Residential Solar AI: 4.1x Booked Consults in 45 Days
A Texas residential installer was spending $46,000/month on Google Ads and only reaching 22% of leads. Within 45 days of deploying our AI voice + WhatsApp + HubSpot stack, cost per booked consult dropped 61% and monthly installs climbed from 11 to 45.
What was broken
- $46,000/month Google Ads spend producing 187 leads — only 43 contacted
- After-hours and weekend enquiries went to voicemail (approximately 42% of all leads)
- Reps spent 3–4 hours a day on unqualified leads (renters, low bills, shaded roofs)
- No structured qualification meant proposals went out for jobs that never had a chance
- Nurture cadence stopped at touch 3 — 84% of solar closes happen after touch 5
What we built
Deployed AI voice agent that dials every Google Ads Lead Form and website submission within 60 seconds
WhatsApp Business API with green-tick fires in parallel, doubling contact channels
Structured qualification on bill, ownership, roof age and shading before a rep gets the lead
HubSpot-native automation pushes only pre-qualified, booked consults into the sales pipeline
45-day, 12-touch nurture cadence across voice, WhatsApp, SMS and email
Meta CAPI + Google enhanced conversions sync booked consults back to ad platforms
How it runs end-to-end
Lead lands
Google Ads / Meta / website form fires a webhook into n8n router.
Enrichment
Property data, utility, irradiance and dedupe check complete in ~3 seconds.
AI voice + WhatsApp
Parallel voice call (24s) and WhatsApp message (5s) hit the homeowner.
Qualification
AI qualifies bill, ownership, roof, timeline in a ~90 second conversation.
Book to HubSpot
Qualified leads book a virtual consult; HubSpot deal created with all fields populated.
Nurture + CAPI
45-day nurture runs; booked-consult event syncs to Google Ads + Meta for smart bidding.
The stack we used
AI Voice: LiveKit + GPT-4o Realtime + ElevenLabs Flash v2.5
Telephony: Twilio Voice (10DLC registered)
Chat: Custom GPT-4o chatbot on Next.js
WhatsApp: 360dialog WhatsApp Business API (green tick)
CRM: HubSpot Sales Hub Professional
Router: n8n self-hosted on Cloudflare
Conversion API: server-side GTM + native Meta CAPI + Google Enhanced Conversions
Analytics: PostHog + Metabase dashboard
What it looks like in production
Live call transcripts with sentiment and qualification scoring.
45-day nurture across proposal, permit and install stages in one thread.
Every field auto-populated by AI — no rep data entry required.
Cost-per-booked-consult tracked daily across every ad channel.
The numbers that moved
"Our cost per booked consult on Meta dropped from $184 to $71 in six weeks. Same ad spend, 2.6x the installs on the calendar."
Answers to the usual questions
How long did full deployment take?
Ten business days from kickoff to full go-live across voice, WhatsApp and HubSpot integration.
Did homeowners realise it was AI?
Roughly 8% asked at some point. The AI is transparent when directly asked and warm-transfers to a human on request.
What was the total investment?
$8,500 one-time build and $780/month all-in run cost including LLM tokens, WhatsApp, telephony and monitoring.
How is ROI tracked?
Booked-consult and closed-install events flow into a shared dashboard against ad spend. Payback was 42 days.
Other solar deployments
Commercial Solar AI: 2 x 400kW Deals Won From Weekend RFQs
A Pune-based C&I solar developer was losing weekend RFQs from procurement teams. Deploying an AI receptionist with intelligent routing captured every enquiry, qualified MW target and financing structure, and pinged the right named account executive on Slack — winning two 400kW deals in month one.
ReadSolar Audit AI: Free Online Audit Booked 340 Consults in 60 Days
A Florida installer replaced their generic 'get a quote' form with an AI-powered instant solar audit. Homeowners input their address and bill; the tool returned an estimated system size, savings and payback in 45 seconds — and 3.2x more visitors booked a consult.
ReadSolar Lead Qualification AI: 68% Reduction in Rep Time on Dead Leads
A UK MCS-certified installer was drowning in unqualified leads from Google Ads. Deploying AI qualification (ownership, bill, EPC rating, roof type) filtered 71% of leads before a rep ever saw them, freeing sales for high-intent conversations that closed 2.4x more often.
ReadReady for a similar deployment?
Book a 20-minute audit. We will show a live demo trained for your service area, panels and CRM — with real numbers, not slides.