AEO for Small Teams: Answer Engine Optimization with 1-2 People

At a 12-person machine manufacturer in Pécs, Zsófi, the marketing manager, suddenly heard this from the managing director: "I asked ChatGPT who the best suppliers in the region are. It didn't mention us." There's no AEO specialist and no SEO agency on a monthly retainer, just Zsófi and a part-time designer. At that point the question isn't whether to deal with this, but how to do it with one or two people, a few hours a day.

The short answer: for a small team, AEO (Answer Engine Optimization) is not a separate project but a reorganization of your existing content work. First, you give short, quotable answers to your 10-15 most important questions on your own site, in a structured way (lists, tables, FAQ, schema), and you add the external evidence that AI systems use to decide whether to trust you. You don't need a new toolset or a big budget, just discipline and the right order.

Key takeaways

• AEO doesn't replace SEO, it builds on it: AI answer engines typically cite pages that are already well indexed and clearly structured.
• For a small team, the secret is narrowing down: build pages around 10-15 real buyer questions instead of trying to answer 100 topics.
• A 40-60 word direct answer at the top of the page is far more likely to be cited than a line of reasoning that only unfolds after three paragraphs.
• AI systems also take external mentions into account (industry sites, reviews, professional forums), so your presence outside your own website is part of AEO too.
• You can measure results manually with monthly test questions; you don't need expensive monitoring software for that.

What is AEO, and how is it different from classic SEO?

The goal of classic SEO is to get your website ranked high in Google's results so that users click on it. The goal of AEO is for ChatGPT, Perplexity, Google AI Overviews or Copilot to quote you in their own answers, or at least mention you. These are two different games, but they're played on the same field.

The practical difference is that an AI system isn't looking for the most colourful or most-linked page, but for the passage it can safely use in an answer. That requires concise, clear, verifiable statements in a clean structure. On a fifty-person company's website, this often means taking the knowledge you've so far hidden on a long "About us" or "Services" page and breaking it down into specific questions and answers.

The good news for small teams: you don't have to rewrite everything. Most Hungarian SMEs already have the expertise; it's just packaged badly. So the first phase of AEO is repackaging, not producing new content.

How much time does it take for one or two people?

Realistically, it's worth setting aside 4-6 hours a week, in other words a part-time but disciplined resource. In the first two months it usually takes more, because the foundations (question list, schema, page structure) only need to be laid once. After that, maintenance takes less: adding new questions, updating old answers, measuring monthly.

Don't try to do everything at once. Anyone who tries to rewrite 30 pages in the first week gives up by the third. What works for a small team is taking one topic through the full process each week: question, answer, structure, schema, publishing, testing. After a month you have four finished pages, after a quarter twelve, and that already gives you a noticeable presence in a narrow professional field.

If there are two of you, split the work naturally: one is responsible for the professional content and collecting questions (the one who hears the customers), the other for the technical side (schema, page structure, measurement). If you're on your own, you do the same things in sequence rather than in parallel.

Which 10-15 questions should you start with?

The question list is the foundation of all AEO work, and this is where most mistakes happen. The mistake: adding generic questions copied from a keyword tool ("what is marketing"), which every website in the world already answers, so you will never be the cited source. A good question meets three conditions at once: your buyers actually ask it, you can answer it credibly within your field, and being in the answer has business value for you.

Where should you collect them? From your own channels, not from a tool:

• the questions your salespeople and customer service hear most often (ask them to note them down for a week);
• the subject lines and text of quote requests and emails;
• question-style searches in Google Search Console;
• Google's "People Also Ask" box for your main services;
• ChatGPT's or Perplexity's answers to questions in your field: who appears there now, and from which sources.

Turn this into a simple table: question, business value (1-3), current status (is there a page for it?), priority. In the first round, only include the 10-15 questions where the value is 3 and you don't yet have a good page. Leave the rest for later.

How should an AEO-friendly page be structured?

AI answer engines typically look for the answer at the top of the page. That's why the page structure should always be the same: the question (H1 or H2), immediately followed by a 40-60 word direct answer, and then the detailed explanation. If you give the answer only after three paragraphs of "setting the mood", the system simply skips it.

This structure is easier to understand with a concrete example. On a law firm's website, under the question "How long does it take to set up a company?", the first two sentences should be: "Setting up a limited liability company (kft.) typically takes 3-8 working days if the documentation is flawless and the court of registration doesn't request corrections." Then come the detailed explanation, the steps and the exceptions. A statement like this is quotable; a decorative introduction is not.

The table below shows the most common mistakes and the better approach:

The author and the update date may seem like small details, but for AI systems they are trust signals. A nameless page that hasn't been updated for years is less likely to be quoted than material recently maintained by a named expert.

What do you need to set up technically if you don't have a developer?

There are three technical basics that really matter for AEO, and all three can be handled without a developer or with minimal help.

The first is structured data (schema markup). The most important types are FAQPage for FAQ pages, Article for articles and Organization for your company. On WordPress, most SEO plugins (Yoast, Rank Math) handle this, and it isn't hard to add manually either. It's important that the schema content matches the visible text and doesn't include anything that isn't on the page.

The second is indexability and speed. If your site is slow, loads badly on mobile, or the content is locked behind JavaScript, neither search engines nor AI crawlers can see it properly. Check your robots.txt too: if you've accidentally blocked AI crawlers there (for example GPTBot or PerplexityBot), it's no wonder they don't quote you. Decide consciously who you let in, and don't leave it to chance.

The third is internal linking. Link every important question page to the related service page and back, with clear, descriptive anchor text. This helps crawlers and also presents your topic area as a connected whole.

How can you build external credibility on a small budget?

