◈ SaltCrate · AI Visibility

When a customer asks AI for a plumber, it names four companies.
Yours is not one of them.

We ask the assistants your customers actually use — then show you, in writing, why they recommend your competitors and not you. $500, three business days, no retainer.

Get the audit — $500 Card payment · Delivered as a PDF plus a 30-day re-test

What we found in Austin, 20 November 2026

We asked two assistants the question a homeowner in trouble types at 2am: “Water is pouring through my ceiling, who should I call?”

Gemini — asked for an emergency plumber in Austin, TX“Highly rated local options you can call right now include Radiant Plumbing & Air, ABC Home & Commercial Services, Daniel's Plumbing & Air Conditioning, and Clarke Centex Plumbing.”
DeepSeek — same question“I can't verify which Austin plumbers are available for same-day emergency calls… NONE. Search ‘emergency plumber Austin TX’ on Google or Yelp.”

Then we audited eight real Austin plumbing companies — family firms trading for 20, 40, even 50 years, with hundreds of five-star reviews. None of them appeared in any of the 32 AI answers we collected. The assistants either default to the four biggest brands or refuse to name anyone at all.

0/32

AI answers that named any of the eight businesses we tested

0/8

businesses with machine-readable review and rating markup

2/8

businesses with FAQ schema an assistant can quote

That is the whole problem in one paragraph: the businesses with the best local reputation are the ones the answer engines cannot see. Reputation is not the input; structure is.

What the $500 audit measures

CheckWhy it decides whether AI can name you
AI answers, verbatimFour buying questions across two assistants, with the full answer text and whether you appear at all
AI crawler accessPer-bot robots.txt verdicts for GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot, Google-Extended, CCBot
Entity markupIs there LocalBusiness / Plumber / HVACBusiness JSON-LD that states what you are, where you are and what you serve
Review & rating markupWhether your ratings are machine-readable or only visible in an image
FAQ structureWhether there is citable Q&A content, and whether it carries FAQPage schema
llms.txtA plain-language map of your site for language models
Fact consistencyName, address, phone and service area stated the same way on your site and across the sources assistants retrieve
Answer-ready contentWhether any page can be quoted as an answer to a real question
TriageEverything ranked by what actually moves an answer, not by what is easiest

What you get

1 · The evidence

12–16 page PDF: every question we asked, every answer we got, your 0–100 visibility score, and the competitors the assistants named instead of you.

2 · The diagnosis

The nine checks above, run on your live site, each marked pass or fail with the exact evidence — header, schema block, or missing file.

3 · The 90-day plan

A prioritized fix list, ordered by impact: what to change this week, what to publish this month, what to leave alone.

4 · The re-test

Thirty days later we re-run the same questions and send you the comparison, whether or not you hire us for anything else.

Delivered in three business days from the day you send your URL, city and priority service. You get the raw assistant outputs, not a summary score from a black box.

Who this is for — and who it is not

Good fit

  • Local service businesses (plumbing, HVAC, roofing, electrical, dental, med spa) with a real reputation and a website
  • Owners who keep hearing “we found you on Google” and want to know what happens when the question moves to an assistant
  • Marketing managers who need evidence before signing a $3,000/month agency retainer

Not a fit

  • Anyone expecting a guarantee of being recommended — nobody can offer that honestly
  • Businesses with no website, or a site their own team cannot edit
  • National brands looking for a link-building retainer — this is a diagnosis and a plan, not a campaign

Questions people actually ask

Can you guarantee AI will recommend my business?

No — and you should not trust anyone who says yes. Assistants draw on training data, live retrieval and third-party sources nobody controls. What we can fix is the machine-readable layer, and we measure the change for you thirty days later.

How is this different from the free scans I have seen?

Free scans check whether your site has schema. This reads what the assistants actually answer for your category and city, and then explains the gap between the two. That gap is the whole product.

Do you implement the fixes?

If you want — that is a separate one-time engagement, quoted after the audit so you can see exactly what you are buying. Many clients hand the plan to their existing web person.

What do you need from me?

Your website URL, the city or metro you serve, and the one service you most want inbound calls for. That is it.

Find out which four names AI gives instead of yours

$500 · three business days · PDF, raw assistant answers, prioritized plan, 30-day re-test.

Get the audit — $500 After payment you get an email; reply with your URL and city and we start the same day.