How AI reads Molly Maid
We scanned Molly Maid and found a 42/100 AEO score. The home cleaning brand has solid agent discoverability at 85/100, but collapses at the interface and authentication layers — both sitting at zero — leaving AI assistants unable to act on behalf of customers trying to book or compare residential cleaning services.
What AI sees
An AI agent landing on Molly Maid's homepage today walks away with brand basics but nothing it can act on.
Molly Maid's homepage gives AI crawlers a coherent picture of a residential cleaning franchise: service tiers, recurring visit options, and a local-franchise lookup model come through clearly in the content structure, which scores 70. The robots.txt is present and parseable. What the agent does not get is any machine-readable interface — no Markdown alternative, no Link response headers pointing to structured resources, no MCP endpoint declaring what the site can actually do. For a service brand where customers increasingly ask AI assistants to compare cleaning companies or request a quote, that gap is commercially meaningful. The site reads like a brochure; it does not behave like a bookable service resource.
Where it loses points
Agent interfaces and identity authentication both score zero — a paired failure that reduces Molly Maid to read-only status for every AI assistant that visits.
How to fix it
Three targeted changes would move Molly Maid from a passively crawlable site to one that AI agents can query, read cleanly, and trust.
Publish an MCP Server Card
Give AI agents a structured declaration of what Molly Maid offers and how to connect to it programmatically.
No MCP Server Card was found at /.well-known/mcp/server-card.json, so agents have no declared entry point for tool-based interaction with the brand.
Serve /.well-known/mcp/server-card.json containing serverInfo, a transport endpoint, and capabilities such as get-quote or find-location. This single static file lifts the agent_interfaces score off zero and signals to AI assistants that the brand supports structured queries.
Return Markdown when agents ask for it
Let AI agents receive a clean, markup-free version of page content without parsing full HTML.
The site returns HTML regardless of the Accept header, so agents must strip navigation, promotional banners, and franchise locator widgets to reach the actual service content.
Add server-side detection for Accept: text/markdown requests and respond with Content-Type: text/markdown, delivering just the core content — service descriptions, coverage areas, pricing structure. Browsers are unaffected and continue receiving HTML.
Add Content-Signal directives to robots.txt
Declare explicitly how Molly Maid's content may be used by AI training pipelines and answer engines.
The robots.txt file carries no Content-Signal directives, leaving AI crawlers to infer permissions — a gap that some answer engines treat conservatively by deprioritizing ambiguous sources.
Append Content-Signal lines covering ai-train, search, and ai-input preferences. Even a straightforwardly permissive declaration improves trust signals with engines that require explicit policy before surfacing brand content in AI-generated answers about home cleaning services.
Common questions
Why does Molly Maid score 85 on agent discovery but zero on agent interfaces?
How does a 42/100 AEO score affect Molly Maid's appearance in AI-generated answers about house cleaning?
What is the fastest single improvement a cleaning-services franchise site can make for AEO?
Is your own site ready for AI?
Run the same five-category analysis on any URL. Free, no account needed to start.
Check your own website free