AI systems don't only look at your own website. If other sources in your field that they consider trustworthy also mention you, your chances improve a lot. This is where small companies often fall short, because they think it requires a PR agency. It doesn't. A few small steps make a difference:

• be present in your field's industry directories and on the websites of associations and chambers, with a consistent company name and description;
• collect genuine customer reviews (Google, industry platforms), because these are recurring sources for AI answers;
• provide expert commentary for industry articles, newsletters and local media where you are mentioned by name;
• take part in your field's forums and communities (LinkedIn, professional groups) with useful, specific answers rather than advertising.

An important professional opinion: bought links and mass, low-value "guest posts" now do more harm than good. Ten genuine, relevant mentions are better than a hundred suspicious ones. AI systems look for consistency and source quality, not quantity.

How do you measure results without expensive tools?

AEO measurement isn't as mature as classic SEO measurement yet, but for a small team it doesn't need to be complicated. A simple monthly routine that takes about an hour is enough.

Put together a list of your 10-15 priority questions, and once a month ask all of them in ChatGPT, Perplexity and Google AI Overviews (where available). Record in a spreadsheet whether you appeared in the answer, whether your website was cited, and which competitors were mentioned. It's worth doing this the same way each time, in incognito mode or without logging in, so that the month-to-month comparison stays clean.

In addition, watch visitors arriving from AI sources in Google Analytics 4 and Search Console (for example referral traffic from chatgpt.com or perplexity.ai), as well as impressions for question-style searches. Don't expect a quick miracle: the first meaningful movement usually shows up within 2-4 months, and AI answers fluctuate, so always look at the trend rather than the result of a single test.

What should you do now?

1. Over one week, collect 20-30 real buyer questions from your salespeople, your emails and Search Console, then choose the 10-15 with the highest business value.
2. Test these questions in ChatGPT and Perplexity and note who appears in the answers before you change anything (this will be your baseline).
3. Rewrite or write the pages for the first three questions so that each starts with a 40-60 word direct answer, followed by an explanation structured with subheadings.
4. Set up FAQPage and Article schema, check your robots.txt, and make sure every page shows the author and the date of the last update.
5. Block out one slot a week in your calendar to take a new question through the full process, and repeat your test questions once a month.

What mistakes do small teams make most often with AEO?

The most common mistake is treating AEO as a one-off cleanup. A team spends three weeks rewriting pages, adds some schema, and then goes back to normal work. Two months later nothing has moved, and the conclusion is that 'AI search doesn't work for us'. In reality the work never reached the point where it could show results. AEO rewards steady, repeated effort on a narrow topic area, not a single burst of activity.

The second mistake is going wide instead of deep. A 12-person machine shop does not need to answer questions about marketing, finance and HR. It needs to be the obvious answer to the 10-15 questions its buyers ask when choosing a supplier: lead times, tolerances, certifications, minimum order sizes, material options. Ten pages that fully own a niche beat a hundred pages that say what everyone else already says.

The third mistake is writing for the machine and forgetting the person. Pages stuffed with FAQ blocks, repeated phrases and keyword-heavy headings read badly, and AI systems are increasingly good at telling padded text from useful text. The practical test is simple: would a buyer who landed on this page be satisfied with the first paragraph? If yes, an answer engine will probably be satisfied too.

Can AI tools speed up AEO work, and where should they stop?

For a one- or two-person team, AI tools are the difference between a realistic plan and an impossible one. Used well, they remove the slow, mechanical parts of the work. Used badly, they produce exactly the generic text that answer engines ignore.

Good uses of AI in your AEO workflow:

• clustering the raw question list you collected from sales emails and Search Console into themes, so you can see which topics deserve their own page;
• drafting the short, direct answer at the top of a page, which you then fact-check and rewrite in your own voice;
• generating schema markup (FAQPage, Article) from your finished page, which you validate before publishing;
• summarising the competitor answers that appear in ChatGPT or Perplexity, so you can see what they cover and what they miss.

Where it should stop: anything that needs real expertise or a real claim. Delivery times, prices, technical specifications, legal conditions and case results must come from you. An invented number on a supplier page does more damage than a missing page, because both customers and AI systems will eventually notice it. Treat AI as a fast junior assistant: useful for first drafts and structure, never the final authority on facts.

How do you keep answers fresh without a content team?

Freshness matters more for AEO than many teams expect. An answer that was correct two years ago and mentions an outdated regulation, price range or tool name is a liability. But a small team cannot re-audit every page every month, so you need a lightweight system rather than good intentions.

Start by giving every question page a review date and an owner. Put the date in a simple spreadsheet next to your priority list. Pages about fast-changing topics (regulations, pricing, software) get a quarterly check. Evergreen pages (how a process works, what a material is) get a check every six months. A check is not a rewrite: you read the first paragraph, verify the numbers, update anything that changed, and change the visible 'last updated' date only when you actually changed something.

Second, let your customers tell you what is stale. When a salesperson hears a question your page does not answer, or a customer says the page contradicts what they were told, that goes straight into the backlog. Your sales and support conversations are the best, cheapest quality-control signal you have.

The table below shows a sensible maintenance rhythm for a team of one or two:

What does a realistic 90-day plan look like for one or two people?

A 90-day plan keeps the work sequenced and prevents the 'everything at once' trap. It assumes about 4-6 hours a week and a site you can edit yourself.

1. Weeks 1-2: collect questions from sales, support and Search Console, score them by business value, and pick your top 10-15. Run each through ChatGPT and Perplexity and record who is cited today.
2. Weeks 3-4: fix the technical basics. Check robots.txt for accidentally blocked AI crawlers, set up Article and Organization schema, confirm mobile speed,