AI for Solar Contractors: Design, Sales, and Install

Residential solar customer acquisition costs hit a five-year low of $0.60 per watt in 2025 and are projected to surge 40% to $0.84 per watt in 2026 per Wood Mackenzie. Over 100 solar companies closed in 2024 and 2025. The market is contracting, CAC is rising, and every soft cost is now under a microscope.

Solar is the highest-ACV home service category that an AI tool can meaningfully help. Average residential systems run $15K to $40K installed and the sales cycle averages 60 to 90 days. That means every leak in the design-to-install workflow compounds into real dollars fast.

This guide is for residential solar contractors doing $1M to $10M in revenue. What the AI tools actually do, where the soft-cost leverage is, and what to stop buying in a tougher market. The adjacent high-CPL vertical to study is AI agents for roofing contractors, which shares the photo-to-proposal and inspection-to-bid flow.

The market context every solar operator should know

The U.S. residential solar market installed 4,647 MWdc in 2025, down 2% year-over-year per SEIA's 2025 Year in Review, ending virtually flat after four years of growth. The Section 25D tax credit expiration pulled forward demand into 2025 and collapsed forecasts for 2026.

85% of all solar installation inquiries now start online per Solar Builder, which is why the front-office AI conversation has shifted from "nice to have" to "required to compete."

Lead-source close-rate data from Sunvoy puts referrals at 37.5% close rate and 29.2% net conversion, the best-performing channel by a meaningful margin. Yelp sits at 23.5% net conversion. Paid leads trail both.

The soft-cost problem AI targets

20% of residential solar business costs are customer acquisition per the National Renewable Energy Laboratory's 2022 report cited by Aurora Solar, and CAC rose 18% from 2020 to 2022. As hardware costs continue to fall, soft costs (sales, design, permitting, financing) are becoming a bigger share of total installed cost.

This is exactly the surface AI is attacking. Design automation, proposal generation, lead qualification, and sales follow-through are all soft-cost buckets. A 20% compression in soft cost on a $25K average project is $5,000 of margin recovered per install.

Where AI moves the needle in solar

AI-driven design. Aurora Solar's AI generates 3D models in under 10 seconds (down from 30 seconds in 2024) using training data expanded from 500K to 2 million designs per TechBrew. Aurora AI, Sales Mode, and Lead Capture AI reduce a 30-minute design cycle to 60 seconds.

End-to-end proposal automation. Enerflo's integration with Aerialytic enables sales reps to generate AI-driven solar designs and interactive proposals in under 2 minutes per Enerflo's published FAQ. The same photo-to-proposal AI estimating pattern is landing across high-ACV trades. OpenSolar 3.0, released late 2025, has evolved from a design tool into a full solar CRM with "Ada" AI assistant for auto-design and lead generation.

Lead qualification and response speed. Harvard Business Review's oft-cited data: leads contacted within one hour are 7x more likely to attend a sales meeting. LeadFuze data puts it more bluntly: 78% of customers choose the solar company that contacts them first. Every minute of response-time delay is revenue walking out the door. AI lead qualification is the leverage point for a referral-close-rate business.

Pipeline follow-through. The sales cycle is 60 to 90 days. Every proposal sits in an inbox for long stretches. Automated, context-aware follow-up at day 3, day 10, day 21, day 45 dramatically changes close rate on the long tail.

The actual tool landscape

Aurora Solar is the dominant design platform with an AI layer that now handles design generation, lead capture, and sales-mode guided selling. Best fit for mid-to-larger installers and companies running a dedicated design team.

OpenSolar launched AI Auto System Design in February 2025 to ~750 dealers and 3.0 (late 2025) repositions the product as a solar CRM with integrated design. "Ada" AI assistant handles auto-design and lead generation. Free tier still exists, which makes it friendly for smaller shops.

Aerialytic handles AI-driven roof analysis and design. Usually integrated via Enerflo or as a design-layer component rather than bought standalone.

