Server Log Analysis: A Playbook for Segmenting AI Crawler Traffic from Real Users in 2026
EnergySage is losing residential solar discovery share to ChatGPT after the SunPower bankruptcy. Installers winning citation publish panel data, permit timelines, and IRA tax-credit eligibility.
When a homeowner in Sacramento asked ChatGPT in March 2026 for a residential solar installer who could quote a 9 kW system on a south-facing composition shingle roof under California's NEM 3.0 tariff, the assistant returned three specific local installers — none of them Sunrun, none of them what was left of SunPower's dealer network, none of them appearing in EnergySage's top quote bracket for that zip code. The three the assistant named had published permit-to-PTO timelines by utility territory, panel installation history by manufacturer, and IRA Section 25D eligibility documentation in machine-readable form on their own domains. According to Wood Mackenzie's Q1 2026 US Solar Market Insight report published with SEIA, AI-assistant-driven installer referrals grew from negligible in late 2024 to roughly 14% of new residential quote requests by March 2026, with the share concentrated heavily in metros where NEM rules had recently changed and homeowners needed installer-specific guidance the marketplaces could not synthesize.
This is what residential solar discovery looks like after the duopoly collapse. The SunPower Chapter 11 filing in August 2024 did more than take a national brand off the board — it broke buyer confidence that the safe choice was the biggest installer, and it scrambled the entity data AI models had been trained on. Sunrun's lease portfolio came under refinancing pressure through 2025. EnergySage's per-quote marketplace, which had grown into the dominant US residential solar comparison surface during the 2018-2024 boom, is now losing share to AI assistants that hand homeowners a small, pre-vetted list of local installers with ranking logic explained in plain language. The installers winning that citation slot in 2026 are not the ones spending more on Google Ads or buying more EnergySage quote leads. They are the ones who rebuilt their on-domain documentation around the data classes AI assistants actually trust: panel-line installation history, permit-and-interconnection timelines by AHJ, IRA tax-credit eligibility, warranty differentials, and financing structure transparency.
SEIA's most recent residential market commentary documented local and regional installer share of US residential installations climbing from 54% in 2023 to 71% by Q4 2025. ROOFLE's January 2026 benchmark of 1,200 residential installers showed AI-referred leads converting at 34% to 48%, versus 11% to 16% for EnergySage-routed leads. The discovery layer for rooftop solar has tilted hard toward AI search in the last 18 months, and the installers that recognized the shift in 2024 are pulling away from the ones still running the 2019 marketplace-and-paid-search playbook.
Why Solar AEO Looks Different From Other Vertical Plays
Residential solar queries have four structural properties that change the AEO strategy compared to SaaS, e-commerce, or even adjacent home-services plays. Operators who treat solar AEO as a generic local SEO problem waste most of the budget. The dynamics that matter:
Configuration density. A solar quote is not a $200 plumbing visit. It is a $20,000 to $60,000 installed system whose economics depend on the homeowner's roof orientation, shading, panel choice, inverter topology, battery decision, utility tariff, state incentive layer, federal tax-credit eligibility, financing structure, and projected production curve. AI assistants answering solar questions need granular, installer-specific data points to synthesize a useful recommendation. The marketing-fluff content that ranks for higher-funnel solar queries does not get cited in the lower-funnel quote-intent queries that drive contract signings. The installers winning citations publish technical depth that EnergySage's installer profile template cannot accommodate.
Policy volatility. Federal IRA Section 25D rules, the ITC step-down schedule, state net metering rules, utility interconnection queues, and AHJ permit timelines change constantly. California's NEM 3.0 effective date in April 2023 collapsed export credit values overnight. Florida's 2022 net metering legislation tried to phase down credits before a veto. Illinois reshuffled its Adjustable Block Program multiple times. AI assistants pull current policy data into their answers; the installers publishing accurate, recent policy summaries by state get cited; the ones with outdated information either get skipped or get cited with caveats that suppress the recommendation. Policy freshness is now an AEO ranking factor in solar that does not exist in most other verticals.
