Answer Engine Optimisation for Small Business: Get Cited

Open ChatGPT right now and test answer engine optimisation for small business: type “best [your trade] in [your city]” and hit enter. You’ll get three or four business names, a sentence about each, and a couple of links. If your business isn’t in that list, the person asking never sees your website, your reviews, or the ad you’re paying for.

That is the entire case for this approach. Buyers have stopped scrolling through ten blue links and started asking a machine for a shortlist. The machine gives them one answer, not a page of options, and you are either in it, or you’re invisible.

This post covers what answer engines actually do when they build that shortlist, why they pick some businesses and skip others, a practical checklist you can work through this month, and how to tell whether any of it is working.

Key Takeaways

• Answer engine optimisation for small business is about being named in the AI answer, not ranking tenth on a page nobody scrolls to.
• Answer engines cite what they can verify. Consistent business details across the wider web matter more than any single page you own.
• Write in answer blocks: a real question as the heading, a complete 40 to 80 word answer underneath, before any long build-up.
• Specifics survive summarisation. Ranges, timeframes, suburb names and real numbers get quoted. Adjectives get thrown away.
• Check your robots.txt. Blocking GPTBot, ClaudeBot or PerplexityBot removes you from the answer entirely, no matter how good the page is.
• Run ten buyer questions through ChatGPT, Claude and Perplexity every month and log who gets named. That log is your real ranking report.
• AI referrals arrive pre-qualified and impatient. If nobody answers within minutes, the visibility work pays your competitor instead.
• Schema markup, a clear About page and a proper FAQ block do more for citations than another thousand words of general advice.

Answer Engine Optimisation for Small Business Is Not SEO With a New Name

An answer engine is anything that reads the web and hands back a written answer instead of a list of links. ChatGPT, Claude, Perplexity, Google’s AI Overviews, Gemini, Copilot. Some of them work from what they learned in training. Most of them now run a live search in the background, read the top handful of pages, and write a summary with citations.

The difference from search engine optimisation is not cosmetic. SEO competes for position. Answer engine optimisation competes for inclusion. There is no page two. A traditional search results page gives ten businesses a shot at the click. An answer gives three names, sometimes one, and everyone else may as well not exist.

The click economics have already shifted underneath this. Pew Research Centre found that when an AI summary appears in Google results, users click a link on about 8% of visits, compared with 15% when there’s no summary. Roughly half the clicks are gone. The traffic that remains is the traffic that goes to whoever got named in the summary.

SEO gets you ranked. Answer engine optimisation gets you quoted.

Here’s the practical distinction. A page can rank well and never get cited, because it’s written to satisfy a crawler rather than to be lifted out and repeated. A page can also get cited without ranking especially well, because it happens to contain the one clean, checkable sentence the model needed.

Ask yourself what a model would actually extract from your services page. If the answer is “a paragraph about being passionate about quality and committed to customer service,” there’s nothing to quote. That sentence describes every competitor you have. It carries no information, so it gets discarded.

Now ask what it would extract from a page that says “We service hot water cylinders across the North Shore, same-day callouts before 2 pm, and we’ve been doing it since 2009.” Three facts, all specific, all verifiable. That’s quotable.

You’ll also see this called generative engine optimisation, or GEO. Same discipline, different label. I’ll stick with answer engine optimisation because it describes the thing more honestly: you’re optimising to be the answer.

Why Answer Engines Name Some Businesses and Ignore Others

Most owners assume the model has a preference or a hidden ranking. It doesn’t. It has a retrieval problem and a confidence problem, and understanding both tells you exactly what to fix.

It has to know what you are

Before a model can recommend you, it needs to resolve you as an entity. That means it needs to be confident that “Mitchell Plumbing” is a plumbing business, in a specific place, serving specific customers, and that the Mitchell Plumbing mentioned in a directory is the same Mitchell Plumbing with the website.

Entity confusion is the single most common reason small businesses get skipped. Your Google Business Profile says one trading name. Your website footer says another. Your invoices use the legal entity. A directory listing from 2019 has your old address. A human resolves all of that in two seconds. A model treats it as four possibly different businesses, none of which it trusts enough to name.

It has to verify you somewhere other than your own site

Your website is a claim. Answer engines weight claims made about you elsewhere far more heavily than claims you make about yourself. Reviews, directory listings, industry association pages, local news coverage, supplier pages, “best X in Y” roundups, forum threads, LinkedIn.

This is the part that frustrates people. You can write the best page on the internet about commercial fitouts in Christchurch, and if nothing else on the web corroborates that you do commercial fitouts in Christchurch, the model will cite a directory listing for a competitor instead. Third-party corroboration is the currency.

