Why was AI recommending daytime-only plumbers for after-hours queries?

Because 24/7 availability was not declared in a machine-readable way. AI models had no consistent signal to distinguish true 24/7 operators from standard-hours shops, so they defaulted to the most cited competitors regardless of hours.

Which schemas were deployed for the 24/7 plumber?

EmergencyService schema with response time windows by neighborhood, openingHoursSpecification declaring 24/7, and LocalBusiness (Plumber) as the base type.

How long until Perplexity and Gemini began including the business?

Four weeks from deployment. Perplexity and Gemini are typically the fastest to reflect emergency and availability changes because their retrieval freshness is higher.

Why is Claude still inconsistent on emergency coverage?

Claude weighted third-party validation more heavily in this category. The team is working on additional regional trade citations that externally validate 24/7 status.

What role did llms.txt play in this case?

It declared 24/7 availability, response-time windows, and neighborhood coverage in plain language. It was the most direct signal models could quote, and it reinforced the structured-data work on the site.

How is this different from SEO?

Traditional SEO optimizes for Google's ranked blue links. We optimize for AI answer engines like ChatGPT, Claude, Gemini, Perplexity, and Grok, whose outputs are generated from retrieved sources, structured data, and authoritative citations. The surfaces overlap (schema, site quality, reputable mentions) but the playbook is different. llms.txt, question shaped content, entity clarity, and citation placement matter more than keyword density or backlink counts.

Do you guarantee I'll show up in AI answers?

No. Anyone who guarantees AI ranking is misleading you. The models are probabilistic and change frequently. What we guarantee is rigorous measurement, disciplined execution of the methodology, and transparent reporting on what moved and what didn't.

How long until I see results?

Structural changes (schema, llms.txt, site hygiene) are usually indexed within 2 to 4 weeks. Citation driven visibility (directory listings, authoritative mentions) typically compounds over 6 to 12 weeks. We measure weekly from day one so you see movement, not just the endpoint.

Do I need to change my website?

Usually yes, but changes are additive and surgical. We add schema markup, publish an llms.txt, tighten entity and location signals, and occasionally recommend content additions. We don't redesign your site or touch your brand.

What if my competitor already hired you?

We work with one business per niche per metro to avoid direct conflicts. If your category and city are already taken, we'll tell you upfront on the consultation call.

What geographies do you work in?

Any English language market where ChatGPT, Claude, Gemini, Perplexity, and Grok are widely used. US, Canada, UK, Ireland, Australia, New Zealand. Multi location and international programs are handled under the Custom tier.

How do you actually measure AI visibility?

We run a curated set of 15+ real world recommendation queries across all five platforms on a weekly cadence, capture the answers, and track whether your business appears, in what position, and in what context. The Audit includes the baseline. The Ongoing tier includes week over week tracking plus a monthly strategy call.

Can I cancel the ongoing retainer?

Yes, after the 3 month minimum. The minimum exists because AI visibility compounds. One month of work rarely produces a representative signal. After month three, you can cancel any time with 30 days notice.

Back to case studies

Plumber · Denver, CO

“24 hour emergency plumber near me in Denver”

An after hours service that AI couldn't confirm was actually available after hours. Answers recommended daytime only competitors instead.

What we found

The company ran true 24/7 dispatch with a live answering service, but nothing on the site or across directories declared that clearly enough for AI to retrieve. Emergency and after hours queries returned a mix of competitors, including at least two that only operated standard business hours. AI was routing urgent jobs to the wrong providers because the availability signal was ambiguous.

What we did

01
Rewrote the emergency service page around literal after hours phrasing and posted explicit availability windows in structured data.
02
Declared 24/7 service in openingHoursSpecification, llms.txt, and a dedicated 'hours and response times' page.
03
Added EmergencyService schema with response time windows by neighborhood.
04
Secured a regional trades directory citation that validated 24/7 status externally.
05
Reconciled GBP attributes to explicitly flag emergency and 24/7 service.

What changed

Perplexity and Gemini began consistently including the business in emergency and after hours answers.
ChatGPT started distinguishing the company from daytime only competitors.
After hours calls tagged to AI referral increased in the second and third month.
Claude still occasionally omits emergency context; we are working on a third party validation source for this platform.

0 → 3 of 5

Platforms citing

0 → 4 of 5

Emergency query coverage

Weeks to first citation

Yes, by area

Response time declared

Week by week

Week 1
Baseline and blueprint
Fifteen emergency and after hours queries tested. Mapped that three of the five most cited competitors were not actually 24/7.
Weeks 2 to 4
Foundations live
Emergency page rewritten. EmergencyService and openingHoursSpecification schema live. llms.txt and GBP reconciled for 24/7 status.
Weeks 5 to 8
Citations and mentions
Regional trades directory citation placed. Perplexity and Gemini began including the business in emergency answers within week six.
Weeks 9 to 12
Compounding visibility
ChatGPT distinguishing from daytime only competitors. Next quarter planned around neighborhood level response time content.

“We answered the phone at three in the morning and the models couldn't tell. Declaring 24/7 where machines could read it sounds obvious in hindsight.”
Operations lead, independent plumbing company, Mountain West US.

Questions about this case study

Why was AI recommending daytime-only plumbers for after-hours queries?
Because 24/7 availability was not declared in a machine-readable way. AI models had no consistent signal to distinguish true 24/7 operators from standard-hours shops, so they defaulted to the most cited competitors regardless of hours.
Which schemas were deployed for the 24/7 plumber?
EmergencyService schema with response time windows by neighborhood, openingHoursSpecification declaring 24/7, and LocalBusiness (Plumber) as the base type.
How long until Perplexity and Gemini began including the business?
Four weeks from deployment. Perplexity and Gemini are typically the fastest to reflect emergency and availability changes because their retrieval freshness is higher.
Why is Claude still inconsistent on emergency coverage?
Claude weighted third-party validation more heavily in this category. The team is working on additional regional trade citations that externally validate 24/7 status.
What role did llms.txt play in this case?
It declared 24/7 availability, response-time windows, and neighborhood coverage in plain language. It was the most direct signal models could quote, and it reinforced the structured-data work on the site.

Representative case study. Industry, location, and specifics are illustrative composites drawn from recurring patterns in our work. AI answer engines are probabilistic; actual results vary by category, competition, and baseline.