How To Get AI To Recommend And Cite Your Business
By David Sessford · 13 Aug 2026 · 12 min read
- AI answers do not have a page two. You are either named in the response or you are invisible to that buyer, so the target is the sentence, not the ranking.
- The biggest lever is third-party trust. AI reads what other sites say about you, so getting into best-of lists, roundups and directories matters more than polishing your own homepage.
- Reddit is the single most cited domain in AI search, per Search Engine Land, so genuinely useful presence in the right subreddits pulls your brand into answers.
- Structure and consistency win the crawl. Answer-first blocks, valid schema and one exact NAP across the web make you easy to trust and easy to quote.
- llms.txt is a five-minute nice-to-have, not a magic switch. As of early 2026 the major AI firms have not committed to reading it and adoption sits near 10 percent.
What answer engine optimisation actually is
Answer engine optimisation (AEO), which some people call generative engine optimisation (GEO), is the work of getting AI assistants to name your business when someone asks them for a recommendation. That is the whole game. When a buyer types "best damp proofing company in Leeds" or "who should I hire to redo my kitchen" into ChatGPT, Gemini, Perplexity or Google's AI answers, one or two businesses get named out loud. You want to be one of them. Everyone else is invisible.
Old-school SEO fought for the blue links. AEO fights for the sentence. The AI reads the web, decides who it trusts, and hands the buyer a short answer with a couple of names in it. There is no page two. There is no scrolling. You are either in the answer or you do not exist to that buyer. That is a scarier game and a much bigger prize.
Here is the part lazy marketers keep missing. You are not optimising for a keyword any more. You are teaching a machine that your business is a real, trusted thing that answers a specific question better than the next guy. Different job. Different playbook. Get this wrong and you can rank on Google all day and still never get spoken aloud by the assistant your buyer actually asked.
Why this matters right now, not next year
The audience is already massive. ChatGPT crossed 800 million weekly active users in late 2025, and Exploding Topics tracked it past 810 million by November 2025. Google's AI Overviews passed 2 billion monthly users by July 2025 and now appear on close to half of US searches. People are not "going to" ask AI who to hire. They are doing it today, on the exact search you want to win.
And here is the kicker. When the AI names a business, the buyer treats it like a tip from a mate, not an ad. No "sponsored" label. No banner blindness. The trust jumps straight from the machine to you. That is worth more than any position one ranking you have ever chased, because it arrives pre-approved.
Meanwhile most of your competitors are asleep. They are still arguing about meta titles while the answer box gets handed to whoever showed up first and looked most trustworthy. Good. That gap is your opening. We drill this exact land-grab inside The Dojo, because the window where this is still easy will not stay open forever.
How AI actually picks who to cite
Stop treating this like magic. These systems are not reading tea leaves. They retrieve a handful of pages, rank them, and quote the ones that are easiest to trust and easiest to lift. Get inside that process and you can win it on purpose. Here is what actually moves the needle.
1. Answer-first content it can lift in one clean block
AI pulls disproportionately from the top of your page. Citation studies from 2025, including work from Profound, found the engines quote heavily from the first slice of body content and mostly ignore the fluff further down. So put the answer first. Turn the question into a heading, then answer it in 40 to 60 words directly underneath, in plain language, with no wind-up. That block is what gets quoted.
Bury your answer under 600 words of throat-clearing about your "journey" and you hand the citation to someone clearer. The machine is lazy in a predictable way. Feed it the exact sentence you want it to repeat.
2. Entities, not just keywords
AI models think in entities: named things they can connect to other named things. Your business needs to be one clear entity with one clear job. "Acme" means nothing to a model. "Acme, a damp proofing specialist serving Leeds and West Yorkshire since 2011" is an entity it can file, trust and recall. Say who you are, what you do, where you do it and who you do it for, in plain words, the same way, everywhere. Repetition is not boring here. It is how a machine learns you exist.
