An API manufacturer can tell whether ChatGPT recommends them by running structured buyer-intent prompts across five journey themes: discovery, evaluation, product/capability, commercial and support. Typing your own company name into ChatGPT does not measure this. AEO Dose's documented methodology tests 10 real buyer questions across 2 platforms, and that structure is what produces reliable signal.

Disclosure: AEO Dose provides AI Visibility and AEO Readiness audits for pharma businesses, including the structured testing approach described in this article. We are naming that relationship so readers can weigh the recommendations with that context. The testing method described here can be run by any API manufacturer without engaging AEO Dose.

Why typing your company name into ChatGPT is not a visibility test

Searching your own company name confirms only that ChatGPT has some record of you, not that it recommends you to a buyer who does not already know you exist. The first thing an API manufacturer usually does when they hear about AI visibility is open ChatGPT and type their own name. The assistant returns a paragraph, the name appears, and the team concludes they are visible. That conclusion is wrong, because no buyer sourcing a solid dose intermediate or a GMP-certified API starts by naming a supplier they have never heard of.

The mechanism this test misses is simple. A buyer asks ChatGPT an open question about a capability or a product category. ChatGPT then selects which suppliers to surface. When you prompt it with your own name, you have removed the selection step entirely. You are no longer testing recommendation. You are testing recall, and recall is not procurement signal.

This matters because buyers increasingly cross-check suppliers across multiple sources before issuing an RFQ, and AI assistants are becoming one of those sources. If you want to understand why AI answers carry procurement weight at all, our work on why answer engines differ from traditional search for pharma buyers covers how a missing brand becomes a commercial problem, not a cosmetic one.

What pharma buyers actually ask ChatGPT about API suppliers

Buyers ask ChatGPT capability and comparison questions, not branded ones. A procurement lead or formulation scientist does not type a supplier name. They type the problem they are trying to solve. The pattern of buyer prompts clusters around a few recognizable shapes: questions that ask who supplies a specific API category with a given certification, questions that ask who has regulatory documentation on file for a particular molecule, questions that compare CDMOs across intermediates and finished API at a given scale, and questions that ask who can support regulatory filings in specific geographies.

Notice what these prompts have in common. They are built around certifications, regulatory documentation, dose form, scale and geography. These are the attributes a buyer qualifies against. When a company assumes its regulatory certifications automatically translate into AI visibility, this is where the gap appears: the certifications exist, but if the company's own pages do not describe them in a way ChatGPT can extract and attribute, the assistant has nothing to surface in response to the exact question a buyer is asking.

The five buyer journey themes and why they frame the test

The five buyer journey themes (discovery, evaluation, product/capability, commercial and support) each cover a distinct stage of how a buyer moves from an open question to a shortlist, and testing across all five is what turns a single lucky result into a reliable signal. AEO Dose's documented methodology tests 10 real buyer questions across 2 platforms, and those questions are distributed across these themes so that no single stage of the buyer journey is left unmeasured.

Here is what each theme covers:

  • Discovery prompts are the open category questions: who supplies a given type of API at all. This is where a buyer with no existing shortlist starts.
  • Evaluation prompts compare named options against each other: reliability, track record, regulatory standing.
  • Product/capability prompts probe specifics: dose forms, intermediates, molecule classes, manufacturing scale.
  • Commercial prompts touch terms a buyer weighs before contact: minimum order quantities, lead times, geographic supply.
  • Support prompts cover the post-shortlist questions: documentation support, regulatory filing assistance, change control.

The reason one prompt is not enough is that each theme can produce a different result. A company might surface cleanly on a discovery prompt and vanish entirely on a product/capability prompt because its pages name the certification but never name the dose form. Testing all five themes shows you where in the buyer journey the signal breaks, which a single search can never reveal. This five-theme structure is AEO Dose's own testing framework, not an industry standard, and it reflects founder M Netrajit Singha's pharma background: a B.Pharma graduate with over twelve years in Pharma ERP implementation and Computer System Validation, which is what shaped the decision to design prompts around regulatory and capability attributes rather than generic brand terms.

Reading the three possible outcomes: named, named without link, not named at all

Every prompt you run returns one of three outcomes, and each one means something different for your procurement visibility. Understanding the distinction is the whole point of the test.

Named with a link or source. ChatGPT surfaces your company and attributes it to a source the buyer can follow. The assistant both knows you and can point the buyer to verification.

Named without a link or source. ChatGPT mentions your company but offers no citation. A buyer evaluating suppliers wants to verify a claim before acting on it, and an unattributed mention gives them nothing to check. It can also mean ChatGPT is working from older or thinly sourced information about you, which is harder to influence.

