AEO prompts: get your content cited by ChatGPT and Perplexity in 5 steps

Your buyers still search. More of them now ask ChatGPT, Perplexity or Google's AI Overviews and read the answer instead of a page of links.

Those answers cite sources. If your site is one of them, you get the click and the credibility. If it isn't, a competitor does.

Getting cited is called answer engine optimisation (AEO), or generative engine optimisation (GEO). It builds on SEO, with more weight on direct answers, clear formats and other sites mentioning you.

Start the AEO content engine at step 1, or read on to see the whole journey first.

What makes a page get cited

AI assistants pull from pages that answer a question cleanly. That usually means a short direct answer near the top, a format that's easy to lift (a list, a table, a definition, a comparison) and a page other sites already link to.

The five prompts in this series cover each of those, in order. Run them in one chat so every step can read the last one's output. Steps 1 and 3 need web search switched on.

Our worked example: a small B2B SaaS company selling scheduling software to physiotherapy clinics, up against three bigger rivals.

Step 1: Tear down your competitors

Step 1 casts the AI as a senior SEO and answer-engine strategist. You paste in your URL, a line on what you do, and three competitor URLs.

It searches their blogs and returns, per competitor, the topics likely to pull the most traffic, their clusters, the angles that win clicks and the gaps they've left. Every point carries a source URL, and estimates are labelled as estimates.

In our example: all three rivals write about online booking in general. None covers no-show policies for physio clinics, or how to move patient records between systems. Those land in the "Gaps I can own" table.

Step 2: Turn gaps into a 30-post plan

Step 2 takes that gaps table and builds 30 blog ideas: 12 top of funnel, 10 middle, 8 bottom. Each has a working title, a keyword, the intent, the format AI answers tend to quote and the cluster it belongs to.

You add your goal and how many posts a month you can publish. It ends with pillar pages and the first five posts to write.

In our example: one pillar becomes "Reducing no-shows in physio clinics", with posts on deposit policies, reminder timing and a template cancellation policy.

Step 3: Find the questions AI actually answers

Step 3 flips to the buyer's side. The AI plays a customer about to spend money and writes 25 questions the way people type them into an assistant.

For each question you get the sources an assistant would likely cite, the phrase your page should target, the answer format it needs and which of your 30 ideas covers it. The 10 with the clearest buying intent are marked.

In our example: "What's the best booking software for a two-room physio clinic?" needs a comparison table. "Should physios charge a no-show fee?" needs a short direct answer with a pros and cons list.

Step 4: Write the articles, then earn the links

Step 4 writes one article at a time from the step 3 table. You give it the target question, the format, your related pages and any real proof you have: data, customer stories, prices.

It returns a brief, the full article opening with a 40 to 60 word direct answer, an FAQ, Article and FAQPage schema, internal link suggestions and a description of the outside pages that would link to it naturally. Any number you need to supply is marked [ADD SOURCE], so nothing gets invented.

Content alone rarely earns citations. Assistants lean on pages other sites point to, so the second half of step 4 is links.

That's where 3way.social comes in. It's a backlink marketplace where you exchange, buy or sell link insertions and guest posts with other site owners. Its link finder suggests pages that already cover your topic, deals are settled in credits, and links are checked to confirm they stay live.

In our example: the no-show fee article goes live, and the team uses 3way.social to find physio and clinic-management blogs that already discuss cancellations and could link to it.

Step 5: Track citations and fix what's missing

Step 5 is the weekly loop. Ask your 25 questions in ChatGPT and Perplexity, and log whether you're cited and who is cited instead.

Paste that log into the prompt. For every miss it compares your page with the cited sources across structure, depth, schema, internal links, external links and freshness, then gives a rewrite checklist per URL with the fastest fixes first.

It closes with a list of pages that need more links, which goes back into 3way.social before the next check.

In our example: the comparison page loses to a rival whose table sits higher on the page and includes prices. The fix is moving the table up and adding a pricing column, then finding two more sites to link to it.

How the steps feed each other

1. Step 1's gaps table feeds step 2's plan.
2. Step 2's 30 ideas get matched to step 3's 25 questions.
3. Step 3's table picks each article step 4 writes.
4. Step 4's pages are what step 5 checks every week.
5. Step 5's fixes and link list send you back to step 4 and 3way.social.

Start the engine

Pick your three closest competitors and run step 1 now. When the articles are live, use 3way.social to earn the links that help AI assistants trust them, and run step 5 every week.