect their AI visibility gap compared to their top three competitors?

Run the same five to ten buyer questions across the major AI platforms for the prospect and their three named competitors, then present the results as a simple side by side grid. The visual gap usually speaks louder than any narrative around it.

How do you train a traditional SEO specialist to think in prompts instead of keywords?

Have them practice converting a list of existing target keywords into full natural language questions a real person would ask an assistant, then compare the differences in phrasing and intent side by side until the shift starts to feel natural.

What should be in a new hire's first week checklist for an AEO focused role?

Have them run the same 10 prompts across all six major platforms by hand before touching any tool, so they experience firsthand how differently each engine answers the same question. That hands on exposure teaches more than any onboarding document alone.

How do you keep the whole team updated when a major AI platform changes its citation behavior overnight?

Assign one person to monitor platform changelogs and industry news each week, then share a short summary in a standing team channel. A single point of ownership prevents important changes from getting missed while everyone assumes someone else is watching.

Should every SEO on the team learn AEO, or should it stay a specialized role?

Every SEO benefits from a working baseline understanding, since it is now part of a basic client conversation, but deep specialization in a couple of people keeps the quality high without diluting focus across the whole team.

What internal documentation actually helps when a client asks a technical question about how AI models retrieve content?

A simple, plain language explainer for each major platform, written once by your most technical team member and reused across client conversations, saves enormous time compared to explaining the same retrieval concepts from scratch every time it comes up.

How do you avoid burnout on a team that has to check six AI platforms daily for every client?

Automate the actual checking with a tracking tool and reserve human time for interpreting the results and deciding what to do about them. Manually querying six platforms by hand every day for multiple clients is not a sustainable long term process for any team.

What is a reasonable ramp up time before a new AEO hire can run a client account solo?

Around four to six weeks for someone who already has a solid SEO or content background, closer to two to three months for someone entirely new to search marketing. The field moves quickly, so ongoing learning matters more than a fixed ramp period.

How do you standardize prompt writing so different team members do not produce wildly different tracking sets?

Create a short prompt writing guide covering tone, length, and the mix of informational versus commercial intent questions expected in every set, then have a second team member review new prompt lists before they go live for any client.

Should account managers be trained to explain AEO basics themselves, or should they always loop in a specialist?

Account managers should handle the basics confidently on their own, since clients ask simple questions far more often than technical ones. Reserve the specialist's time for the deeper technical questions that actually require their expertise.

What internal SLA should the team have for reacting to a sudden visibility drop for any client?

A same day acknowledgment and an initial cause check within 48 hours works well for most agencies. Clients tolerate a temporary drop far better when they know someone is already looking into it than when they discover it themselves in a monthly report.

Does a local service business, like a dentist or a plumber, even need AI visibility tracking yet?

Adoption for local service search is still lower than for broader product research, but it is growing steadily, so a light touch tracking setup now is a reasonable, low cost way to stay ahead without a large budget commitment.

How do you track AI visibility for a client with locations in ten different cities?

Pick a small set of representative cities rather than tracking all ten in full depth, unless the client's budget supports it. Focus first on the cities with the highest revenue or the most competitive local landscape.

Do AI engines answer the same local query differently depending on the user's actual location?

Yes, location can meaningfully change which local businesses a model surfaces in its answer. Always track from a location close to where the client's actual customers are searching, not from wherever your agency happens to be based.

How do you adapt AEO strategy for a client whose customers mostly speak a language other than English?

Track prompts written natively in that language rather than translated English prompts, since translation can shift both intent and phrasing in ways that change what a model retrieves. Content should also be written natively, not translated after the fact.

Does a small ecommerce brand with a limited budget get any real value from AI visibility tracking?

Yes, particularly for product category and comparison prompts where AI recommendations increasingly shape purchase decisions. Start with a small, focused prompt set covering the brand's top selling categories rather than trying to cover the full catalog at once.

How do you handle a highly regulated client, like healthcare or finance, where AI answers must stay strictly accurate?

Prioritize monitoring for factual errors and misrepresentation over pure visibility growth, since an inaccurate AI answer about a regulated product can create real compliance risk. Loop in the client's legal or compliance team before publishing any new AEO focused content.

What changes about AEO strategy for a B2B client with a six month sales cycle versus a B2C client?

B2B prompts skew toward comparison and evaluation questions used deep in a research process, so content should support that longer decision cycle with detail and evidence. B2C prompts skew toward quicker, more immediate purchase decisions, so answers can be shorter and more direct.

Do AI platforms treat multi location franchise brands as one entity, or as separate local businesses?

This varies by platform and by how consistently the franchise's own web presence represents each location. Strong, consistent entity data across every location's listings tends to help models understand the brand as one connected entity rather than scattered separate businesses.

How do you prove value to a niche B2B client whose buyers rarely use consumer facing AI chat apps?

Focus the value story on how those buyers actually research, which increasingly includes AI features built into Google search itself, even if they never open a dedicated chat app. Track the platforms your specific buyers actually touch, not the platforms with the most general hype.

What is different about tracking AI visibility for a marketplace or aggregator brand instead of a single product brand?

Marketplace brands need prompts that test whether the model recommends the marketplace itself as a place to shop, separately from prompts testing whether specific sellers or products on that marketplace get mentioned individually. Track both layers, since they answer different business questions.