Sample · Vertical
Campgrounds — 4 in-county, 55 in 30-mi radius
The campground audit is illustrative, not part of the county rollup. It demonstrates how a single category of business looks under the same Tier 1 / Tier 2 lens applied countywide.
Tier 0 · Tier 1 · 271 Businesses
When a customer asks ChatGPT, Perplexity, or Claude about a local business in Hancock County, the answer depends on what AI agents can actually find and read. We audited every business of record in the county against two questions: can an agent act on it, and do answer engines cite it. The averages are low. The opportunity is wide.
† "Website" = any URL in the business's Google Business Profile, including company landing pages (e.g. Hostfully, Family Dollar) and Facebook pages. Owned domains alone would be lower.
Tier 0 · Agent Readiness
A technical audit of the signals AI agents and crawlers use to discover, read, and trust a website — robots.txt, sitemap, llms.txt, structured data, MCP server cards, OAuth discovery. Each check returns pass, fail, or unable-to-check.
Tier 1 · Answer Engine Visibility
We submit customer-style questions to seven major answer engines — Perplexity, Claude, ChatGPT, Google AI Mode, Google AI Overviews, Gemini, Grok — and count how often each business is cited as a recommended source.
§ 01 · Tier 0
The Agent Readiness Score measures how easy it is for AI tools to find, read, and use a business's website. Higher scores mean AI engines are more likely to pick up the site when building answers. Businesses without a website are capped at half the possible Tier 1 score because there is nothing to crawl.
robots.txt, sitemap, link response headers — basics every site should already do.
Markdown content negotiation: serves a clean markdown version of pages when agents ask for one. Universal failure across the county.
AI-specific rules in robots.txt + Content Signals. Web Bot Auth checked but neutral on most sites.
API Catalog, OAuth/OIDC discovery, OAuth Protected Resources, MCP Server Card, Agent Skills Index, WebMCP. Where the deepest gaps live.
x402 / MPP / UDP / ACP / NFD surfaced for context but not scored — they apply to commerce sites narrowly.
Roughly three-quarters of the county has no website. They appear on Google Business Profile but are effectively invisible to web-based AI agents from the start.
38% of identified businesses are not listed on the Chamber website at all — and 34% of those that are listed have incomplete name, address, or phone data.
§ 02 · Root causes
Two patterns drive the low readiness scores. First, the majority of Hancock County businesses have no website at all. Second, even sites that exist are missing the structured signals AI agents look for — eight technical checks fail across the board.
§ 03 · Tier 1
Readiness is a precondition; visibility is whether engines actually cite you. We submit real customer questions to seven answer engines and count citations. The long tail is silent: most businesses are not surfaced by any engine for any prompt.
§ 04 · Methodology sample
To make the methodology inspectable, we ran both audits end-to-end on a single vertical before scaling. The sample reports below show every score, every check, and every answer-engine citation for one business — so the rollup numbers above can be traced back to source data.
The campground audit is illustrative, not part of the county rollup. It demonstrates how a single category of business looks under the same Tier 1 / Tier 2 lens applied countywide.
An individual business view shows the Tier 1 readiness score, every passing and failing check, and the Tier 2 answer-engine breakdown — including which engines cite the business and which queries it appears for.
§ 05 · Method
Two automated audits, run against every business of record in Hancock County, Tennessee.
We run the Agent Readiness Audit scanner against each domain, checking for the technical signals AI agents and crawlers use to discover and use a website — robots.txt, sitemap, llms.txt, structured data, MCP server cards, A2A agent cards, OAuth/OIDC discovery, and commerce protocols. Each check returns pass, fail, or unable-to-check.
A fixed set of customer-style questions is submitted to seven major answer engines — Perplexity, Claude, ChatGPT, Google AI Mode, Google AI Overviews, Gemini, and Grok. Every website each engine cites as a recommended source is recorded. Citation count is the total across all engines and prompts.