AirCentral was making all the calls. Now the calls come to them.
AirCentral was competing on Angi and losing on price — every spring and fall dead slow. Deep-Y built an AI-search visibility system: they now appear on ChatGPT, Perplexity, and Google AI Overviews when homeowners ask for HVAC in their market. ~6 qualified inbound calls per day. People find them, decide, and call — nobody competing on price.
~6/day
Inbound Leads from AI Search
Lower CPL
vs Ad-Dependent Model
Higher CV
Meeting → Client Rate
$240K
Documented Pipeline
Every HVAC owner knows the price war. Angi sends the lead to whoever answers first. Google Ads costs climb every year. Competitors multiply. The phone rings for the cheapest quote, not the best service. AirCentral had real skills and great reviews — but zero presence where buyers had already decided. In 2026, that decision happens on ChatGPT, not Google. We built the system that put them there. ~6 qualified inbound calls per day. Customers who called because they were recommended — not because AirCentral outbid someone on a lead marketplace.
The ceiling wasn't their service. It was their invisibility on the channel where buyers had already decided.
Before
Angi
Competing on price. Slow seasons. Every lead shared with three competitors.
→
After
~6/day
Inbound leads from ChatGPT + Perplexity + Google AI Overviews. Customers already decided.
Deep-Y built the AI-search visibility system: content structured for LLM citation, query-cluster coverage across ChatGPT and Perplexity, and Google AI Overview optimization. AirCentral now appears as the recommended answer — not one option in a list.
The Challenge
Good service. Great reviews. Invisible where buyers were already deciding.
Before Deep-Y
Angi · price competition
Every lead shared with three competitors. Whoever answered first, offered the lowest price, won. Spring and fall nearly silent. Google Ads cost rising every year. No way to reach buyers who'd already decided before calling anyone.
After Deep-Y
~6 inbound leads/day
An AI-search visibility system AirCentral owns: structured for ChatGPT, Perplexity, and Google AI Overviews. Appears as the recommended answer for HVAC queries in their market. No retainer to keep it running.
In 2026, 63% of local service buying decisions start with a conversational AI query — not a Google search. Homeowners ask ChatGPT "best HVAC company in [city]" and get a ranked recommendation before they ever touch a search result. AirCentral wasn't in those answers. Now they are.
~6/day
Qualified inbound leads from AI search. ChatGPT · Perplexity · Google AI Overviews — customers who already decided, calling to confirm.
~6 qualified inbound leads per day — from AI search alone
~6/day
When a homeowner asks ChatGPT "best HVAC company in [city]," AirCentral appears in the answer. When they search Perplexity for HVAC repair, AirCentral is cited. When Google generates an AI Overview for heating and cooling queries in their service area, AirCentral is recommended. These callers have already decided — the call is to confirm, not to compare.
"We were competing with every HVAC company in our market on Angi and losing on price. Deep-Y built a system that puts us in the answer when someone asks ChatGPT for the best HVAC company in our area. We're getting 6+ qualified calls a day from people who've already decided. Conversion is higher, average jobs are larger, cost per lead is down. It runs without us doing anything."
AC
AirCentral
Residential & Commercial HVAC · US market
The Approach
How we built the AI-search visibility system AirCentral now owns
Deep-Y built a 3-layer AEO/AIO system — AI citation architecture, query-cluster content, and inbound routing — and handed it over to run in AirCentral's name. No retainer required. AirCentral's team didn't write a single piece of content.
AI Citation Architecture
~6 Inbound Leads/Day
01
AI citation architecture
Restructured AirCentral's digital presence to be cited by large language models: entity-anchored content formatted for LLM extraction, brand authority seeded across local directories and review platforms that LLMs pull from, schema markup aligned to how AI systems process local service providers.
02
AI Overview query-cluster domination
Reverse-engineered the specific queries triggering Google AI Overviews for HVAC — urgency queries ("AC not working"), seasonal readiness queries ("AC tune-up before summer"), and research queries ("best HVAC company in [city]"). Each cluster got dedicated AIO-optimized content built to be cited as the primary recommendation.
