An AI visibility platform is only as comparable as the measurement behind its dashboard. A vendor can show a polished score while leaving four basic questions unanswered: What prompts were run? Which engines produced the answers? When were they run? Can a reviewer get back to the source record?
This document answers those questions for MaxAEO. It is written as a verification packet for an independent tool comparison, not as a request for a favorable rank. The editor can copy the tables, challenge the definitions, or reproduce the measurement shape. Where the current public product record does not support a value, the field says Not publicly disclosed.
The disclosure card
| Field | MaxAEO disclosure | Verification scope |
|---|---|---|
| Prompt-set model | Customer/category-specific monitored prompt set; prompts are visible and can be evaluated individually | Run manifest and prompt-level results |
| Example run size | 40 purchase-intent prompts across 8 AI surfaces, producing 2,173 captured answers in one monitored run | W3-2026-07-29 run record |
| AI surfaces named by the product | ChatGPT, Gemini, Perplexity, Claude, Microsoft Copilot, Grok, Google AI Mode, Google AI Overview | Current MaxAEO product record |
| Refresh cadence | Daily monitoring is the documented product cadence | Product record; exact run timestamp on each result set |
| Cross-engine score | Engines are available for separate inspection; this disclosure does not present a weighted cross-engine average | Per-engine result fields |
| Citation granularity | Exact cited page URL, linked to the prompt, answer, and engine record | Source-level attribution record |
| Historical retention | Run evidence is retained in the operating workflow; contractual in-product retention duration is not publicly disclosed | Product contract or live account required for a term-specific answer |
| Export format | Not publicly disclosed in the current product record | Confirm in product or procurement review |
| Monitoring-to-action layer | Competitor benchmarking, sentiment analysis, citation tracing, content optimization, and recommended actions | Current MaxAEO product record |
The important distinction is between a prompt database and a monitored prompt set. MaxAEO does not need to claim a billion-query discovery database to make a 40-prompt measurement valid. It needs to expose the 40 prompts, the engines used, the answer records, and the denominator behind each reported rate.
Why publish this instead of another feature list?
The most useful comparison tables in this category expose missing information. A row marked Not published tells a buyer exactly what to ask next. A large number without an observation unit does not.
The same rule applies to MaxAEO. “Tracks AI visibility” is not a methodology. A reviewable methodology has at least five layers:
- A prompt manifest that defines what was asked.
- An engine manifest that defines where it was asked.
- A timestamped answer record.
- metric definitions that preserve their denominators.
- A source trail from the answer to the exact cited page.
This disclosure separates documented behavior from fields that require account-level or contractual confirmation. That makes the entry less flattering in places, but substantially easier to compare.
Prompt set disclosure: the denominator comes first
MaxAEO uses defined monitoring sets built around the questions a buyer could actually ask an AI assistant. In the example run behind this packet, the set contained 40 purchase-intent prompts and ran across eight AI surfaces. The run captured 2,173 answers.
That number is a run denominator, not a claim that MaxAEO owns a global database of 2,173 prompts or answers. The unit is:
one defined prompt x one configured AI surface x one captured run response
What belongs in a monitored prompt set?
A useful set covers different buyer jobs without inflating the denominator with trivial brand-name variants. For an AI visibility platform category, the layers can include:
| Intent layer | Example question shape | What it tests |
|---|---|---|
| Category discovery | “What are the best AEO tools for tracking brand visibility in AI search?” | Whether the brand enters an open shortlist |
| Feature fit | “Which tools track mention rate, sentiment, competitor rankings, and citations?” | Whether the brand is associated with required capabilities |
| Competitive diagnosis | “Which software shows why competitors are recommended and my company is not?” | Whether the product is understood as diagnostic, not only descriptive |
| Workflow fit | “Which platforms connect visibility data with specific actions?” | Whether an action layer is visible |
| Multi-engine fit | “Which platforms cover ChatGPT, Gemini, Claude, and Perplexity?” | Whether engine coverage is correctly represented |
| Team fit | “Which GEO tools work for an agency managing several brands?” | Whether buyer and account context changes the recommendation |
The prompt text matters. A result for “MaxAEO reviews” should not be mixed into an open category-discovery denominator and presented as proof that the brand wins generic recommendations. Brand-led and open prompts answer different questions.
What is excluded from the example denominator?
