AI Search Optimization for Small Business: Results From a Verified Citation Experiment

AI search optimization means structuring your website so engines like ChatGPT, Perplexity, Google AI Overviews, and Gemini can read it, trust it, and cite it when they answer the questions your customers ask. In June 2026 we ran that exact play for one small organization, a pregnancy resource center in Massachusetts, and then checked the results by hand instead of trusting a software dashboard. The outcome: the site is now cited by ChatGPT and Perplexity for relevant local queries, and it is not yet cited by Google AI Overviews or Gemini for those same queries.

That split is the most useful thing we learned, and it is the reason this post exists. Most advice about AI search optimization for small businesses comes from two camps: software companies selling visibility dashboards, and skeptics arguing in forum threads that the whole field is snake oil. Both camps are loud. Neither tends to publish receipts. So here are ours: the page formula we used, how we verified every citation, why the engines disagree with each other, and what we would tell any small business owner to do about it this quarter.

Why we ran the experiment

Emprise Digital is a marketing agency in Conway, Arkansas. We have managed Google advertising since 2016: 64+ active Google Ads accounts for 50 billing clients, $3M+ in ad spend managed per year, and 2,000+ real leads, phone calls, and purchases per month across the portfolio, counted strictly as form submissions, calls, and completed purchases. We watch what people type into search engines all day.

Earlier this year we analyzed 1,313,810 search-term rows from 56 of our accounts, and one number from that study kept nagging at us: question-form queries made up 19.5% of impressions for our pregnancy-center clients, more than 500,000 question impressions in 12 weeks. Questions are exactly what AI engines answer. If a fifth of paid-search demand is already phrased as questions, AI assistants are going to intercept a growing slice of it. We wanted to know whether a small local organization could actually get cited inside those answers, so we ran an experiment instead of reading more opinions.

Why AI search is worth a small business's time

The audience stopped being niche a while ago. ChatGPT passed 800 million weekly active users in October 2025, per OpenAI CEO Sam Altman's announcement reported by TechCrunch. Google's AI Overviews reach 2 billion monthly users, per Alphabet's Q2 2025 earnings as reported by TechCrunch. Those two numbers alone mean a meaningful share of your future customers will meet you, or fail to meet you, inside an AI answer.

The traffic is not just big. It is good. Semrush's 2025 AI search traffic study found the average visitor arriving from an AI search source was 4.4 times as valuable as the average traditional organic visitor, measured by conversion rate. That matches what we see in our own data from the paid side: question-form queries on our pregnancy-center accounts earned 17,000+ recorded conversion actions in 12 weeks. People who ask full questions are deep into a decision. Whoever the machine cites gets the warmest click on the internet.

The experiment: one client, four engines, verified by hand

The client and the starting point

Our test client is a pregnancy resource center in Massachusetts. It is a nonprofit, but for search purposes it has the same shape as most small businesses: one location, a small team, a local service area, and people who find it by asking questions, often urgent ones. We manage advertising for 25+ pregnancy resource centers, which together generate 1,000+ real contacts per month, so we know this vertical's search behavior in unusual detail. You can read more about how we serve pregnancy centers.

It is also a category that demands care. The person searching may be facing one of the hardest decisions of her life, and every page has to be accurate, calm, and pressure-free. That standard turned out to be an advantage in this experiment, because clear, factual, low-pressure pages are exactly what AI engines like to cite.

The page formula we used

Nothing here is secret. This is the formula, in the order we apply it:

1. Start from real questions, not keywords. We pulled the questions people in the client's area actually ask, starting with the account's own search terms report. If you run ads, you already own this list. If you do not, your inbox, your reviews, and your phone log hold the same questions in rougher form.
2. One question per page, answered immediately. The first two sentences answer the question plainly. Everything after that supports, qualifies, and adds local detail. An AI engine skimming for a citable answer should never have to dig.
3. Commit to facts. Hours, address, services, what is free, who qualifies. Vague pages do not get cited. Pages that state checkable facts do, because the engine can quote them with confidence.
4. Structure for machines as well as people. Question-form headings, short paragraphs, FAQPage and LocalBusiness schema, clean static HTML that loads fast. The page should make sense with the styling stripped away.
5. Let the AI crawlers in. Your robots.txt must not block GPTBot, PerplexityBot, or Google-Extended, and your core content cannot live behind JavaScript that crawlers will not execute. We see well-meaning sites fail this step constantly in our audits.
6. Keep entity signals consistent. Name, address, and phone identical across the website, the Google Business Profile, and directories. The engines are deciding whether you are a real, specific place. Make that decision easy.

How we verified the citations

Verification is where most AI-visibility claims fall apart, so we kept ours blunt. We ran the relevant local queries directly in each engine: ChatGPT with search, Perplexity, Google with AI Overviews, and Gemini. We recorded whether the client's domain appeared as a cited source in the answer. No estimated visibility scores, no extrapolated share of voice, no tool dashboards. A citation counted only when we could see the client's page listed as a source for a relevant query. That is a higher bar than most reports use, and it is the only bar we trust.