3. Getting mentioned on sites that are not yours
This is the big one, so read it twice. AI does not just read your website and take your word for it. It reads what other sites say about you. Being listed in "best X in [city]" articles, roundups, directories and comparison pages is often the thing that tips the model into naming you. If ten third-party pages call you one of the best plumbers in Bristol, the machine believes it. If only your own homepage says it, the machine shrugs and moves on.
Third-party mentions are the currency of AEO. This is exactly where our link building and digital PR work earns its keep, and it is the single most common reason a good local business gets ignored by AI: nobody else on the web is vouching for them.
4. Reddit and real human discussion
Reddit is the single most cited domain across AI search. Search Engine Land reported studies putting Reddit, YouTube and LinkedIn at the top of the citation pile, and in some analyses Reddit makes up close to half of Perplexity's top-source citations. The engines treat Reddit threads as honest, lived-in experience, not marketing. So if your brand keeps coming up helpfully in the right subreddits and forums, you get pulled into the answers.
You cannot fake this with drive-by spam. Drop a link and run and you get banned and buried. You earn it with genuinely useful replies over weeks, from an account that looks like a human being, in the places your buyers already hang out.
5. Structured data and schema
Schema is the label on the tin. It tells the machine, in a language it never has to guess at, exactly what your page is: a business, a service, a review, a set of questions and answers. Pages with valid structured data get pulled into AI summaries noticeably more often than identical pages without it. FAQPage, HowTo, LocalBusiness and Organization schema are the ones that matter for most small businesses. It is not glamorous. It just works, and most of your rivals cannot be bothered to add it.
6. Consistent NAP and entity signals
NAP means Name, Address, Phone. If your business name, address and phone number are written three different ways across your site, your Google Business Profile, Yelp, Facebook and a dozen directories, you have just told the machine you might be three different businesses. Confusion kills citations. Pick one exact format and enforce it everywhere, down to "Street" versus "St". Boring? Yes. Non-negotiable? Also yes. This is the backbone of local AEO, and we walk through the full clean-up step by step inside the Dojo.
7. llms.txt, with honest expectations
You will hear a lot of noise about llms.txt, a little markdown file you drop in your site root that lists your key content for AI crawlers. Here is the truth so you do not get sold a dream. As of early 2026 the major AI companies have not committed to reading it, and SE Ranking found only around 10% of sites even use it, many of those being empty plugin stubs. It costs nothing to add and it may pay off later. Just do not treat it as a magic switch. It is a nice-to-have, not the strategy.
The exact step-by-step: do this today
Enough theory. Here is the order I would run it in for a small business starting from zero. Do it top to bottom.
- Fix your entity. Write one clear sentence: who you are, what you do, where, and since when. Put it on your homepage, your About page and your Google Business Profile, word for word.
- Lock your NAP. Choose one exact name, address and phone format. Update your site, Google Business Profile, Bing Places, Facebook and your top ten directories so they all match exactly.
- Rebuild your key pages answer-first. On every money page, lead with the buyer's question as a heading and a 40 to 60 word direct answer. Then go deeper below it.
- Add schema. Put Organization or LocalBusiness schema sitewide, plus FAQPage schema on pages with real questions. Check it in Google's Rich Results Test.
- Get on the lists. Find the "best [your service] in [your city]" articles already ranking. Get yourself added, pitched or featured. This is digital PR, and it is where citations are actually won.
- Show up on Reddit and forums. Pick two or three subreddits where your buyers live. Answer questions properly for a few weeks. Mention your business only when it genuinely fits the thread.
- Publish one proper answer asset. A single deep, genuinely useful guide on the thing you are best at, built with clear question headings, beats ten thin blog posts nobody cites.
- Add llms.txt. Five minutes. List your best pages. Move on. Do not overthink it.
- Test yourself. Ask ChatGPT, Gemini and Perplexity the exact question a buyer would ask. See if you show up. Write down who does. Those names are your targets to displace.
Do the first three today. Genuinely, this afternoon. They cost nothing but a couple of hours and they move you further than most agencies will in a month of invoices.