It has to be able to lift a clean answer out of your page

Models summarise. Summarising means throwing away everything that isn’t load-bearing. Long build-ups, brand storytelling, three paragraphs before you answer the question in the heading, all discarded. What survives is the sentence that directly answers something a person asked.

If your page takes 400 words to get to the point, the model will find a page that took 40 and cite that one instead.

It has to be allowed in

This one catches people constantly. A lot of sites, particularly ones built by agencies during the scraping panic of the last few years, block AI crawlers in robots.txt. GPTBot, ClaudeBot, PerplexityBot, Google-Extended. If those are disallowed, your site is not in the retrieval pool. You can do every other thing on this list perfectly and still never appear.

Go and look at yoursite.co.nz/robots.txt right now. It takes fifteen seconds. I’ve found this blocking on more client sites than any other single technical issue.

The Answer Engine Optimisation Checklist for Small Business

This is the practical part. Work through it in order. The first three items take a couple of hours and fix most of the damage.

1. Make your business details identical everywhere

Pick one trading name, one address format, one phone number format, one one-sentence description of what you do. Then push that exact set across your website, Google Business Profile, Facebook, LinkedIn, every directory you’re listed in, and any industry body page. No variations. No “Ltd” on some and not others.

Write the one-sentence description as though a stranger will repeat it word for word, because that’s the job. “Mitchell Plumbing is a residential plumbing and gasfitting business serving Auckland’s North Shore, specialising in hot water systems and bathroom renovations.” Boring, specific, repeatable. Perfect.

2. Let the crawlers in and add schema markup

Fix robots.txt first. Then add structured data: Organization or LocalBusiness markup on your homepage with your name, address, phone, hours and service area, and FAQPage markup on any page with a question-and-answer section.

Schema doesn’t magically get you cited. What it does is remove ambiguity. It tells the machine, in a format built for machines, exactly what you are and where you operate, so it stops guessing. For a small business site, that’s usually two blocks of JSON-LD and half an hour of work.

3. Rewrite your key pages as answer blocks

Take the ten questions your customers actually ask before they buy. Not keyword phrases, actual questions, the ones that come up on the phone every week. “How much does it cost to replace a hot water cylinder?” “Do you charge for quotes?” “How long does a bathroom reno take?”

Each one becomes a heading. Directly underneath, answer it completely in 40 to 80 words. No preamble, no “great question,” no building to a reveal. Answer, then expand if you want to.

That structure exists for one reason: it hands the model a self-contained, extractable block. It also happens to be better for humans, who were skimming for the answer anyway.

4. Publish the specifics nobody else will

This is where most small businesses win, and it’s also where most refuse to play. Answer engines reward information that isn’t available elsewhere. Price ranges. Real timeframes. Which suburbs you cover and which you don’t. What you won’t take on. How long a job actually takes when it goes wrong.

Your competitors have all written, “affordable, reliable, professional.” None of that gets quoted, because none of it is information. A page that says “cylinder replacements typically run four to six hours and we quote in a range because access is the variable” gets quoted, because it tells the reader something they couldn’t have guessed.

If you’re worried about publishing ranges, publish the range and the reason it’s a range. That’s more honest than the competitor who won’t say anything, and the model can see the difference.

5. Get corroborated off-site

Claim and complete every listing that matters in your industry. Ask for reviews on the platforms models actually read, not just the one your CRM defaults to. Get onto local business association pages, supplier partner lists, industry directories. If a local publication does a roundup in your category, be reachable and be quotable.

One genuinely useful move: write the “best X in Y” content yourself, honestly, including competitors. It sounds counterintuitive. It works because those roundup pages are exactly what answer engines retrieve for shortlist questions, and being the source of the list is better than being absent from someone else’s.

6. Keep it current

Model weight freshness for anything with a time-sensitive element. Pricing, regulations, availability, hours. A page last updated in 2022 that mentions pricing is worse than no page, because it will be cited with numbers that are now wrong and you’ll spend the call correcting it.

Pick a quarterly cadence. Update the dates, update the numbers, kill the pages that no longer reflect what you do. This is dull, recurring, low-status work, which is precisely why it never gets done in a founder-led business. Hold that thought.

How to Tell If Answer Engines Are Naming You Yet

There’s no rank tracker for this that I’d trust yet. The tools are early, and the answers vary by user, by session, by phrasing. So do it manually. It takes twenty minutes a month.

Build a ten-question prompt set

Write the ten questions a buyer would ask an AI instead of Google. Mix them up:

Shortlist questions (“who are the best commercial electricians in Hamilton”), problem questions (“my hot water is running out after five minutes, who do I call in Wellington”), comparison questions (“what should I ask before hiring a bookkeeper”), and category questions (“what does a mortgage broker in Auckland actually charge”).