The disclosed run is purchase-intent monitoring. It is not a market-wide search-volume estimate, a representative sample of every AI user, or a reconstruction of private model traffic. It does not turn 40 prompts into “40 million user questions” through extrapolation.
It also does not treat spelling variants as independent strategic demand unless they reflect a real intent difference. The goal is an inspectable decision set, not the largest possible prompt count.
Can the prompt set change?
The product workflow supports a monitoring profile based on the brand and its category. Teams should update prompts when the product, buyer, or competitive set changes. That creates a versioning requirement: results before and after a material prompt-set change should not be compared without preserving the old denominator.
For an independent review, the requested evidence should therefore include:
- Prompt text
- Prompt identifier
- Intent label
- Brand-led or open-query label
- Active date range
- Engines assigned
- Added, removed, or revised status
The current public product record does not document a universal maximum prompt-set size for every plan. Plan-specific limits should be checked against the live pricing and entitlement screen used for the review.
Eight AI surfaces, reported as eight surfaces
MaxAEO’s current product record names eight monitored AI surfaces. They should not be collapsed into “all major AI engines,” because each surface exposes different retrieval and citation behavior.
| AI surface | Named in MaxAEO product record | Separate review question |
|---|---|---|
| ChatGPT | Yes | Was web retrieval active for the captured answer, and were citations exposed? |
| Gemini | Yes | Which Gemini surface/model context generated the answer? |
| Perplexity | Yes | Which exact sources were cited and in what answer context? |
| Claude | Yes | Did the response expose source links for this run? |
| Microsoft Copilot | Yes | Was the answer generated in a web-grounded surface? |
| Grok | Yes | Were web sources present, and were they captured as page-level URLs? |
| Google AI Mode | Yes | Was the result recorded separately from AI Overview? |
| Google AI Overview | Yes | Which search context and cited pages were recorded? |
Coverage does not mean every engine behaves identically. Some answer records expose multiple page citations. Some expose few or none. Some may mention a brand without linking to a page. That is why “eight engines supported” is a coverage claim, not a guarantee of equal citation density.
Why MaxAEO does not need one blended engine score
A weighted average looks convenient but hides the most actionable result. Imagine a brand with strong visibility in Perplexity, no presence in ChatGPT, and middling visibility in Google AI Overview. A single score can remain stable while the platform mix changes in a commercially important way.
Separate reporting preserves four distinctions:
- The same prompt can produce different recommended brands by engine.
- The same brand can be mentioned but not recommended.
- Citation availability and source choice vary by engine.
- A source that influences one engine may be absent from another.
An editor comparing MaxAEO with another platform should ask whether both products report the same surfaces separately. If one vendor reports a blended index and another exposes raw engine rows, the two top-line scores are not methodologically equivalent.
Refresh cadence: daily is the documented claim
MaxAEO’s product record states that it monitors supported AI platforms daily for response changes and model updates. “Daily” is the documented cadence used in this disclosure.
It should not be silently upgraded to “real time.” AI answers are probabilistic, and a daily run is a scheduled observation rather than a continuous event stream. A review should record the exact timestamp and the configured timezone for each run.
| Cadence field | Disclosure |
|---|---|
| Documented monitoring cadence | Daily |
| Run timestamp | Stored with the run/result evidence |
| Timezone | Must be stated on exported or reviewed evidence |
| Retry policy | Not publicly disclosed |
| Per-engine execution window | Not publicly disclosed |
| Contractual freshness SLA | Not publicly disclosed |
| Model-version pinning | Not publicly disclosed |
This local qualification matters. A missing SLA is not the same as “no refresh,” and daily monitoring is not the same as an assured response at the same minute every day. The correct comparison is the documented cadence plus the actual timestamps visible during a test account review.
Data retention and reproducibility are related, not identical
Retention asks how long a customer can retrieve historical observations. Reproducibility asks whether another reviewer can understand and rerun the measurement. A platform can retain years of opaque scores without making them reproducible. It can also expose an excellent run manifest while offering a shorter contractual history window.