Not named at all. ChatGPT surfaces other suppliers in response to a question you should answer, and your company is absent. When competitors appear and you do not, the mechanism is usually that their pages describe the queried capability in extractable terms and yours do not. The assistant surfaces what it can confidently attribute. The absence is not a verdict on your manufacturing quality. It is a verdict on whether your content gave the assistant something to work with.

How ChatGPT differs from Gemini, Perplexity, Claude and Grok in what it surfaces

ChatGPT, Gemini, Perplexity, Claude and Grok draw on different sources and surface suppliers differently, so a result on one does not predict a result on another. Each assistant has its own retrieval and sourcing behavior, and those behaviors change over time as the platforms update.

The practical consequence for an API manufacturer is that visibility is not a single number. You can be named cleanly on one platform and absent on another for the same buyer question. This article is not arguing that ChatGPT visibility matters more than visibility on the other assistants. It is arguing that whatever you measure, you measure it per platform, because the platforms do not share one answer. That is why the Free Check tests across 2 platforms rather than treating ChatGPT as the whole picture.

What to do after you find a visibility gap

Once you find a prompt where competitors appear and you do not, the next step is diagnosis of why your pages gave the assistant nothing to surface, not a rush to publish more content. A gap in the product/capability theme points to pages that do not describe dose forms, molecule classes or scale in extractable terms. A gap in the evaluation theme points to a thin third-party footprint the assistant can cite. The gap tells you where to look, not what to write blindly.

This is where a structured audit replaces guesswork. Our walkthrough of what a Free AI Visibility Check includes and how a full readiness audit builds on it covers how the diagnostic moves from a Free Check into a full readiness assessment and an ongoing measurement cycle, so you are fixing the specific mechanism behind each gap rather than optimizing everything at once.

Here is the action to run this week. Pick five buyer questions, one from each theme: a discovery question about your core API category, an evaluation question comparing suppliers in your space, a product/capability question naming your specific dose form or molecule class, a commercial question about lead times or geography, and a support question about regulatory documentation. Run all five on ChatGPT without ever typing your own name. Record which outcome each returns: named with a link, named without a link, or not named. That record is your first real visibility signal.

To see where you stand across platforms with questions designed for pharma procurement, Get Your Free AI Visibility Check.

Frequently asked questions

Why does typing my company name into ChatGPT not tell me if I'm visible to buyers?
Typing your own name tests recall, not recommendation. It confirms ChatGPT has a record of you, but no buyer sourcing an unfamiliar supplier starts by naming them. Buyers ask open capability and comparison questions, and ChatGPT then selects which suppliers to surface. When you prompt with your own name, you remove that selection step entirely, so you learn nothing about whether the assistant would surface you to a buyer who does not already know you exist.
What kinds of questions do pharma buyers actually ask ChatGPT about API suppliers?
Buyers ask capability and comparison questions built around certifications, regulatory documentation, dose form, scale and geography. The pattern clusters around who supplies a given API category with specific GMP certification, who has US DMF documentation on file for a molecule, which CDMOs handle both intermediates and finished API at commercial scale, and who can support regulatory filings in specific geographies. They do not type supplier names. They type the sourcing problem they are trying to solve.
What does it mean if ChatGPT names competitors but not my company?
It usually means competitors' pages describe the queried capability in terms the assistant can extract and attribute, while yours do not. ChatGPT surfaces what it can confidently source. The absence is not a judgment on your manufacturing quality or certifications. It is a signal that your content did not give the assistant something extractable to work with for that specific buyer question. The fix starts with diagnosing which theme the gap sits in.
How many prompts do I need to run to get a reliable signal?
One prompt is not enough because each buyer journey theme can produce a different result. You might surface on a discovery prompt and vanish on a product/capability prompt. AEO Dose's methodology tests 10 real buyer questions distributed across the five themes so no stage of the buyer journey is left unmeasured. For a quick self-test, run at least five prompts, one from each theme, to see where in the journey your visibility breaks.
How is ChatGPT visibility different from visibility on Gemini, Perplexity, Claude or Grok?
Each assistant draws on different sources and surfaces suppliers differently, so a result on one does not predict a result on another. Each platform has its own retrieval and sourcing behavior, and those behaviors change over time. The practical consequence is that visibility is not a single number. You measure it per platform, because the platforms do not share one answer, and the Free Check tests across 2 platforms rather than one.
What should an API manufacturer do after they identify a visibility gap?
Diagnose why your pages gave the assistant nothing to surface before publishing more content. The theme where the gap appears tells you where to look: a product/capability gap points to pages missing dose form or scale descriptions, an evaluation gap points to a thin third-party footprint. A structured audit then replaces guesswork by matching each gap to its underlying mechanism, so you fix the specific cause rather than optimizing every page at once.