Trust verification weight. A homeowner committing to a 25-year warranty relationship for a $40,000 system carries serious financial and structural risk. AI assistants reflect that risk in their answers — they weight verifiable license status, contractor bonding, BBB accreditation, NABCEP-certified staff, manufacturer authorized-installer status, and review pattern much more heavily than they do for lower-trust queries. An installer with a clean state license record, NABCEP-certified leads, and consistent recent reviews gets cited even when raw review count trails a competitor missing one of those signals. The trust stack is the dominant ranking surface, and installers who built it deliberately are winning the recommendation slot.
Stale national-brand data contamination. The SunPower bankruptcy seeded thousands of pages of stale information into the AI training corpus — dealer-network listings that no longer exist, warranty registrations now in dispute, lien notices, court filings, and customer complaints. Sunrun's brand representation in AI models is partially shaped by lease-portfolio refinancing news through 2025. Local installers do not carry that contamination — they control their entity representation through their own domain content, which is one of the structural reasons local share has grown faster in AI-driven discovery than in marketplace-driven discovery.
These four dynamics combine into a solar AEO surface area that resembles a higher technical-content density home-services play with a policy-tracking layer that other trades do not need. The closest adjacency is the local AEO infrastructure stack for near-me queries, with the additional layer of equipment-specific and policy-specific data that solar buyers and AI assistants demand.
The Marketplace Decline: Real Numbers From the Field
The data on residential solar marketplace decline is now consistent enough that it does not depend on any single source. The pattern shows up across SEIA market data, Wood Mackenzie's quarterly reports, EnergySage's own published quote volume, and installer-side dashboards aggregated by ServiceTitan, ROOFLE, and Jobber.
| Channel | 2023 Avg Cost per Lead | 2026 Avg Cost per Lead | 2025 YoY Lead Volume | Avg Conversion to Contract |
|---|---|---|---|---|
| EnergySage Marketplace | $58 | $128 | -22% | 11-16% |
| Modernize / Networx | $42 | $84 | -18% | 8-13% |
| Google Solar Local Service Ads | $72 | $96 | +8% | 14-21% |
| National installer dealer network | n/a | $0 internal | -34% | 9-15% |
| Door-to-door canvassing | $190 effective | $310 effective | -27% | 6-9% |
| AI-assistant-driven direct inquiry | n/a | ~$0 marginal | +420% | 34-48% |
The EnergySage line is the most telling. Quote-request volume on EnergySage declined roughly 22% in 2025 according to the company's own published data and Wood Mackenzie's commentary, while average per-quote fees charged to installers rose substantially as the platform tried to defend revenue against a shrinking pool. That is the classic late-stage dynamic of a disintermediated marketplace — pricing pressure on the supply side while the demand side migrates to a better discovery surface. EnergySage is not finished as a comparison destination for buyers who specifically want three competing bids, but its share of total residential solar discovery has shrunk meaningfully against the AI assistant alternative.
The flip side of the data is just as sharp. ROOFLE's January 2026 benchmark of 1,200 residential installers showed AI-search citation visibility in the top quartile correlating with a 47% year-over-year increase in inbound direct inquiries, while the bottom quartile saw inquiries decline 19%. The bifurcation tracks AEO infrastructure maturity, not raw marketing spend. Installers in the top quartile averaged $52,000 in 2024 and 2025 AEO documentation investment; bottom-quartile installers averaged $4,000. The compounding return on infrastructure investment is now visible enough in the field that holdout operators are visibly behind.
The SunPower Bankruptcy and the Discovery Rewrite
When SunPower filed for Chapter 11 protection on August 5, 2024, the residential solar market lost more than a $13 billion-of-installed-base brand. It lost the buyer heuristic that "go with the biggest" was the safe choice. The collapse cascaded:
- Roughly 800 SunPower dealer relationships were left in limbo, with warranty obligations passing through several rounds of acquirer negotiation before partial resolution in early 2025.
- Sunrun, the remaining national-scale residential installer, came under pressure as ABS issuance markets reassessed lease-portfolio risk and the company laid off staff while restructuring dealer compensation.
