A technically strong garage door site was still nearly invisible to AI assistants.

This business had a lot going for it: a fast modern website, city-page architecture, clean schema, clear phone CTAs, active analytics, and a real review base. It did not look broken from the outside.

But the audit score told a sharper story: 62/100 overall, with 80/100 conversion fundamentals and 80/100 structured data, but only 20/100 AI-assistant visibility and 25/100 content velocity. The site had good bones. The market-facing flywheel was not spinning.

What we found

AI assistants did not know the business existed. Category and urgent-intent prompts named aggregators and competitors instead. For a local service business, this is the new version of not ranking in Google: customers ask "who should I call?" and you are absent from the answer.

Branded search was not owning the brand. The business had recently changed how its name appeared online. The site, directories, and search results were not aligned, so entity signals were split across old and new versions.

The content engine existed but was not live. The audit found one public blog post and more than twenty drafted posts sitting outside the public site. That matters because AI assistants need citation-worthy pages to read, quote, and use as evidence.

The service-area story was inconsistent. Schema and profile settings said the business had expanded into Connecticut. The visible homepage still told customers the company served only its older market. That mismatch loses real visitors before they call.

Review velocity had stalled. The business had a strong total review count, but the latest visible reviews were many months old. For local search, freshness is not cosmetic; it is a ranking and trust signal.

Calls could not be attributed. Analytics was installed, but there was no call tracking or dynamic number insertion. For a phone-driven business, that means the owner could see traffic but not confidently connect visitor source to booked work.

The visitor lesson

The useful finding was not "get more traffic." The useful finding was that the business already had a site capable of converting visitors, but several high-intent visitor paths were leaking before they reached the phone: customers in the new geography could not see themselves in the homepage, AI assistants could not cite the business, review freshness weakened trust, and calls could not be tied back to source.

That is the kind of opportunity we look for: not vanity ranking work, but specific places where a real visitor with intent is being lost.

The fix order

1. 30 days: align the business name everywhere, request re-indexing, update the homepage service-area message, remove inaccurate trust claims, and publish the drafted content.
2. 60 days: add missing city pages, wire post-job review automation, claim missing directory profiles, and implement call tracking.
3. 90 days: maintain a weekly content cadence, track AI-assistant mention rate, add brand/category pages, run reactivation SMS, and review lift in impressions, calls, and profile interactions.

This case study is anonymized. The underlying audit was performed on a real local-service business; names, identifiers, and private details were removed before publication.