Stage 2 of AI visibility. AI has found your business. Now it must correctly read what you sell, who you serve, and why a buyer should choose you. Most businesses fail here because their content is written for humans — not for AI interpretation.
Being Findable™ means AI knows you exist. Being Understandable™ means AI knows what you do, for whom, and why it matters. These are separate problems. A business can be perfectly findable and completely misunderstood.
Most business websites are written for human persuasion — narrative, emotional, and context-dependent. AI reads differently. It looks for explicit structured signals: service type, target sector, client size, problem solved, and outcome delivered. Implicit information that humans infer from context is invisible to AI.
If AI cannot answer these six questions accurately, you are not Understandable™.
Not a general category — specific service names with descriptions. AI must be able to generate an accurate list of your services without visiting your website.
Sector, company size, geography, and role. AI uses this to match your services to buyer queries. Without explicit buyer persona data, AI guesses — and guesses wrong.
AI matches services to buyer intent by problem category. If your service description does not explicitly state the problem it solves, AI cannot include you in problem-specific recommendation queries.
Buyers increasingly specify delivery format in AI queries. A consulting firm that does not declare remote capability will be excluded from queries for remote consulting services.
Language coverage determines which buyer markets AI assigns the business to. Without explicit language signals, AI defaults to the primary website language only — excluding all other markets.
Buyers use AI to filter by project size and timeline. Without explicit scope signals, AI cannot match your business to queries that specify budget range, project duration, or company size fit.
Numbered actions. Each with the exact file name and location.
Add a structured services array to ai.json. Each service must include: name, description, targetClient, problemSolved, deliveryFormat, and priceRange.
📄 /ai.json → services[]Add JSON-LD Service schema to every service page. Include serviceType, provider, areaServed, audience, and hasOfferCatalog properties.
📄 /services/*.html → JSON-LDMap each service to the exact buyer queries that should trigger a recommendation. Include primary intent, secondary intents, and negative intents (queries you should NOT appear for).
📄 /intents.jsonDefine buyer personas as structured entities. Include sector, company size, geography, role, and pain points. Connect each persona to the services that address them.
📄 /entities.json → buyerPersonas[]Convert existing FAQ content to FAQPage schema with explicit Question and Answer pairs. Each answer must directly address the question — no narrative padding.
📄 /faq.html or service pages → FAQPage JSON-LDExplicitly declare all operating languages, service geographies, and market segments in llms.txt. AI uses this to determine which buyer markets to include the business in.
📄 /llms.txt → languages, marketsCreate a machine-readable service catalogue file that mirrors the service structure in ai.json with additional detail: case examples, typical outcomes, and delivery specifications.
📄 /ai-ready.jsonService page headlines and meta descriptions must be explicit and structured. Replace narrative headlines with declarative ones: "Accounting Services for Romanian SMEs" not "Your Financial Growth Partner".
📄 /services/*.html → <title> <meta description> <h1>