AI visibility research

AI market research for enterprise brands

Buyers now ask ChatGPT, Google AI, Perplexity, Claude, and Grok which products are worth a look. This study records those answers every day for 2 to 4 weeks and shows where each product in the line stands, who wins the recommendation, and which sources AI trusts.

Fixed price per study Dashboard from day one No subscription

Study scope Sample proposal · set per study
Product lines 3
Markets 4 countries, 3 languages
Buyer questions 40 a day, reviewed by the brand team
Fieldwork 21 days, every day
Assistants
ChatGPTGoogle AIPerplexityClaudeGrok
40 questions × 5 assistants × 21 days
4,200 recorded answers, each stored with its sources
Deliverable: written report and dashboard access Fixed price
The research gap

The channel current research skips

Enterprise teams already buy studies for every channel where buyers used to make up their minds. AI answers fall outside all of them.

Study What it reads How often Status

Brand tracking

Awareness and sentiment, read from panel surveys

2 to 4 waves a year In the budget

Usage & attitude

Why buyers choose, switch, and stay

Once a year In the budget

Media share of voice

Press, social, and advertising volume

Monthly In the budget

The AI conversation

What assistants tell buyers about the brand's products, with the sources behind each answer

Daily, for 2 to 4 weeks This study
The deliverable

What the AI visibility report covers

One chapter per question a brand team will ask. Product by product, market by market.

  1. 01

    AI share of voice

    How often each product gets named across assistants and markets, next to every competitor in the same answers.

  2. 02

    Positions per product

    Where each product lands when AI ranks the options, and who holds the top spots in each category.

  3. 03

    The sources AI reads

    The pages, directories, and publications AI reads before answering, and which of them work for competitors today.

  4. 04

    Assistant breakdown

    ChatGPT, Google AI, Perplexity, Grok, and Claude answer differently. The chapter shows where the brand leads on each of them and where it is invisible.

  5. 05

    Markets and languages

    The same questions asked the way local buyers ask them, so a strong home market never hides a weak one.

  6. 06

    Recommendations to brief out

    What to change, on which pages, in what order. An agency starts from a task list instead of a discovery phase.

Methodology

How we measure AI visibility

Assistants answer the same question differently from day to day. One afternoon of prompts produces an opinion. Weeks of daily answers produce a measurement.

Before day 1

Question set

Built from what buyers ask about each product line, in each market's own language. The brand team reviews it before the study starts.

Day 1

Dashboard opens

The brand team watches the data arrive while it collects, with every answer readable as it lands.

Days 1 to 21

Daily collection

The same questions go to ChatGPT, Google AI, Perplexity, Grok, and Claude every day, so one odd answer never skews the picture.

After the last day

The report

The written study. Any figure in it, whether a share of voice or a single quote, traces back to answers anyone can read.

The unit of evidence

What one recorded answer holds

Every figure in the report is a count over records like this one, and any of them can be opened and read.

  • Day Which day of the fieldwork
  • Assistant One of the five
  • Market Country and language it was asked in
  • Question Verbatim, as a buyer asks it
  • Answer Full text, as the assistant gave it
  • Products named Every product in the answer, the brand's and its rivals'
  • Position Order of first mention, counted from the text
  • Sources Every page the assistant cited
Built for enterprise

Made for how large organizations decide

A finding is only useful if it survives the company's internal process. The study is shaped for that from the start.

Procurement

One document to align on

The findings arrive as a report the team can discuss and approve at its own pace, without buying seat licenses before anyone has seen value.

Execution

A brief contractors can run with

Recommendations name the page, the claim it needs, and the order of work. The agency starts from a task list, and the discovery phase is already paid for.

Follow-up

A second measurement when it matters

After the changes ship, the same study runs again on the same questions and shows what moved. Decisions get a before and an after.

FAQ

Questions about AI visibility research

What is an AI visibility research report?

An AI visibility research report is a study of how AI assistants answer real buyer questions about a market. Over 2 to 4 weeks, the same questions go to ChatGPT, Google AI, Perplexity, Grok, and Claude every day. The report shows AI share of voice for every product in the line, positions against competitors, and the sources assistants rely on, with every figure traceable to a recorded answer.

How is this different from an AI visibility audit?

An AI visibility audit is usually a manual snapshot. Someone asks an assistant twenty or thirty questions on one afternoon and writes up what they saw. That is a fine first look, and it carries the mood of that one afternoon, because assistants answer the same question differently from day to day. A research report is built on daily measurement over weeks, so a number in it reflects a pattern, and a change in it means something moved.

How do you measure AI share of voice?

AI share of voice is the portion of recommendations a product wins across all answers to the questions buyers ask in its category. We record every product and competitor an assistant names, where in the answer it appears, and whether it is recommended or only mentioned. Because the same question set runs daily across all five assistants, share of voice can be compared between products, between assistants, and between markets.

Can we see what ChatGPT says about our brand before commissioning a study?

Yes. The free AI visibility checker at reachd.ai/ai-visibility-checker shows in about 30 seconds whether ChatGPT recommends a business and who it names instead. It is a single-question look rather than a study, and a fair way to decide whether the topic deserves a research budget.

Does the study work for regulated industries like pharma?

Yes, and it changes nothing on the brand's side while it runs. The study reads what AI assistants already tell the public, so there is nothing for legal or medical review to approve before it starts. The finished report then moves through the company's normal internal process, and the recommendations are written so existing contractors can act on them once approved.

What does an AI market research study cost?

Pricing depends on how many product lines, markets, and languages the study covers, which is why the proposal comes first. A study can scope a single product category or a whole portfolio. Every proposal states the question volume, the collection period, and a fixed price for the deliverable.

Is this a subscription?

No. It is a one-time study with a defined start and end, plus dashboard access while the data collects. Many teams commission a second measurement after their changes ship to see what moved, and some later switch to continuous AI brand monitoring, which is a separate product.

See what AI tells buyers

Tell us the product lines and markets. The proposal comes back with the question volume, the collection period, and a fixed price for the study.