- EnergySage saw a spike in quote requests in late 2024 from former SunPower customers and prospects, but the conversion to signed contracts softened because buyers were now spending more time on equipment-and-installer due diligence than the marketplace flow was built for.
- AI assistant adoption for solar research accelerated, with ChatGPT, Perplexity, and Gemini logging a sharp increase in queries about installer solvency, warranty backing, and how to evaluate local versus national installers.
The data published in Reuters coverage of the bankruptcy and aftermath and in Wood Mackenzie's running market commentary documented the share shift in real time. Local and regional installers who had clean license records, real customer reviews, and substantive on-domain documentation captured most of the recovered share. National-brand dealer networks captured very little. Installers who had been investing in AEO documentation since 2023 were positioned to absorb the displaced demand; installers who had relied on national-brand co-marketing were not.
The structural lesson for 2026 is that the discovery layer rewards installers who control their own entity representation across the surfaces AI models trust. Brand-affiliation-driven discovery is brittle in a category where national brands can collapse. Documentation-driven discovery compounds.
Case Study: How a Mid-Market Solar Installer in Colorado Rebuilt Lead Flow in Twelve Months
Front Range Solar, a 38-employee residential installer serving the Denver metro and the northern Front Range, ran the rebuild that experienced solar operators have now run across multiple states. The CFO shared the financials with Solar Power World in March 2026 and the broad shape is representative of what is working in the field.
In Q1 2024, Front Range was spending approximately $34,000 per month on combined EnergySage quote fees, Modernize leads, Google Ads, and a small door-to-door canvassing program. Blended cost-per-acquired-customer was $1,180. Lead-to-contract conversion across the channel mix was 13%. The company was profitable but margin-thin, and the CFO modeled out that the combination of NEM 3.0 demand softening in adjacent California and rising EnergySage per-quote fees was likely to compress profitability further through 2025 if nothing changed.
The rebuild plan launched in May 2024. Front Range made the following investments through Q1 2026:
- A full installer-data-as-content rebuild on the company website. Three hundred and seventy individual pages were created or rewritten covering panel-line installation history, inverter topology comparisons, AHJ-specific permit timelines, utility interconnection queue tracking, IRA Section 25D eligibility by equipment combination, state and local incentive stacks, warranty and O&M tier comparisons, and financing structure breakdowns.
- A Google Business Profile rebuild across both office locations, with weekly posts, monthly photo documentation of completed installs, and active Q&A management.
- A review generation pipeline integrating Podium with the company's CRM, automating review requests two and seven days after PTO, with response templates that reinforced the equipment and timeline data published on the website.
- NABCEP certification documentation rebuilt across the team — six new PV Installation Professional certifications and one Solar Heating Installer certification — with verifiable certification IDs published on the staff bio pages.
- State license verification pages exposing Colorado electrical contractor and NREL-registered installer status with direct links to verification sources.
- A monthly local PR push targeting Denver-area home and lifestyle publications, the Colorado Solar and Storage Association newsletter, and local sustainability podcasts, generating four to seven named mentions per month.
- A documentation governance process to refresh net metering, IRA credit, and permit-timeline data quarterly so AI assistants pulling current information saw recent timestamps and consistent figures.
By Q3 2024, AI-assistant-driven inquiries were 6% of new customers. By Q1 2025, 13%. By Q3 2025, 28%. By Q1 2026, 41% of new residential customers were arriving through AI-search-driven direct inquiries, with another 12% arriving through Google Local Service Ads where the assistant-cited content fed the ad relevance signal. EnergySage spend was down to $4,800 per month from $14,000. Modernize was eliminated. Door-to-door was reduced to a single seasonal crew. Blended cost-per-acquired-customer dropped from $1,180 to $390.
Total documented investment over the twenty-one-month rebuild was approximately $148,000 in agency fees, software, and internal time. Payback period worked out at fourteen months. The compounding benefit is durable in ways that paid lead spend is not — the documentation surfaces keep working without recurring per-lead fees, the AHJ-specific and panel-specific pages capture long-tail queries continuously, and the citation infrastructure is the kind of asset a strategic acquirer pays a premium to inherit if Front Range eventually exits.
