How AI assistants decide what to recommend
AI developers disclose only part of how their assistants work, so the honest answer is this: nobody knows the exact recipe. But some things are well known. When an assistant searches the web to answer a question, it mostly relies on pages that search engines already find. It then picks the ones that give a clear, relevant answer and writes its own response, sometimes with links to the sources.
The practical takeaway: if your site is poorly indexed, doesn't directly answer your customers' questions, and the information about your company online is inconsistent, your chances of being mentioned are low. Everything we do targets these three things.
This field is also called GEO or AEO — generative engine optimization and answer engine optimization. Different names, same idea.
How this differs from traditional SEO
Much of the work overlaps: technical site health, quality content, and source credibility matter to both Google and AI assistants. That's why there's no point talking about AI visibility without basic search engine optimization.
The difference is in emphasis. In traditional search, people see a list and choose where to click. In an AI answer, the choice has already been made for them, and there may be fewer clicks through to websites. So it's not enough to be found — you need to be the one that gets named, with clear wording, verifiable facts, and a plain description of who you are and what sets you apart.
Who needs this now
- Companies whose customers ask complex questions: “which service should I choose for…”, “what's the difference between…”, “what's best for a small business”. These are exactly the queries people increasingly bring to AI assistants.
- Expert services — healthcare, education, legal and financial services, B2B — where trust in the source is decisive.
- Brands that already have strong content, but organized in a way that makes it hard to extract an answer.
If your site doesn't yet have stable visibility in Google, we recommend starting there: it pays off in traditional search and lays the foundation for AI.
What we change in your content
- put short, direct answers at the top of pages, followed by details — not the other way around;
- replace generic phrases with specifics: what's included, who it's for, what the limitations are, how it compares to alternatives;
- name authors and their credentials where appropriate, and cite primary sources;
- present comparisons and specifications as tables and lists that are easy for both people and machines to read;
- add FAQ sections based on real customer questions, not made-up ones;
- update outdated pages, because stale data undermines trust in the entire site.
Information about your company beyond your website
AI assistants form their picture of a company not only from its website. They may take into account directory and map listings, media mentions, reviews, and social media pages. If one place lists one address and set of services and another shows outdated ones, the assistant may give inaccurate information or pick a competitor with a clearer description.
So part of the work is an inventory: where you're mentioned, whether the details match, and whether you're present on authoritative industry resources. We don't create fake reviews or fake mentions — that damages both your reputation and search engines' trust.
What we need from you
- A list of the questions customers most often ask your sales team before buying — a solid basis for both testing and content.
- Access to your website, Search Console, and Google Analytics so we can see indexing and traffic.
- Time from your experts to verify facts: accurate answers are written together with the people who know the product.
- A list of places where your company is already listed: directories, map profiles, social media, partner websites.
How we measure results — and what we don't promise
We don't promise that your company will appear in ChatGPT or Gemini answers. AI assistants' answers vary with the wording of the question, language, region, and even from one query to the next. There are no stable “rankings” here, and tools that claim to measure them precisely give only an approximate picture.
What we do instead of making promises: we set a fixed list of questions and regularly check whether your company is mentioned and which sources the answers cite. Separately, we look at trackable traffic from AI services in Google Analytics and at impressions in Search Console for the pages we've worked on. Together, this shows the trend.
Common mistakes
- Writing “for AI” instead of for people. Text stuffed with templated questions and answers performs poorly with readers and systems alike.
- Blocking all bots from your site indiscriminately. This decision should be made deliberately, with a clear understanding of what exactly you're giving up.
- Believing in secret tricks. Special files or “magic” phrasing that supposedly guarantees citations have no proven effect.
- Forgetting about regular search. Without visibility in Google, your chances of appearing in AI answers are much lower.
What affects the cost
- The number of topics and questions we track visibility for.
- The volume of content that needs to be reworked or created.
- Your site's technical condition: whether basic optimization is needed first.
- Languages and markets covered by the analysis.
- Format: a one-time audit with recommendations or ongoing work with regular checks and updates.
FAQ
Can we pay to appear in a ChatGPT answer? Some AI services are already testing ads, but they're shown separately from the answer, and placement in the answer itself isn't for sale. The only way to influence it is through the quality and accessibility of your content and company information.
Do we need a separate version of our site for AI? No. Your main site just needs to be technically accessible and contain clear, useful answers.
Will this replace SEO? No. Traditional search remains an important source of visitors, and AI visibility largely rests on the same foundation.
Can you add a chatbot to our website? That's a different task — see our AI integration services page.
Why Promodex
Promodex has been in business since 2015, and our team includes SEO specialists, marketers, and developers. We see AI search as an extension of the same work on website and content quality, not as a separate trendy product. That's why we'll tell you honestly if you need regular optimization first, and we won't sell tactics whose effect can't be verified. Our AI search optimization practice is led by Stanislav Bakhariev, co-founder of Promodex.