A realistic timeline
Anyone promising overnight AI citations is lying to you. Here is roughly how it actually plays out when you run the steps in order.
| Timeframe | What to expect |
|---|---|
| Week 1 | NAP cleaned, key pages rebuilt answer-first, schema live. Your foundation is done and your entity is finally clear. |
| Weeks 2 to 4 | First third-party mentions land and your Reddit presence builds. Engines start re-crawling and re-reading you. |
| Weeks 6 to 8 | You begin appearing in some AI answers, usually for narrower and local questions first. |
| Months 3 to 6 | Consistent citations on your core questions as the trust signals stack up and compound. |
| Ongoing | Defend and expand. Add mentions, refresh your answers, and track new engines as they launch. |
How the big engines pick their sources
They are not identical twins. Knowing where each one leans tells you where to spend your effort.
| Engine | Leans hardest on | Your priority |
|---|---|---|
| ChatGPT | Wikipedia, established sources, its own web index | Entity clarity plus broad third-party mentions |
| Google AI Overviews and AI Mode | Google's index, Reddit, high-authority pages | Classic SEO plus Reddit plus schema |
| Perplexity | Fresh pages, Reddit, well-structured content | Answer-first blocks, schema and Reddit presence |
| Gemini | Google index and the Knowledge Graph | Google Business Profile and strong entity signals |
| Claude | Trusted, well-structured reference content | Clear, factual, cleanly structured pages |
Notice the pattern. Clean structure, clear entity, outside trust. Nail those three and you show up across all of them, not just one.
Who this suits, and who it does not
AEO is a monster fit for service businesses, local operators, consultants, tradespeople, SaaS tools and anyone whose buyers ask "who is best for X" before they commit. If people research you before they hand over money, you need to be inside that answer.
It is a weaker fit if you sell purely on impulse, compete only on being the cheapest, or trade in something nobody ever asks an assistant about. And it is a bad fit if you are not willing to earn outside trust. If your plan is to tweak your own homepage and call it done, AEO will humble you. The entire point is that the machine listens to everyone except you, so you have to give it something to listen to.
Tools worth using
- The AI engines themselves. Free. Ask them your buyer's question every week and track who gets named. This is your scoreboard, and it is the only one that counts.
- Google's Rich Results Test. Free. Confirms your schema is valid before you rely on it.
- Google Business Profile. Free. Your single most important entity signal if you are local. Fill in every field.
- An AI visibility tracker. Tools like Profound or Semrush's AI features track how often you get mentioned across engines. Worth it once you are past the basics.
You do not need a big budget to start. You need the right order of operations, and that is exactly what we lay out inside the Dojo, so you are not handing an agency $2,000 a month to do what you could run yourself in an afternoon. If you want the wider strategy first, our AEO playbook for 2026 goes broader on the theory.
A quick real-world example
Say you run a small kitchen fitting company in Manchester. A buyer opens ChatGPT and types "who are the best kitchen fitters in Manchester". Right now the AI names two firms. Neither is you. Why? Because those two show up in three local "best kitchen fitters" roundups, their details are consistent across the web, and their sites answer the exact question in a clean block near the top.
Your move is not a mystery. Get featured in those same roundups. Tidy your NAP. Rebuild your services page answer-first. Add LocalBusiness and FAQ schema. Earn a few honest Reddit and forum mentions from happy customers. Six to twelve weeks later you ask the same question and your name is in the answer. That is not a fairy tale. That is the mechanism, run in order, on a business exactly like yours.
Mistakes that keep you invisible
- Writing for robots, not readers. Keyword stuffing still does nothing. AI wants clear, useful, human answers it can quote.
- Only talking about yourself. If the only site calling you great is your own, the machine ignores it. Earn outside mentions or stay invisible.
- Burying the answer. Forcing the AI to dig through 700 words to find your point. It will not dig. It will quote someone clearer.
- Messy details everywhere. Different names and numbers across the web make you look like a ghost, not a business.
- Chasing llms.txt as a shortcut. It is a five-minute add, not a strategy. Do not let it distract you from the real work.
- Spamming Reddit. Drop a link, get banned, torch your reputation. Be useful or stay out.