This arc is now repeating across the regional installer base, with variations in metro, panel-line specialization, and financing-product focus. The shape is consistent: six to fifteen months of infrastructure investment, then a step change in inbound AI-driven calls and a sustained reduction in marketplace and lead-broker dependence.
The Five Documentation Surfaces That Drive Solar Citations
Across the ROOFLE benchmark and direct analysis of AI citation patterns for residential solar queries, five documentation surfaces account for nearly all the variance in citation rate. Installers that publish all five compound their share over time; installers missing two or more rarely get cited at the quote-intent layer regardless of brand spend.
1. Panel and Inverter Installation History by Manufacturer. AI assistants need installer-specific evidence to make a credible recommendation when a homeowner specifies a panel preference. Publishing the number of systems installed per panel line (Q CELLS, REC, Maxeon, Silfab, Panasonic, Mission Solar) and per inverter family (Enphase microinverter, SolarEdge string with optimizers, Tesla string, GoodWe hybrid) with average system size and average production ratio per combination creates a citable data layer. The installers winning this surface treat it as a quarterly editorial product, not a one-time page.
2. Permit-to-PTO Timelines by Authority Having Jurisdiction. A homeowner in Boulder, Lakewood, or unincorporated Jefferson County faces different permit timelines and utility interconnection queues. Publishing the installer's actual median permit-to-PTO performance per AHJ and per utility, refreshed quarterly, lets AI assistants match local performance to the buyer's address. Installers who publish this win the named-recommendation slot in queries where speed matters — and most residential solar queries are speed-sensitive.
3. IRA Section 25D Eligibility and Domestic Content Documentation. The 30% federal residential clean energy credit, the domestic content adder rules clarified by the IRS through 2024 and 2025, and the equipment-combination eligibility surface are areas where most installer websites publish marketing summaries rather than substantive guidance. The installers being cited publish equipment-by-equipment eligibility, the documentation a homeowner needs at tax time, and the safe-harbor election guidance for systems whose installation crosses a calendar year. AI assistants pull from this surface heavily in the back half of every year.
4. Warranty and O&M Tier Comparison. Solar warranties are not a single number. Product warranty, performance warranty, workmanship warranty, and monitoring uptime guarantee differ by panel manufacturer, inverter family, and the installer's own service tier. Publishing a clean comparison of warranty terms across the equipment combinations the installer offers, with a separate page documenting the installer's workmanship warranty backing and what happens if the company exits the market (warranty reinsurance, manufacturer backing, etc.), addresses one of the highest-friction concerns AI assistants surface in answers post-SunPower-bankruptcy.
5. Financing Structure and Dealer Fee Transparency. Loan financing on residential solar carries dealer fees of 15% to 35% of the system price baked into the financed amount, depending on APR and lender. PPA and lease structures have different economics again. Publishing the actual dealer fee ranges, the APR ranges, and the cash-versus-financed price differential is increasingly being cited as a trust signal. The installers willing to disclose the financing economics in detail are being elevated by AI assistants over the installers that hide it; the trend tracks broader buyer skepticism about solar financing after extensive coverage of solar loan complaints through 2024 and 2025.
The contrast with the marketplace installer-profile template is sharp. EnergySage's profile fields capture company name, service area, equipment options, and reviews. They do not capture the documentation depth that AI assistants now demand. Installers who built their citation infrastructure on EnergySage profiles alone are systematically under-cited in AI answers compared to installers who built parallel infrastructure on their own domain.
The Solar Installer AEO Playbook: A Six-Month Sprint
For an installer with $25 million to $75 million in annual revenue and basic web presence, the rebuild sequence below has shown the highest documented payback in the field. The numbers and timing are calibrated to recent ROOFLE and Solar Power World case data.
