Summary
Eight existing accounts, one pitch each, would normally mean eight discovery cycles. On July 24, 2026, I ran a single live scan across all eight domains and deployed nine client-ready proposals the same day. It cost about 2,100 units of a 400,000-a-month rank-data allowance, 16 AI-assistant queries and six page scrapes.
Eight discovery cycles, or one scan
Discovery is where most agency time goes: a call, a tool export, a slide deck, repeated per prospect. With eight accounts to pitch at once, that was eight weeks of work for a set of documents that would mostly say the same things.
I ran my Audit Engine in prospect mode instead, once, across all eight.
What the scan pulled
For every domain, the same set: domain rating, backlinks, organic keywords, competitors and referring domains from a rank dataset. A live fetch of the robots file, the AI-crawler file, the sitemap and the homepage’s structured data. A crawler service where the direct fetch was blocked.
Two buyer-intent questions per domain put to an AI assistant, with the verbatim answers stored. And first-party Search Console for the four properties I already had access to, so those four could be pitched on their own click data rather than third-party estimates.
What it found
Three of the eight targets ranked for zero keywords. That reframed three pitches from optimization to foundation.
An AI-focused health startup’s site was a crawl trap: every path tested, including a nonsense filename, returned a success code with the identical 467,604-byte homepage, zero internal links and one discoverable URL. Search engines would see one page pretending to be a site.
An AI-conference business blocked GPTBot, ClaudeBot and Google-Extended with a blanket disallow, shutting the AI assistants out of a company whose product is AI events.
A quick-service restaurant chain had zero location pages: 17 URLs in its page sitemap, none of them a store. Its store locator sat on a subdomain that blocked all crawlers and ranked for nothing.
A DTC apparel brand with first-party data turned out to earn 84% of its clicks from its own name. A QR-code SaaS with a domain rating of 78 had its blog hub returning a 404. A pre-FDA medical device company had one page and no ranking keywords.
What shipped
Nine proposals deployed the same day, one per domain plus a capability page, each on a hosted subdomain with its findings linked to the tool call or Search Console row they came from. The budget was disclosed inside the research file: roughly 2,100 rank-data units, 16 assistant queries, six scrapes, one search. Search Console and direct fetches are free.
At a glance
- The niche: Business development for an SEO and AI-search practice; eight existing accounts across retail, SaaS, healthcare, events and food service.
- The problem: Eight discovery cycles for eight pitches, on third-party estimates.
- What was built: A one-day outside-in scan: rank data, live crawler-file and structured-data fetches, stored AI-assistant answers, and first-party Search Console where access existed.
- Implementation: My Audit Engine in prospect mode across all eight domains, July 24, 2026; every number a live tool call with the budget recorded in the file.
- Results: Nine client-ready proposals deployed the same day for about 2,100 rank-data units; three targets found ranking for zero keywords; a crawl trap, a blanket AI-crawler block and a chain with no location pages surfaced.
- Built with: my Audit Engine.
