Google is no longer the only search engine that matters. Here is the exact system we use to get local businesses cited by ChatGPT, Gemini, Claude, and Perplexity — in plain English, with the six steps that actually move the needle.
More than 400 million people use ChatGPT every week. Perplexity crossed 20 million monthly users. Google's AI Overviews now sit above the classic 10 blue links for most questions. When your business is the one the AI cites, you win the click, the call, and the customer — often before your competitor is even considered.
Every page needs a real <title>, a real meta description, one H1, and clean semantic HTML. AI crawlers throw pages away that render as an empty JS shell. If your site needs JavaScript to show text, you are invisible to Perplexity, ChatGPT, and most Claude fetches.
Organization sitewide, Article on every guide, FAQPage on any Q&A section, and BreadcrumbList on deep pages. Ship it as JSON-LD in the page head. This is the single biggest lever for getting quoted in AI answers.
Name people, places, brands, tools, and numbers explicitly. Instead of 'our team', write 'Justin Whitaker, founder of Forge Growth'. Instead of 'a lot of clients', write '47 local service businesses across 12 states'. Entities are what AI engines index and cite.
Create /llms.txt at your site root. One H1 with your business name, a one-line summary, and a flat list of your most important pages with short descriptions. ChatGPT, Perplexity, and Claude read this before crawling your JS.
Pull the top 10 questions your customers actually ask. Write one clear, factual answer for each — 3 to 5 sentences, no fluff, no hedging. Group them into a FAQ section with FAQPage schema. These are the exact strings AI engines quote back.
AI engines weigh citations from Wikipedia, industry associations, news outlets, and G2/Trustpilot-style directories very heavily. One citation from a trusted source beats 50 low-quality backlinks for AI visibility.
Pair this with strong local SEO and a fast website and you get compounding visibility across Google, ChatGPT, and every AI engine that matters.
AI search optimization (sometimes called GEO or answer engine optimization) is the practice of getting your business cited inside answers from ChatGPT, Google Gemini, Perplexity, and Claude. It combines classic SEO with entity-rich content, structured data (schema.org), and an llms.txt file that tells AI crawlers what your site is about.
It overlaps but is not the same. Google ranks pages by keywords and links. AI engines answer questions by pulling entities, facts, and citations. A page that ranks #5 on Google can still be the #1 citation inside ChatGPT if it has clean schema, a clear author, and factual, quotable answers.
Publish clear, factual pages that directly answer real questions, mark them up with Article, FAQPage, and Organization schema, add a /llms.txt file at the site root, and earn citations from trusted sources. AI engines strongly prefer content that is well-structured, dated, and attributed to a named author.
Yes. llms.txt is a short markdown file at your domain root that tells AI assistants what your site is about and where to read more. ChatGPT, Perplexity, and Claude increasingly use it to skip the JS shell and read your key pages directly. It takes 15 minutes to write and stays useful for years.
Most sites start showing up in Perplexity and ChatGPT within 4–8 weeks of shipping schema, llms.txt, and 5–10 entity-rich pages. Google's AI Overviews are slower and follow classic ranking cycles, usually 8–16 weeks.
How roofing, HVAC, and plumbing companies get named inside AI answers — and what to publish first.
The same six steps, aimed at job-level questions. A contractor's wins come from pages like 'how much does a roof replacement cost in Austin' or 'signs your AC compressor is failing' — written plainly, marked up with FAQPage schema, and backed by a Google Business Profile with real reviews. AI engines quote the clearest factual answer they can find, and most contractor sites do not have one.
Start with five pages: AC repair cost in your city, furnace repair cost, how often to service a system, emergency AC troubleshooting, and financing options. Add real numbers and dates. Those five cover most of what homeowners ask an AI assistant before they call anyone.
Roofers should publish storm-damage and insurance-claim answers, because that is what people ask AI after weather hits. Include your service radius, the materials you install, and typical timelines. Named specifics get quoted; vague marketing copy does not.
Yes, indirectly and strongly. AI engines lean on sources they trust — Google, Yelp, BBB, industry directories — and those sources rank you partly on review volume and recency. Steady reviews raise your odds of being the business the model names.
Most of it, yes. Schema, llms.txt, internal linking, question mining, and publishing can all run on a schedule. That is exactly what we run for our own site and for clients: articles, links, and AI-visibility checks ship daily without anyone clicking a button.
The services, trade playbooks, and proof behind this system.
Have us run this playbook for you, tracked monthly.
Map Pack rankings built on the same entity and schema work.
The heating and cooling version of this playbook.
The listing data AI assistants read before naming anyone.
How citations and rankings turn into booked jobs.
70.8% revenue lift from a rebuild, schema, and reviews.
Every Forge Growth plan ships schema, llms.txt, entity-rich pages, and monthly AI visibility checks — done for you.