How We Made LGX Appear in ChatGPT Results — and How We Do It for Clients

In 2025, we sat down and did what we now do for every new client: we queried the major AI tools with the exact questions a potential client might ask. Questions like: “What’s the best strategic growth agency for small businesses?” “Who does SEO and AI search optimization for local businesses?” “What agencies specialize in LLM visibility and generative engine optimization?” “Who can help me get found on ChatGPT?”

The results were humbling. Despite being in business since 2010, with a fully optimized site, strong search visibility, and years of content — we were barely showing up in AI-generated answers. One tool named us in a generic list of agencies. Most gave advice without naming anyone. A couple recommended competitors.

We knew why. Our website was built to rank on Google, not to be understood by AI. Our content was keyword-optimized but not structured for machine readability. We had no llms.txt file. Our FAQ content was buried and unformatted. Our schema markup was minimal. We hadn’t given AI tools a clear reason to trust us as the authoritative answer.

So we got to work.

Five changes. Three months. Here’s exactly what we did.

The first thing we did was create and publish an llms.txt file at growwithlgx.com/llms.txt. We structured it to answer the specific questions a potential client might ask an AI tool: what we specialize in, what industries we serve, what results we’ve produced, what our approach is, and where to find our most important content. The file has to be specific and declarative — not marketing language, but plain-language descriptions that an AI model can parse with confidence.

AI tools cite FAQ content more than almost any other content type. Our existing FAQ section was thin and generic. We rebuilt it from scratch — identifying the actual questions our target clients ask when researching agencies, writing clear and specific answers, and formatting everything with FAQPage schema markup so AI tools could identify and cite it accurately. We added questions across several categories: how retainers work, what SEO actually does, how AI search optimization differs from traditional SEO, what industries we serve, and how our pricing compares.

Schema markup is code that tells AI tools and search engines exactly what type of content they’re looking at. We expanded ours to include Organization and LocalBusiness schema with complete location, contact, and service information; Service schema on each of our six service category pages; FAQPage schema on all FAQ content; Article schema on blog posts; and PriceSpecification schema on our packages page. This gives AI tools a machine-readable map of who we are, what we do, and why we’re a credible source — without requiring them to interpret paragraphs of marketing copy.

We went through our key service pages and homepage and restructured the content to be more semantically clear — plain language that answers specific questions, organized hierarchically, with natural-language summaries at the top of each section. We also built entity references into our content: consistent, factual mentions of our business name, service categories, founding year, team structure, and areas of expertise across multiple pages. AI models build understanding through consistency — the more specifically you describe yourself across your site, the more confidently an AI can cite you.

Before we started, we documented our baseline: what came up when we queried major AI tools with relevant questions. We screenshotted the results, noted which competitors appeared, and documented the specific gaps in how we were represented. We ran the same audit 60 days after implementation. The difference was substantial. LGX now appears in AI-generated answers for relevant queries about growth agencies and marketing services — with accurate, specific descriptions that match what we actually do.

When we run this process for a client, the sequence follows the same logic — adapted to their industry, their service area, and the specific questions their customers are asking AI tools.

We sit down with real AI tools and query them the way a buyer in your market would. What comes up? Who gets recommended? Are you in the answer at all? If you are, how accurate is the representation? This gives us a baseline — and it’s usually enough to make the value of the work immediately clear.

For every new retainer client, it’s one of the first deliverables. Published to the site root, structured around your specific services and geography, and updated as your offerings or service areas change.

For local track clients — home services, medical, legal, contractors — we prioritize geo-specific FAQ content: “Who does [service] in [city]?” “What should I expect to pay for [service]?” For authority track clients — professional services, tech, SaaS — we prioritize expertise-driven content that answers the questions someone would ask an AI before deciding which provider to trust with a complex problem.

Every new page we build for a client gets appropriate schema markup from day one. For existing sites, we audit what’s in place and expand it — especially FAQPage and Service schema, which have the most direct impact on AI citation.

We document what changed — not just what we built, but whether you now appear in AI-generated answers, how you appear, and what questions are still unanswered by your content.

Short answer: if your business has a website and serves customers, you probably do.

The longer answer: ask yourself whether your potential customers are the type of people who might ask an AI tool a question before making a buying decision. If you serve homeowners, business owners, professionals, or anyone already using ChatGPT or Google’s AI Overviews — they are asking AI tools about businesses like yours. Whether your business is in that answer is a different question.

You can test this yourself right now. Open ChatGPT or Perplexity and type: “Who are the best [your service] businesses in [your city]?” or “What [your business type] would you recommend near [your market]?” Document what comes up. Note whether your name appears — and if it does, whether the description is accurate.

If you’re not in the answer — or if the answer is wrong — that’s what we fix. And the free audit we run is exactly this process, done systematically, across multiple AI tools, with the specific questions your customers are actually asking.

We took three months to get the LGX result we were satisfied with. That’s not the standard client timeline — that’s what happens when you’re building the methodology from scratch while running a business at the same time.

For clients, the core implementation work — llms.txt, FAQ architecture, schema markup, on-page restructuring — typically takes two to four weeks depending on the size of the site. AI tools vary in how frequently they re-index content, but most clients begin to see improved AI representation within 60 to 90 days of completing the core work.

For retainer clients, it compounds. Every new page we build comes with appropriate schema markup. Every FAQ we add is another potential citation point. Every content update is another opportunity to give AI tools a clearer, more specific picture of your business. The clients who will have the strongest AI presence in 12 months are the ones who start now and keep building.

We document the before-and-after with every engagement. You’ll know exactly what changed — not just what we built, but whether you’re now showing up in the answers your customers are getting. That’s the only metric that matters.