The 2026 view
Two things are happening at once. AI answers are eating more of every search, and the engines are getting fussier about who they trust. That means the businesses that build real entity signals now, the outside mentions, the clean data, the honest reputation, will compound a lead that is genuinely hard to catch later. The ones sitting back waiting for "proof it works" will end up buying their way in at ten times the cost, if they can get in at all.
The tactics will keep shifting. llms.txt might grow teeth, or die quietly. New trackers will appear. But the core never changes: be a clearly defined, widely vouched-for, easy-to-quote business, and the machines will recommend you. Everything else is detail.
You do not win the AI answer by shouting louder. You win it by being the business the rest of the web already agrees is the answer.
Your move
Pick three steps from the list above and do them this week. Fix your entity. Lock your NAP. Rebuild one page answer-first. Then check whether the AI names you. That alone will put you ahead of most of your market.
If you want the full system, the schema templates, the outreach scripts, the exact Reddit playbook and the weekly tracking we use, that is what the membership is for. Join The Dojo. $67/mo and get the complete AEO build plus everything else we run to win in AI search. If you would rather we ran it for you, look at our done-for-you options and let the team do the heavy lifting.
Frequently asked questions
How do I get ChatGPT to recommend my business?
Make yourself easy to trust and easy to quote. Lock one consistent name, address and phone across the web, rebuild your key pages so the answer sits in a clean block near the top, add Organization and FAQ schema, and get mentioned on third-party sites and roundups. ChatGPT names businesses that the wider web already agrees are the answer, not the ones that only praise themselves.
Is AEO the same as SEO?
No, but they overlap. SEO fights for a ranking in a list of links. AEO fights to be named inside the AI answer itself. Good SEO still helps, because engines like Google AI Mode lean on their own index, but AEO adds entity clarity, third-party mentions, structured data and answer-first writing on top.
How long does it take to show up in AI answers?
Foundations go live in week one. First mentions on narrow and local questions usually start around weeks six to eight, with consistent citations on your core questions building over three to six months as trust signals compound. Anyone promising instant results is selling you something.
Do I need llms.txt to get cited by AI?
No. As of early 2026 the major AI companies have not committed to reading llms.txt, and SE Ranking found only around 10 percent of sites use it, many of them empty stubs. Add it in five minutes because it is harmless and may pay off later, but do not treat it as your strategy.
Why does Reddit matter so much for AI recommendations?
Because the engines treat Reddit threads as honest human experience rather than marketing. Search Engine Land reported studies putting Reddit at the top of the AI citation pile, and it makes up a large share of Perplexity's cited sources. Being mentioned helpfully in the right subreddits pulls you into answers, but only if you earn it, not spam it.
Can a small local business compete with big brands in AI search?
Yes, and often more easily than in classic SEO. Local questions have smaller, less contested pools of sources. A consistent entity, a filled-out Google Business Profile, a handful of local best-of mentions and clean schema can get a small firm named ahead of bigger, lazier competitors.
How do I check if AI is already recommending me?
Ask the engines directly. Type the exact question your buyer would ask into ChatGPT, Gemini and Perplexity, and note who gets named. Do it weekly and track it. That list of names is both your scoreboard and your list of competitors to displace.
David Sessford. Founder of The KillerEdge, 14+ years in SEO & online marketing More about the founders →
Sources
- Exploding Topics: Number of ChatGPT Users
- TechCrunch: Google's AI Overviews have 2B monthly users, AI Mode 100M in US and India
- Search Engine Land: AI search engines cite Reddit, YouTube and LinkedIn most
- Profound: AI Platform Citation Patterns
- SE Ranking: LLMs.txt adoption and why it does not work
- Semrush: The Most-Cited Domains in AI (study)
Grab the free AI-Search Cheat Sheet
The exact moves to make ChatGPT, Gemini and Google AI name-drop YOU. Plus the weekly intel. Free.
Instant download. No spam. Quit anytime.
Want it done for you?
Learn it in The Dojo, or have us run it. Either way, you win.
Join The Dojo. $67/mo → ← Back to the blog