Run all ten through ChatGPT, Claude, Perplexity and Google AI Overviews. Same wording each month. Log which businesses get named and which sources get cited.

That log is your report. Not your keyword rankings. Which names come out of the machine, and whose pages the machine read to produce them.

Watch for AI referral traffic in your analytics

In GA4, filter your referral sources for chatgpt.com, perplexity.ai, claude.ai, copilot.microsoft.com and gemini.google.com. The volumes will look small next to organic search. Look at the behaviour rather than the volume: session duration, pages per session, and conversion rate.

In almost every account I’ve looked at, AI referral traffic converts at a materially higher rate than general organic. That makes sense. The model already did the filtering, the comparison and the recommendation. By the time someone clicks through, they’ve been sold to by a third party they trust.

Track branded search as the lagging indicator

The other signal worth watching is people searching your business name directly. When an answer engine names you, a good chunk of those people don’t click the citation. They go and search your name separately, because that’s the habit. A rise in branded search with no matching ad spend usually means you’re being named somewhere you can’t see.

The Part Nobody Warns You About: What Happens After You Get Named

Say all of this works. You do the entity cleanup, you rewrite the pages, you get the listings sorted, and three months later ChatGPT starts naming you when someone asks who to call in your area.

Now the enquiry lands. And here’s the thing about an AI-referred lead: they’ve already compared you against two competitors, already read the summary, already decided you’re worth contacting. They are further down the line than any lead you’ve ever got from a Google ad. They are also, for exactly the same reason, in a hurry. They asked a machine and got an answer in four seconds. Their tolerance for waiting on you is close to zero.

Research on lead response has held steady for years: the first business to respond wins the deal roughly 78% of the time. Most small businesses take four hours or more. Some take days. Which means the visibility work you just did gets converted by whoever picks up the phone, and that isn’t always you.

I saw this play out with a dental practice I worked with. Dr Claire had two receptionists and was still missing close to half her inbound calls, because calls arrive in clusters and people are mid-appointment. After putting an AI receptionist on the line to catch everything the humans couldn’t, booked appointments went up 44%. Nothing changed about her marketing. The demand was already there. It was falling through the gap between arriving and being answered.

Answer engine optimisation makes that gap more expensive, not less. More visibility into a business that can’t respond is just a bigger leak. If your after-hours enquiries currently go to voicemail, fix that before you go chasing citations. I’ve written about what after-hours call answering actually costs a small business if that’s the closer gap for you.

And it’s another recurring job on your desk

Look back at that checklist. Rewrite pages as answer blocks. Keep listings consistent. Update pricing quarterly. Run ten prompts a month and log the results. Watch referral traffic. Publish specifics your competitors won’t.

None of it is hard. All of it is recurring. And in a business where the owner is the system, recurring work with no deadline attached simply doesn’t happen. It sits on the list under the things that shout louder, and in six months you’re right back where you started, wondering why the machine still names the other guy.

That’s the actual constraint, and it isn’t a marketing constraint. It’s a capacity one. The businesses that will stay visible in AI search over the next two years aren’t the ones with the best content strategy. They’re the ones where content monitoring, listing hygiene and lead response run without the founder remembering to do them.

That’s what an AI brain plus an AI workforce is for. The brain holds the context: what you sell, who you serve, how you describe yourself, what your positioning actually is, so anything written on your behalf is consistent instead of drifting. The workforce does the recurring jobs: checking the prompt set, flagging when a competitor starts outranking you in the answers, drafting the quarterly page updates, and answering the enquiry within ninety seconds when it lands. If you’re weighing that up against hiring someone to own it, I’ve laid out the comparison in AI automation for business and the practical starting points in AI adoption for SMEs.

Conclusion

Answer engine optimisation for small business comes down to three things. Be resolvable, so the machine knows exactly what you are and where. Be corroborated, so something other than your own website confirms it. Be quotable, so there’s a clean, specific sentence worth lifting into the answer.

Everything else is detail. The technical work is a weekend. The hard part is that it never stops being work, and in most founder-led businesses that means it never gets done at all.

The businesses winning here aren’t the ones with the biggest content budget. They’re the ones where this runs as a system instead of as something the owner keeps meaning to get to. Get the visibility, then make sure there’s something on the other end capable of catching what it sends you. A shortlist mention that goes to voicemail is worth exactly nothing.

Ready to find out where you actually stand?

Run your ten prompts first. Twenty minutes, no tools required, and you’ll know immediately whether this is a problem you have.

If the answers come back naming everyone but you, or if they name you and you’re not confident you’d catch the enquiry, book a 30-minute Discovery Call. We’ll go through what the machines are currently saying about your business, where the gaps are, and what’s worth fixing first. No pitch deck, no obligation, just a straight read on where you sit.

Frequently Asked Questions

About Octavius