Mazaya Digital
Request a working session
Menu
Request a working session

Where you stand in the AI assistants' answers

We track how AI assistants answer high-intent customer questions, which businesses they name, and which outside sources they cite. Also known as GEO.

“I've noticed a drop in organic traffic over the past several months, even on pages that still rank reasonably well.”

A question owners ask.

What customers ask

We put twenty real buyer questions to the assistants and record what comes back.

The customer asksWhich orthodontist near Ann Arbor treats adults with clear aligners?

  • ChatGPT

    • Three practices in the area take adult cases.
    • Your practiceNamedNamed, and first
    • Two competitors, both with a booking link.

  • Gemini

    • Lists four practices by distance.
    • Your practice is absentMissingNot returned at all
    • Cites a directory page you do not control.

  • Claude

    • Named in 1 of 3 checksSometimesSame question, same week
    • Absent on the other two.
    • One answer is not a position.
Illustrative exampleThe same customer question, put to three assistants on one dayOne of twenty questions in a scan

Your customers do not always click through a list of websites. They may ask an AI assistant for a recommendation, compare the businesses named in its answer, and make a decision without visiting every source. That creates a second visibility surface: your Google position does not tell you whether an assistant will mention you, and an AI mention does not tell you whether you appear across the local map. You need to measure both.

We call this AI Visibility work. The category is also known as generative engine optimization (GEO). Which label wins matters less than whether your business appears for the questions customers ask, whether the answer is accurate, and what sources support it.

A useful scan begins with the questions connected to a real buying decision. For a clinic, those compare treatment types, location, availability or provider fit; for a restaurant, what is served, what a place suits, and what is open late; for a home services company, the job, the service area, the property type or the urgency. We agree on twenty high-intent questions, the same set is tested on the leading assistants, and each dated run records whether your business is named, where it appears, which competitors are named, whether material details are accurate, and which websites are cited. The scan page shows a run across five assistants. The scan creates a baseline; it does not promise that the next one will contain more mentions.

AI answers do not follow the local pack

A business can lead the local map and be absent from an AI answer, and another can have limited map visibility and still appear prominently in an assistant’s response. We observed this split in our own scans across several cities. Google local results weigh relevance, distance and prominence around a search location, while assistants assemble answers from model knowledge, search results, business websites, lists, directories and articles, and the mix changes by question and system.

For one customer questionVisible on GoogleWeak on Google
Named by the assistantsStrong on both: keep the baseline and watch itAI only: the sources cite you and the map does not. The profile and the local pages get the first look
Absent from the assistantsGoogle only: the entity facts and the outside sources get the first lookAbsent from both: a wider relevance and authority review
The four cases a scan sorts each question into, and where we look first in each

A local rank map therefore cannot substitute for an AI Visibility scan, and the scan cannot substitute for call, form, or booking measurement.

Where the answers come from

An assistant draws on sources, and on what it already holds. Which sources, and whether they agree, is the work.

    • Name
    • Address and service area
    • Services
    • People and credentials
    • Hours
    The business factsOne approved set
    • Your website
    • Your Google Business Profile
    • Directories and associations
    • Publications and lists
    • Reviews
    The surfaces that carry themConsistent, not word for word. Your website is the entity home, usually the about page. The profile is yours to edit; the publisher decides the rest
    • A live search
    • What the model already holds
    The assistant's answerDrawn from whichever of these it reads on the day, from a live search, and from what the model already holds. Which ones it used shows only when it cites them

What has to holdThe same facts, consistent on every surface an assistant reads. A conflict on any of them can be the one an answer repeats.

Where an assistant's answer can come from

Your website needs a clear entity home

One page settles who you are, what you do, and where.

An assistant needs to understand that your business is one identifiable entity with specific services, locations, people and credentials. Your about page usually serves as that entity home: it states the approved facts plainly and connects them to the same organization referenced elsewhere on the site, with structured data supporting the reading through consistent identifiers. Repeated, consistent, verifiable facts are easier for search and answer systems to reconcile than conflicting claims across unrelated pages. What we review, and how conflicts are found and fixed, is on the entity and citations page.

Citations happen away from your website

Some sources an assistant cites are not yours to edit.

Assistants may draw on directories, professional associations, editorial lists, local publications and industry resources as well as your website, so publishing more pages on your own domain cannot fix every visibility gap. We inspect the sources cited or repeatedly surfaced for your question set and look for legitimate ways to correct, earn or strengthen your presence on the ones that matter. Outside publication remains the publisher’s decision, and citations and mentions are not guaranteed.

Priority pages need answerable sections

The first sentence answers the question, and the proof follows.

A customer may ask a narrow question instead of searching a short keyword, so priority pages carry concise, factual answer blocks: the first sentence answers the question, the next lines explain the condition, mechanism or limitation. That is not hundreds of generic question pages, and the page still has to help a person decide, because copy written only so a machine can lift it fails that test.

What we check, and when

The scan is a dated baseline that we run again.

AI answers can vary between runs, because models, connected search systems, source indexes, prompts and location context all change, so one favorable screenshot is not a stable measurement.

Your report uses a dated question set and a repeatable method. It separates:

  • Inclusion from omission
  • Accurate statements from factual errors
  • Direct citations from likely source influence
  • Your own website from third-party sources
  • Google visibility from AI Visibility
  • Visibility from attributed inquiries

Attribution from an AI answer remains imperfect. A prospect may say they found you through an assistant, arrive through an untagged path, or search your name later. We report observed visibility and identifiable source data without turning an absence into a confident attribution claim.

AI Visibility inside your growth system

AI Visibility is part of Market Visibility because it concerns where your business is discovered. It depends on Offer Design when the page must answer a buying question, and Insights when the observation must be repeated and compared.

What this work includes

What to fix first

Come with the project you already have in mind. Before we meet, we review your website, listings and other public-facing material. During the 30-minute session, you show us how the relevant part works today. We check whether the project addresses the real constraint and what it needs to connect to. You can hire us for one part.

Request a working sessionRead about the MOSAIC Six

Frequently asked questions

How do I get my company mentioned in ChatGPT?
Start by measuring the questions for which your company is absent, then inspect the businesses and sources that appear instead. Improvements may involve clearer service pages, consistent entity facts, and legitimate third-party references, but no one can guarantee a mention.
Is this what people call GEO, or answer engine optimization?
They are overlapping names for the same aim. Generative engine optimization (GEO), answer engine optimization (AEO), AI SEO and LLM SEO each describe, with a different emphasis, getting a business named accurately inside an AI-generated answer. We call it AI Visibility because that is what a business owner is asking about.
Is AI SEO different from normal SEO?
Yes, although the work overlaps. Traditional SEO measures positions and visits from search surfaces, while AI Visibility also measures inclusion, accuracy, competitors, and cited sources inside generated answers.
Can you guarantee that an AI assistant will recommend my business?
No. We guarantee the agreed scan and implementation work, not recommendations, citations, leads, or mentions controlled by another system.
How often should AI Visibility be measured?
Monthly measurement is useful when AI Visibility is an active service because it creates dated comparisons using the same commercial question set. A less frequent check can suit a business using it as a secondary part of Local SEO.
Does structured data make ChatGPT cite my website?
No. Structured data can clarify entities and page meaning, but it does not force an AI assistant to retrieve, trust, or cite a page.

Primarily serves the Market Visibility pillar of the MOSAIC Six.