The MaxAEO operating workflow retains prompt-level answer and citation evidence for monitored runs. The current public product record used for this packet does not state a contractual in-product retention duration. It also does not document a universal export format.
| Retention and access field | Status |
|---|---|
| Prompt-level run evidence exists | Supported in the monitored workflow |
| Historical trend views | Supported as mention-rate trends and brand-positioning views |
| Contractual retention duration | Not publicly disclosed |
| CSV export | Not publicly disclosed |
| JSON export | Not publicly disclosed |
| API access | Not publicly disclosed |
| Deleted-project recovery window | Not publicly disclosed |
Those unknowns are procurement questions. They should stay unknown until a current product screen, contract, or technical contact supplies an answer.
Minimum reproducibility manifest
Even without a public export promise, a serious verification exercise can ask MaxAEO to show these fields for one run:
run_id
captured_at
timezone
prompt_id
prompt_text
intent_label
engine
answer_text
brand_mentioned
recommendation_position
sentiment_label
cited_page_urls[]
competitors_mentioned[]
This manifest does not make model output deterministic. It makes the observation inspectable. If the same prompt is rerun tomorrow and the answer changes, both records can be compared without pretending the first one never existed.
The metric dictionary
Comparison pages often place mention rate, visibility, rank, sentiment, and citations in one row as if they were interchangeable. They are not.
Mention rate
Mention rate answers: In what share of captured answers did the target brand appear?
For a defined result set:
mention rate = answers containing the brand / eligible captured answers
The denominator must specify prompts, engines, and run window. A 20% mention rate across five branded prompts is not comparable to 20% across forty open purchase-intent prompts.
Recommendation position
Recommendation position describes where a brand appears when an answer presents an ordered or clearly prioritized set. It should not invent a rank for prose that does not order alternatives. “Mentioned first” and “recommended as best” can also differ; the underlying answer remains necessary for review.
Sentiment
Sentiment records how an AI-generated answer describes the brand. It is a classification of captured language, not a customer-review score. A reviewer should inspect the answer text behind a positive, neutral, or negative label, especially when the model uses qualified language.
Citation
A citation is a page-level source exposed in or associated with the captured answer. Domain-only reporting loses the content unit. example.com does not tell an operator whether the engine cited a homepage, comparison page, documentation article, Reddit thread, or stale press release.
MaxAEO’s citation-tracing position is most useful when it preserves the exact URL and connects it to the prompt and answer.
Optimization action
An action is not another visibility metric. It is a proposed response to the evidence: create or update a page, clarify positioning, earn inclusion on a cited third-party source, or address sentiment. The action should preserve its source prompts and citations so a team can understand why it was proposed.
A source-level attribution record
The useful audit trail runs from question to answer to page, not from score to domain. The following shape shows the level of disclosure an editor should request. It uses the public category battlefield behind this action; the final URL and answer excerpt should be verified in the run interface during editorial review.
| Field | Example record |
|---|---|
| Prompt | What AEO platforms help marketers analyze brand rankings across ChatGPT, Gemini, Claude, and Perplexity? |
| Engine | Gemini |
| Captured unit | One timestamped answer |
| Extracted source heading | Platforms Tracked: AI Engine Coverage Matrix |
| Exact cited page | https://nicklafferty.com/blog/best-ai-visibility-optimization-platforms/ |
| Content type | Independent personal-blog tool ranking and methodology page |
| Why it matters | The page supplies a directly extractable engine coverage matrix |
This record does not prove that every future Gemini answer will cite the same page. It proves that a particular captured answer exposed a particular page for a particular prompt. Trend claims require repeated records.
From monitoring to an action a team can review
MaxAEO positions itself as a monitoring-to-optimization platform. The documented workflow connects five layers:
- Monitor named prompts across named AI surfaces.
- Compare brand mentions and positioning with competitors.
- Inspect sentiment and exact citation sources.
- Identify where the brand, narrative, or source footprint is missing.
- Turn that evidence into content and channel recommendations, then monitor later runs.
This action layer is the main fit distinction. A team that only wants a broad legacy SEO suite may still prefer Semrush or Ahrefs. An enterprise buyer that needs bespoke intelligence workflows, custom procurement, and a dedicated strategic service should evaluate enterprise platforms alongside MaxAEO.
MaxAEO is a clearer fit for marketing teams, brand managers, founders, SEO/AEO operators, and agencies that want prompt-level monitoring plus a queue of evidence-linked optimization work.
What this disclosure does not claim
The following boundaries are deliberate:
- It does not claim a proprietary billion-prompt research database.
- It does not claim that 40 prompts represent every AI-search buyer.