Enerflo is the sales-platform layer, integrating with Aerialytic, Aurora, and other design tools to attach AI-driven designs directly to proposals. Strong fit for door-to-door and in-home sales teams.

Bodhi handles post-install customer experience with AI-driven updates. Niche but well-liked by installers focused on referral velocity.

Solargraf, Scanifly, Sighten round out the design and proposal stack with varying AI depth.

Generic AI receptionists (Rosie, Dialzara) can handle inbound call overflow but do not understand a solar conversation without extensive training.

Operator-level reality check

Solar Builder's installer-focused coverage repeatedly surfaces the same dynamic. Installers who invest in stacks that connect CRM, proposal, and design in one view have "complete visibility across the entire sales process from first call to install" and are pulling away on close rate and installed-cost efficiency. Installers who still do design on a separate platform from CRM, and proposals on a third, are losing 10 to 15 percentage points on close rate to stack friction alone.

The Greenlancer and Wood Mackenzie coverage of the 2024 to 2025 bankruptcy wave names the same pattern. Companies that closed were largely the ones with high CAC, thin operational margin, and heavy reliance on paid leads. Companies surviving the 2026 CAC spike are referral-dense, operationally tight, and automating the soft-cost side aggressively.

The referral math nobody talks about

Referral close rate at 37.5% versus paid-lead close rate at 10% to 15% means a referral lead is worth 2.5x to 3.7x a paid lead at install. On a $25K system with 20% margin ($5K gross profit), a single referral is $5K of margin that costs $0 to acquire.

The AI case for referral-velocity tools (Bodhi-class customer-experience AI, automated review and referral asks) is purely CAC math. In a market where CAC is jumping 40%, every extra referral is a CAC-free install.

Where AI is still weak in solar

Financing qualification. Solar sales are deeply financing-dependent and AI tools cannot replace a human reading the file and matching a buyer to the right lender and product. Use AI for design and proposal, use humans for financing.

Complex roof conditions. AI roof analysis is now good at basic pitch and azimuth. It still struggles on multi-plane, dormer-heavy, or shaded roofs where a human designer's judgment on array layout matters. Site visits still matter on the 20% of jobs that are not standard.

Permit and interconnection navigation. AHJ variability is extreme and permit and interconnection delays are the biggest install-cycle killer. AI has made modest progress here but it is not solved. Plan on human permitting ops for the foreseeable future.

How Sully fits

Most solar-specific tools are vertical specialists. Aurora does design. Enerflo does sales. Aerialytic does roof analysis. OpenSolar is evolving toward CRM. Each solves one slice, and installers end up with five to eight subscriptions that charge $1,500 to $5,000 a month combined, with the integrations varying from great to painful. If you are trying to roll those systems into one view, our comparison of Looker vs Metabase vs Sigma for contractors covers the seven factors that decide it.

Sully (sull.ai) is the pre-built AI operations layer for home service contractors in the $1M to $10M range, solar included. Missed-call follow-up, inbound lead qualification, proposal and bid follow-up, AI chat trained on your actual job history, and a morning brief for the owner. It sits across whatever design and CRM stack you already run.

For solar specifically the highest-value plays are response-time on inbound leads (the 1-hour, 7x stat) and multi-touch follow-up on the 60 to 90 day pipeline. Those are the leakiest surfaces, and where a pre-built agent layer pays back fastest. If you still want the ground-up path, read how to build an AI agent for home services.

Where to start in a tougher market

2026 is structurally harder than 2025. CAC is spiking, install volume is soft, and the long tail of weaker operators is going to close. The operators who come out of the contraction are the ones who did the unglamorous work: compressed soft cost, lifted referral velocity, and automated response-time on every inbound lead.

If you are not yet on a design platform with real AI (Aurora, OpenSolar), that is move one. If your design is good but your follow-up is a CSR with a spreadsheet, move two is automated pipeline follow-through. If both are solved, move three is referral velocity.

Every hour of CSR time and every hour of design time you save compounds directly into margin in a market where margin is shrinking. That is the whole play.

Sources: