How do you get ChatGPT to recommend your company?
There is no switch. No guaranteed schema. No secret prompt you add to your homepage. And no responsible consultant can promise that ChatGPT will recommend your business first.
What you can do is make your company much easier to discover, understand, verify and compare. That is the real work.
Start with a slightly uncomfortable test
Open ChatGPT. Do not search your company name. Ask the questions a real buyer would ask. For example: "Which marketing attribution consultants would you recommend in South Africa?" "Which companies can help a B2B business improve its visibility in AI search?" "Who are the strongest AEO consultants in South Africa?" "Compare three options and tell me which one suits an enterprise business."
Now look at what happens. Are you there? Who appears instead? What are they being praised for? Which information is wrong? Which sources are cited? That tells you far more than asking "Tell me about my company."
There is no single "ChatGPT ranking"
This is important. AI responses are not fixed ten-blue-link rankings. Responses can change based on the prompt, wording, context, product, model, available sources, date and user circumstances. So when somebody says "We ranked you number one in ChatGPT", ask: For what? A serious ChatGPT optimisation programme needs a defined prompt universe. Not one nice screenshot.
1. Make sure ChatGPT can reach the information
This is basic but important. OpenAI says public websites can appear in ChatGPT search and recommends not blocking OAI-SearchBot if you want content to be discoverable, surfaced and clearly cited. So check robots.txt, page accessibility, noindex directives, server responses, canonicalisation and whether important content is actually present in HTML. If the useful information is inaccessible, optimisation ends very quickly.
2. Make the company easy to understand
A surprisingly common problem is that a website never clearly answers: Who is this company? What exactly does it do? Where does it operate? Who is it for? What makes the offer different? Humans can often infer the answer from vague marketing language. Machines also have to reconcile that information with everything else they find. Make the facts boringly clear. Use consistent company names, service descriptions, locations, product names, founder/expert information and commercial terminology. Clarity is not unsophisticated. Clarity is useful.
3. Build pages around questions buyers actually ask
Do not optimise only for keywords. Build around decisions. A buyer may ask: What is AEO? Then: Does AEO replace SEO? Then: Is AEO worth investing in? Then: Which AEO consultant should I hire? Then: Compare these three firms. Those are different information needs. A site covering the entire decision journey gives systems considerably more context than one generic service page. That is exactly why QFD is building this content library.
4. Answer the question before writing the essay
If the heading asks "What is AEO?", answer it. Do not start with "In today's rapidly evolving digital landscape..." A strong answer block might say: Answer Engine Optimisation is the practice of improving how easily search and AI systems can understand and use a brand's information when generating direct answers. Then expand. Good information architecture works for people and retrieval systems. That is not a ChatGPT trick. It is good writing.
5. Publish information other people have reason to cite
This is where the biggest opportunity sits. The internet already has enough articles summarising other articles. Create things that are harder to replace. Examples: original research, benchmarks, proprietary datasets, practical calculators, experiments, real case studies, technical methodologies, useful comparison tables, expert analysis, original frameworks grounded in evidence. If ten sites explain the same idea and one contains the original dataset, which would you rather reference? That is the standard.
6. Improve the information outside your website
This part is frequently overlooked. Your website is not the only description of your company. Other sources may include publications, profiles, directories, customer reviews, partner sites, industry bodies, podcasts, interviews, comparison pages, news coverage and community discussions. Ask: Does the wider web tell a coherent story about us? If your website says "Enterprise attribution consultancy" while every third-party source describes you as "Small PPC agency", you have an information problem. That cannot always be solved with another blog article.
7. Give important claims evidence
There is a difference between saying "We are experts" and demonstrating relevant experience, methodology, case studies, results, research, credible third-party references and identifiable expertise. Do not create fake authority. Document the authority that actually exists. For QFD, that means things such as real campaign-management experience, real attribution work, real AEO implementations, actual case studies and original research as it becomes available. Proof compounds. Claims do not.
8. Keep the information current
AI visibility becomes especially painful when old information survives. Common problems include old prices, discontinued products, former locations, outdated staff and superseded service descriptions and old comparison pages. Finding the stale source can be more important than rewriting the current website. QFD has already dealt with this for a national brand where outdated product and pricing information persisted inside AI answers. The objective was not simply more visibility. It was the correct answer.
See the Making a National Brand the AI Answer case study.
What about schema?
Use it where it accurately describes visible information. Structured data can help machines interpret relationships and facts. It is useful. It is not a magic ChatGPT recommendation switch. Google similarly continues to recommend valid structured data that matches visible page content for its search experiences. The principle is: make information clearer. Not: add every schema type you can find.
What about backlinks?
Authority still matters. But the more useful question for AI visibility is broader: Which sources shape the category? A traditional link may help organic authority. A respected third-party mention may also help establish reputation, category association, expertise and market context. So stop thinking only "How many links do competitors have?" Start asking "Which sources repeatedly explain, compare and validate this market?" That is a much more interesting research question.
Measure recommendations properly
Build a fixed prompt set. Track whether the brand appears, recommendation position, competitors, answer accuracy, citations, repeatability, platform and date. Then retest. Do not change the questions every time performance looks bad. Measurement works because the test remains reasonably consistent.
What not to do
Do not create 500 thin AI-written articles
Google explicitly warns that mass-generating pages without adding user value can violate its scaled-content spam policy. The same strategic problem exists beyond Google. More commodity content does not make the brand more distinctive.
Do not buy fake mentions
Manufactured reputation is not authority.
Do not optimise only for the brand name
If the system already knows the buyer wants your company, discovery has already happened. Optimise the questions that happen before the buyer knows who to choose.
Do not chase one screenshot
One response is not a market position.
Do not promise a guaranteed recommendation
Nobody outside the platform controls the answer.
The real goal
You are not trying to manipulate ChatGPT into saying your name. You are trying to create enough clear, useful and verifiable information that recommending the company becomes a reasonable answer. That is a much higher standard. It also produces a better website.
Find out what ChatGPT is already telling your buyers.
Then decide what actually needs fixing.