How to Get Small Businesses in ChatGPT Results: The 3-Step Playbook

Your own website is not your strongest asset for ChatGPT and other LLMs’ visibility; third-party profiles are. For ChatGPT specifically, it leans heavily on directory data when answering “best [service] in [city]” queries.

BrightLocal’s 2025 study of AI search sources found that Yelp alone is cited as a source in 33% of AI searches across ChatGPT, Gemini, and Perplexity. (Source) The goal is to be present, consistent, and accurate on every directory ChatGPT pulls from.

The directory priority order

The priority order for directories to help show up in ChatGPT results is: Google Business Profile, Yelp, Bing Places, Apple Maps, and finally MapQuest. Your Google Business Profile is first because it feeds into Google’s broader entity graph, and is a very heavy local SEO lever. Yelp is one of the most heavily cited sources for Chat for local recommendation queries, especially in service categories. Bing Places is great because ChatGPT runs on Bing’s index, and Apple Maps and MapQuest are high authority profiles from reputable companies, making them very valuable as well. Claim your business on each of these directories, fill out every field that it provides you, add unique, relevant photos, and list every service as a separate entry when it is allowed.

NAP consistency

NAP consistency, a standard SEO practice, is a rule where the Name, Address, and Phone number must be identical, character-for-character, across every profile. When an AI sees “123 Main Street” on one profile and “123 Main St.” on another, it can read them as two different businesses and lose confidence in citing either. The additional directories mean more places to check, but a general rule of thumb is to pick one version to stick with and audit every listing against it.

Reviews are the single biggest trust signals for AI tools. ChatGPT and other LLMs weigh recency heavily, so a steady trickle of reviews will beat a big one-time influx. Additionally, a sudden pile of reviews looks suspicious. Setting up a system that captures reviews continuously is integral to succeeding in GEO.

The in-person ask

The highest converting tactic is asking for the review in person, while the work is fresh in a customer’s mind and they are happy. The face-to-face ask converts far higher than an email followup, even if the email is sent quickly after the job is done. Two tools we commonly suggest that make the ask easy are NFC review cards (tap-to-leave-review) and business cards with a QR code pointing directly to your GBP review link. This just ensures that the process for the client has no hiccups or delays.

What a great review actually looks like

While Google prohibits telling anyone what to write on a review, ensuring that your reviews include the services you provide is integral to your success in showing up in ChatGPT. Generic short reviews provide little to no value to an LLM’s understanding of your company. A great suggestion to the client to ensure good, helpful, AI-quotable reviews is to have them explain what the problem was before you came and then how you solved that problem.

Diversify across platforms from day one

Diversifying your reviews to multiple directories should be a high-priority task. Don’t wait until you are top 3 in the Map Pack to start collecting on other platforms. ChatGPT looks at Yelp reviews commonly for local recommendations, so when asking customers for reviews, alternate which platform you point them to so reviews build proportionally.

A singular “Services” page that becomes bloated or ends up with vague bullet points will be hard for both Google and AI to find anything worth quoting. LLMs are looking for clean, self-contained chunks of information that they can lift into their own answer. For example, “Professional window cleaning service” tells the AI nothing. A dedicated page on exterior window cleaning with specifics gives it much more to work with.

SE Ranking’s analysis of 129,000 domains found this pattern in the data: articles over 2,900 words average 5.1 ChatGPT citations versus 3.2 for thin content under 800 words, and pages with sections of 120 to 180 words between headings perform best. (Source)

Every service gets its own page

A good rule to follow is that every service and separable sub-service needs its own dedicated page on the website. A window cleaner, for example, shouldn’t have one services page. They should have separate pages for exterior window cleaning, interior window cleaning, screen cleaning, sill cleaning, and so on. Pages that match more specific intents help Google, LLMs, and humans understand what your business actually provides.

What goes on each service page

Every service page should include sections about how your small business does it differently, your actual step-by-step process, mentions of specific names of the tools, brands, or techniques you use, and a FAQ section. The “how you’re different” section gives the AI a unique angle to cite when comparing you to competitors. The step-by-step matters because it creates a clean, extractable chunk that an AI can pull into its direct answers. Being specific is just more proof that you are a professional and know what you are talking about. A FAQ section is a format that LLMs prefer to extract from. Use real questions a curious customer would ask, not keyword-stuffed filler no one cares about. Your FAQ is another opportunity to prove that you are a professional.

A good baseline template for service pages looks like this: an H1 that names the service and the location, an H2 explaining why a customer should choose you, an H2 walking through your step-by-step process, an H2 for reviews or social proof, an H2 covering pricing or memberships if relevant, and an FAQ H2 at the bottom. Most pages on the site can run on this skeleton. Where it makes sense, add a section unique to that service. Service pages are the most important pages to get right because they are what AI tools quote when answering “best [service] in [city]” queries.

A worked example: Dirt Herders’ exterior window cleaning page

The exterior window cleaning page on dirtherders.com is a working example of this template. The H1 is “Professional Exterior Window Cleaning Services In Santa Monica,” which names the service and the city. From there, the page breaks into the following H2s:

• Why Choose Dirt Herders for Exterior Window Cleaning in Santa Monica
• Our 4 Step Process
• See What Your Neighbors Are Saying
• Save Big With Our Dirt Herders Memberships
• FAQ
• Transform Your View With The Dirt Herders Exterior Window Cleaning Service

Each H2 does a specific job. The “Why Choose” section is where the differentiator goes. For Dirt Herders, that differentiator is the four-step process: scrubbing, squeegeeing, detailing, and applying their RainGuard Technology. Naming the process and the product gives ChatGPT something concrete to quote when a customer asks why one window cleaner is better than another. Generic phrasing like “we use the best products” gets ignored. “RainGuard Technology applied as the fourth step of our process” gets quoted.

The “Our 4 Step Process” section is the step-by-step. This is the most extractable section on the entire page because it is structured exactly the way an LLM wants to read it: a numbered list of named, specific actions. When ChatGPT pulls a recommendation for Dirt Herders, this is the section it lifts from.

The FAQ section at the bottom is where most of the long-tail queries get answered. Real questions a curious customer would ask, written out in plain language. The exterior window cleaning page uses these:

• What makes The Dirt Herders exterior window cleaning different?
• Can The Dirt Herders handle high windows?
• How often should I schedule window cleaning for optimal results?
• What if it rains after my window cleaning?

Each answer is short, direct, and references a specific feature of the service. None of them are filler. Every answer reinforces a differentiator that the AI can quote.