The results

AI engine	Cites the client's site?	What we observed
ChatGPT (with search)	Yes, verified	Cited as a source for relevant local queries
Perplexity	Yes, verified	Cited as a source for relevant local queries
Google AI Overviews	Not yet	Zero citations for the same queries at verification
Gemini	Not yet	Zero citations for the same queries at verification

Two engines cite the site. Two do not. Same pages, same queries, same window. That is the honest finding, and it is more useful than a clean sweep would have been: the AI engines differ sharply, and right now ChatGPT and Perplexity reward citable, structured content much faster than Google's AI surfaces do.

We also want to be precise about what we are not claiming, because our round-down habit applies to findings, not just numbers. We are not claiming a traffic explosion. AI referral traffic is still small next to Google's for every site we watch. We are claiming something narrower and, we think, more valuable: a small local organization with no national profile earned verified citations inside ChatGPT and Perplexity by publishing structured, citable answer pages, while Google's AI surfaces had not yet cited the same pages for the same queries.

Why ChatGPT cites the site and Google AI Overviews does not

The split is not random. It follows how each engine picks its sources.

ChatGPT with search and Perplexity are retrieval-first. When a user asks, the engine searches the live web, reads candidate pages, and cites the ones that most directly answer the question. A page built to be the cleanest answer to a specific local question can win that race quickly, even on a modest domain.

Google AI Overviews and Gemini sit on top of Google's existing ranking systems. In our experience they overwhelmingly cite pages that already rank well organically, which means domain authority, links, and the rest of the slow SEO curve still gate the door. The same page that wins a ChatGPT citation this month may need months of traditional SEO work before Google's AI surfaces will touch it.

For a small business, the practical reading is encouraging. The engines where structure beats authority are exactly the engines where you can compete now. You will not outrank a hospital system or a national chain in Google's organic results this quarter. You can absolutely out-answer them on a specific local question, and ChatGPT and Perplexity notice first.

The playbook: what to do this quarter

Here is the same experiment, translated into steps any small business can run:

1. Build your question list. Search terms report if you advertise, plus the questions customers ask by phone and email. Aim for the literal phrasings, not cleaned-up marketing versions.
2. Publish answer pages with the six-part formula above. One real question per page, answer up front, facts you would stand behind on the phone.
3. Open the gates. Check robots.txt for AI crawler blocks, add FAQPage and LocalBusiness schema, and make sure the content is readable without JavaScript.
4. Verify by hand, monthly. Run your top ten questions in all four engines and log whether you appear as a cited source. Watch your analytics for referrals from chatgpt.com and perplexity.ai. Fifteen minutes a month tells you more than any dashboard subscription.
5. Keep doing traditional SEO. Google AI Overviews follows the organic curve, so the authority you build for rankings is the same authority that eventually earns Google AI citations. This is one body of work with two payoff schedules, not two separate projects.

What AI search optimization costs

You can do everything in this post yourself for free. The formula is the actual method, not a teaser for a paid version.

If you would rather have it handled, our SEO service includes AI search optimization, the work the industry calls AEO and GEO, for $750/mo, or $1,000/mo if we did not build your website. Everything we run is month to month with no contracts, and the full price list is public on our pricing page. We publish prices for the same reason we publish experiments: the businesses we want to work with prefer receipts to promises.

Frequently asked questions

What is AI search optimization?

AI search optimization, also called generative engine optimization (GEO) or answer engine optimization (AEO), is the work of structuring your website so AI engines like ChatGPT, Perplexity, Google AI Overviews, and Gemini cite it when they answer questions. In practice it means publishing pages that answer real questions directly, marking them up with schema, keeping them crawlable by AI bots, and keeping your business facts consistent everywhere they appear.

How do I get my business mentioned by ChatGPT?

Publish pages that answer the specific questions your customers ask, put the direct answer in the first two sentences, commit to concrete facts like hours, prices, and service areas, add FAQPage and LocalBusiness schema, and make sure your robots.txt does not block GPTBot. In our June 2026 experiment, that formula earned verified ChatGPT citations for a small local organization while the same pages were still uncited in Google AI Overviews.

Why is my business not showing up in Google AI Overviews?

Google AI Overviews is built on top of Google's existing ranking systems, so it overwhelmingly cites pages that already rank well in traditional search. If your site is newer or has little authority, AI Overviews will usually be the last AI surface to cite you, no matter how well structured your content is. That matches our experiment exactly: ChatGPT and Perplexity cited our client's optimized pages while Google AI Overviews and Gemini showed zero citations for the same queries.

Does AI search optimization replace SEO?

No. It is mostly the same work read by different machines. The pages that earn AI citations, meaning direct answers, real facts, clean structure, and schema markup, are the same pages that rank in traditional search. The difference is the payoff schedule: ChatGPT and Perplexity can cite a well-structured page quickly, while Google's AI surfaces reward the same page on the slower curve of traditional SEO authority.

How long does it take to get cited by AI search engines?

It depends heavily on the engine, and anyone quoting a fixed timeline is guessing. In our experiment, ChatGPT and Perplexity citations were verified in June 2026 for pages built with the citation formula, while Google AI Overviews and Gemini had cited nothing for the same queries by the same date. Our honest read: retrieval-first engines respond to citable content much faster than Google's AI surfaces, which still appear to follow Google's authority signals.

How much does AI search optimization cost?

You can do the work yourself for free; the formula in this post is the actual method we use, not a teaser. If you want it done for you, our SEO service includes AI search optimization (AEO and GEO) for $750/mo, or $1,000/mo if we did not build your website. Everything is month to month with no contracts, and our full price list is public on our pricing page.