03
ChatGPT + Perplexity presence
Structured AirCentral's content and brand signals to appear in conversational AI responses on ChatGPT and Perplexity. Q&A content formatted to match exactly what LLMs surface for HVAC service queries. AirCentral now appears in responses — not as a link, but as the recommended answer.
04
Inbound routing — lead to calendar in 15 minutes
When leads arrive inbound — call, form, or text — they're already warm. They named their problem. They found AirCentral through a recommendation. We built intake routing so every inbound lead hits the right technician calendar within 15 minutes of first contact.
05
89% follow-up open rate (supporting layer)
For leads that don't book immediately, we built follow-up sequences with 89% open rates. Home-age signals as subject hooks, isolated sending domains — same deliverability discipline as the AI-search layer. No cold outreach from lists, just follow-through on inbound intent.
06
Monthly compounding loop
New query clusters added monthly as AI search evolves. Content updated to maintain citation advantage as competitors catch on. AirCentral's AI-search position compounds — each month's investment strengthens the next.
Under the Hood
How the AEO/AIO system was built
The AirCentral AI-search system had four technical layers, each solving a specific gap in how LLMs discover and cite local service businesses.
01
LLM entity mapping (week 1)
Audited how ChatGPT, Perplexity, and Google AI Overviews currently represent AirCentral. Identified citation gaps: which authoritative sources the LLMs pull from for HVAC queries in their market, and where AirCentral was absent. Built the entity map — what needs to exist for LLMs to cite them confidently.
02
Citation architecture (weeks 1-2)
Restructured existing content for LLM extraction: entity-anchored Q&A pages, structured data markup for service provider data, brand presence seeded across the local directories and review platforms that LLMs pull from when forming service recommendations.
03
Query cluster content (week 2-3)
Built dedicated AIO-optimized content for each query cluster: urgency queries, seasonal queries, and research queries. Each piece formatted as the exact answer an LLM would surface — not keyword-stuffed pages, but structured answers to the questions homeowners actually ask AI assistants.
04
Inbound capture and routing (week 3)
Built intake routing so every inbound lead — call, form, or text — hits the right technician calendar within 15 minutes. Follow-up sequences for non-immediate bookers achieve 89% open rate using home-condition signals as subject hooks. Zero cold outreach; pure inbound intent follow-through.
The Result
What changed when the calls started coming in
Inbound leads/day from AI search~6 qualified leads
Lead cost vs ad-dependent modelSignificantly lower CPL
Meeting → client conversionHigher than cold or ad-sourced leads
Average sale valueHigher — inbound buyers don't negotiate, they confirm
Follow-up open rate89%
Documented pipeline (residential)$240K in 90 days
Signed residential jobs22 in 90 days
New headcount requiredZero
Ongoing agency retainerZero — they own the system
AirCentral went from a company that makes calls to a company that receives them. The phone rings with people who've already searched, already been recommended, and are calling to schedule — not to compare.
Documented Pipeline (Phase 1 · Supporting Metric)
Cumulative signed residential pipeline — documented alongside the AEO/AIO system build
Cumulative signed pipeline value · 12-week residential lead gen campaign · AirCentral HVAC
AEO/AIO System Metrics
Metric
Result
Inbound leads/day (AI search)
~6 qualified
Cost per lead vs ad model
Significantly lower
Meeting → client conversion
Higher than ad-sourced
Average sale value
Higher (inbound buyers)
Follow-up open rate
89%
Documented pipeline (90 days)
$240K
The system in motion · AI outreach running on AirCentral domains
Video AI System · Layer 2
1,253 videos. 6 months. Zero video crew.
After the AI-search visibility system was running and generating inbound leads, we layered in a Video AI System. Scripts to distribution, automated — same client, second system, zero added headcount.
Videos produced1,253 in 6 months
Video team requiredZero
Systems running simultaneouslyOutreach + Video
Additional headcountNone
Two systems running simultaneously. Neither required a hire.
Your competitors are invisible on AI search. Most of them don't know it yet.
We'll audit your current AI-search visibility, identify the query clusters your buyers use, and show you exactly what it'd take to appear as the recommended answer — before your competitors figure this out.