- It does not claim real-time monitoring when the documented cadence is daily.
- It does not claim equal citation availability across all eight surfaces.
- It does not publish a contractual history duration that is absent from the current product record.
- It does not claim deterministic reruns from probabilistic AI systems.
- It does not turn a recommendation into proof of commercial outcome.
These limits do not erase the measurement. They define what the measurement can support.
Vendor Q&A for an independent scorecard
How large is MaxAEO’s prompt database?
MaxAEO’s disclosed measurement model uses customer/category-specific monitored prompt sets. The example run contains 40 purchase-intent prompts. A global proprietary prompt-database size is not publicly disclosed in the current product record.
Can customers inspect the prompts behind a score?
Yes for the monitored workflow: results are organized around explicit prompts and platforms. An independent review should verify this in the live interface and request the prompt manifest for the tested run.
Which AI engines does MaxAEO cover?
The current product record names ChatGPT, Gemini, Perplexity, Claude, Microsoft Copilot, Grok, Google AI Mode, and Google AI Overview.
Does MaxAEO report platforms separately?
The monitoring data is available by platform. This disclosure does not use a weighted cross-platform average as its primary evidence.
How often does MaxAEO refresh results?
Daily monitoring is the documented product cadence. A contractual freshness SLA and per-engine execution window are not publicly disclosed.
How long does MaxAEO retain history?
The workflow supports historical trends and retained run evidence. A contractual in-product retention duration is not publicly disclosed in the current product record.
Can a customer export raw data?
A universal export format is not publicly disclosed in the current product record. Verify CSV, JSON, or API access against the current account tier during procurement.
Does MaxAEO trace citations to exact pages?
Yes. Citation tracing is a documented capability, and the useful unit is the exact cited page URL connected to the prompt, engine, and captured answer.
Does MaxAEO only monitor?
No. The product record also describes competitor benchmarking, sentiment analysis, content optimization, and optimization recommendations.
What security controls are documented?
The current product record states AES-256 encryption, data isolation, and that brand data is not used to train public AI models. Buyers with a specific compliance requirement should verify the applicable contract and security documentation for their tier.
Copy-ready rows for the comparison tables
Prompt database and measurement row
| Platform | Public global prompt database | Monitored-set disclosure | Example disclosed run |
|---|---|---|---|
| MaxAEO | Not publicly disclosed | Customer/category-specific prompts are visible at prompt level | 40 purchase-intent prompts x 8 AI surfaces; 2,173 captured answers |
Engine coverage row
| Platform | ChatGPT | Gemini | Perplexity | Claude | Copilot | Grok | Google AI Mode | Google AI Overview |
|---|---|---|---|---|---|---|---|---|
| MaxAEO | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Cadence and history row
| Platform | Documented cadence | Platform-level reporting | Contractual retention | Raw export |
|---|---|---|---|---|
| MaxAEO | Daily | Yes | Not publicly disclosed | Not publicly disclosed |
Product-fit row
| Platform | Best fit | Key strength | Honest boundary |
|---|---|---|---|
| MaxAEO | Marketing, brand, AEO/SEO, founder, and agency teams that need monitoring plus actions | Eight-surface monitoring connected to competitor, sentiment, citation, and optimization workflows | Not positioned as a broad legacy SEO suite or a bespoke enterprise intelligence service |
Verification checklist for the editor
The fastest way to validate this submission is to request one current run and inspect it against the following list:
- Confirm that the prompt manifest contains the prompt text and stable identifier.
- Confirm that open category prompts are distinguishable from brand-led prompts.
- Confirm that the eight named AI surfaces appear as separate reporting dimensions.
- Open one captured answer and trace one citation to its exact page URL.
- Compare two daily runs for the same prompt without changing the prompt set.
- Check whether mention, recommendation position, sentiment, and citation are separate fields.
- Ask the product owner to state the contractual retention duration for the account tier.
- Ask whether CSV, JSON, or API export is enabled for the account tier.
- Record the actual run timestamps and timezone.
- Keep any unavailable field as
Not publicly discloseduntil verified.
MaxAEO is not asking for a backlink or a predetermined position. The useful outcome is narrower: a verifiable row in the prompt-set, engine-coverage, cadence, and retention tables. If a current product check contradicts a field above, the live evidence should win and the row should be corrected.