How an AC Repair Company in Phoenix Got Into Google's AI Overview

Business Background

A mid-size AC repair company serving the greater Phoenix metro had been in business for 11 years. With 8 technicians, they were doing about $1.2 million in annual revenue — solid, but plateauing. Their traditional SEO was working: they ranked #3 for "AC repair Phoenix" and had 140+ Google reviews averaging 4.8 stars.

The problem was invisible. Call volume was flat despite a growing service area. The owner noticed that when he asked his own iPhone "who does AC repair near me," his company didn't come up. He typed the same question into ChatGPT. Again, nothing. Three competitors appeared by name. His business, with a decade of history and top organic rankings, was simply not part of the AI's answer.

The Problem: Ranking #3 But Invisible Where It Matters

When the gaflow AI Search Readiness assessment was run against this business, the issues were clear:

• Entity fragmentation: The business had three different phone numbers across nine directories, two different name spellings ("ABC Cooling" vs. "ABC Cooling & Heating"), and a mismatched address on two major directories after a location change 18 months prior.
• No FAQ schema: The website had solid service pages with good keyword density, but zero structured data. Google's AI Overview couldn't extract structured answers to cite.
• GBP Services section mostly empty: Only 4 of 11 actual services were listed. The Q&A section had no business-posted answers.
• Review velocity stalled: 140 reviews, but only 6 in the prior 90 days. AI systems treat stale review signals as low-confidence.
• No conversational content: The website had no FAQ pages and no content structured to answer the questions homeowners actually type into ChatGPT — "how long does AC replacement take," "what does AC repair cost in Phoenix," "is it better to repair or replace an AC unit."

This is the pattern we see most often in established HVAC businesses. Their traditional SEO is working. They rank. But AI systems are reading entirely different signals, and this company was weak on almost all of them.

The Solution: Claude Skill Bundle — Full Implementation

After purchasing the Skill Bundle and installing it in Claude, they ran the full AI Search Readiness audit using the 12-dimension scoring rubric. Their composite score came back at 34 out of 100 — below the threshold for reliable AI recommendation. The audit generated a prioritized action list. Here's what they actually did:

Step 1: Entity Audit and NAP Cleanup (Week 1)

Using the 30-directory citation tracker included in the bundle, they identified every listing where their name, address, or phone number was inconsistent. They worked through 28 directories over 4 days, standardizing everything to the exact legal business name and current address. They chose one primary phone number and made it consistent everywhere. This is unglamorous work, but it's the foundation — nothing else sticks if the entity is fragmented.

Step 2: GBP Completion and Q&A Population (Week 1–2)

They went through the 50-point GBP audit punch list. Every service was added with a 2–3 sentence description. Business hours were verified. The business description was rewritten to explicitly mention Phoenix, Scottsdale, Tempe, and Mesa (their actual service area). Twelve Q&A pairs were added covering the most common questions: pricing, licensing, warranty, emergency availability, and service area. Photos were refreshed with recent job site images.

Step 3: FAQ Schema on Every Service Page (Week 2–3)

Five service pages got new FAQ sections — AC repair, AC installation, heat pump service, air handler replacement, and duct sealing. Each FAQ section had 6–8 questions with substantive answers. The Claude Skill generated the FAQPage schema JSON-LD blocks for each page. The dev dropped them in. Done.

The questions were drawn from actual conversational patterns: "How much does AC repair cost in Phoenix?" "How long does AC installation take?" "What warranty comes with a new AC unit?" "Do you offer same-day AC repair?" These aren't keyword-stuffed answers — they're the actual answers a knowledgeable tech would give.

Step 4: Review Velocity System (Week 3 onward)

They implemented the review request workflow from the bundle's Review Velocity Strategy template. Every closed job triggered an automated text 24 hours later with a direct link to the Google review page. Technicians were briefed to mention reviews naturally at job completion. Within 3 weeks, they went from 2 reviews per week to 7–9 per week. The recency signal shifted meaningfully within 30 days.

Step 5: Conversational Content (Week 3–5)

Using the 9-Section Strategy Doc as a framework, they identified the top 15 questions their CSR team answers weekly. They created standalone FAQ pages for the top 8: AC replacement cost, repair vs. replace decision guide, heat pump vs. traditional AC comparison, Phoenix-specific summer maintenance tips, and four more. Each page was 600–900 words, structured to directly answer the question, not to rank for a keyword.

Results: 6 Weeks to First AI Overview Appearance

The first AI Overview appearance happened at week 6, for the query "AC repair company Phoenix reviews." By week 8, they were appearing for two more: "best AC repair Phoenix" and "emergency AC repair Phoenix AZ." All three were in the top 3 sources cited in the AI Overview.

Before vs. After Comparison

Week-by-Week Timeline

"We'd been telling ourselves we were doing fine on SEO because we ranked #3. But we were completely invisible to half the people looking for HVAC service now. The audit showed us exactly why, and the steps were specific enough that we could actually execute them — not just 'improve your GBP,' but exactly what to add and why. Two months later we're in the AI results, and the phone is ringing more."

— Composite perspective based on typical outcomes from mid-size Phoenix market HVAC operators

ROI: Investment Paid Back in the First Month

The Skill Bundle is a $1,997 one-time investment. In the first month after AI Overview appearances began, the business attributed 6 new jobs directly to callers who said they "saw them pop up when they searched" — a new source attribution not previously appearing in their call tracking. Average job value in residential AC repair: $380 for a diagnostic/repair, $4,200+ for a replacement. Even with conservative attribution, the investment returned multiple times in month one.

More importantly: the signals built during implementation don't decay. The schema is on the site. The NAP is clean. The review velocity system is running. The conversational content is indexed. These are compounding assets — not a campaign you run once.

Tools Used in This Implementation

• 12-Dimension AI Search Readiness Scoring Rubric — identified the exact gaps
• 30-Directory Citation Tracker — found and fixed NAP inconsistencies across 28 directories
• 50-Point GBP Audit Punch List — walked through every GBP field systematically
• FAQ Schema Generator (Claude Skill) — produced JSON-LD blocks for all 5 service pages
• Review Velocity Strategy Template — built the automated text workflow
• 9-Section Strategy Doc — structured the conversational content roadmap