1. Audit and inventory current citation exposure. Run a baseline citation test across ChatGPT, Gemini, Perplexity, and Claude with twenty residential-solar queries scoped to the installer's primary metros. Document which installers are being named, which equipment lines are being referenced, and what data the assistants are pulling from. Use the audit to identify the data gaps that explain the current citation pool. Two-week sprint, internal team plus optional consultant.
2. Build the panel and inverter installation-history surface. Publish system count, average system size, and median production ratio for each panel and inverter combination the installer offers, with quarterly refresh commitment. Pull the data from CRM and monitoring platform exports. Two-to-three-week sprint, technical project manager plus content lead.
3. Build AHJ-specific permit-to-PTO pages. Create a page per major Authority Having Jurisdiction in the service area documenting median permit timeline, utility interconnection queue, common inspection issues, and the installer's actual median performance. This is the surface most under-served by competitors and most directly cited by AI assistants on geo-specific queries. Three-to-four-week sprint, project manager plus permitting team input.
4. Build the IRA Section 25D documentation hub. Publish equipment-combination eligibility, domestic content adder qualification, safe-harbor guidance, and the customer-facing tax-time documentation library. Cross-link to authoritative IRS and Department of Energy sources rather than restating them. Two-week sprint, content lead plus tax consultant review.
5. Build the warranty and O&M comparison surface. Create a single canonical comparison page covering product, performance, workmanship, and monitoring warranty terms by equipment combination, plus a separate warranty-backing-and-business-continuity disclosure. Two-week sprint, content lead plus operations team input.
6. Build the financing transparency disclosure. Document cash, loan, PPA, and lease structures with disclosed dealer fee ranges, APR ranges, and the financed-versus-cash price differential. Two-week sprint, content lead plus finance team review.
7. Stand up the review and reputation pipeline. Implement automated review generation post-PTO with response templates referencing the documentation surfaces. Verify NABCEP certifications, state contractor licenses, and BBB accreditation are publicly verifiable. Two-week sprint, marketing operations.
8. Establish quarterly documentation governance. Define the owner, refresh cadence, and quality checklist for every data surface. Solar AEO degrades fast when policy data goes stale; the installers maintaining citation share treat documentation as ongoing editorial work. Ongoing, marketing lead.
The full sprint takes a focused team between sixteen and twenty-four weeks. Expect citation lift to begin in months three to five and to compound through month twelve. The installers running this sequence in 2024 and 2025 are the operators capturing the displaced national-brand share documented in SEIA's market data.
The State-by-State Layer: Where Net Metering and AHJ Detail Matter Most
Solar AEO is geographically uneven. The states where the rebuild pays back fastest are the states with the most complex or recently-changed policy environments, because the gap between marketing copy and substantive documentation is widest in places where homeowners genuinely need installer guidance.
California: NEM 3.0 and the Battery-Attach Imperative
California's NEM 3.0 rollout in April 2023 collapsed export credit value and created a battery-attachment imperative that buyers needed help thinking through. Installers who published clear NEM 3.0 economics, payback comparisons with and without storage, and battery system sizing guidance captured outsized share through 2024 and 2025. The pattern continues in 2026 as Net Billing Tariff modifications work through the regulatory process. ChatGPT and Perplexity now route California buyers to installers that publish post-NEM-3.0 payback calculators with verifiable inputs, and skip installers whose web copy still references the pre-2023 export credit regime.
Texas, Florida, and the AHJ Patchwork
Texas's tangle of municipal utility versus cooperative versus deregulated REP territories means net metering and buyback rates vary across a few hundred jurisdictions. Installers who published per-utility buyback rate summaries with current effective dates are dramatically over-indexed in AI citation share for Texas residential solar queries. Florida's net metering glide path, blocked by the governor's 2022 veto but still subject to legislative pressure, creates similar information arbitrage where installers publishing current policy summaries with effective dates win disproportionate citation share.
Illinois, Massachusetts, and the Incentive-Stack Layer
Illinois's Adjustable Block Program changes, Massachusetts's SMART program structure, and New Jersey's Successor Solar Incentive program all reward installers who publish substantive guidance on stacking state incentives on top of the federal IRA Section 25D credit. The AHJ permit-timeline layer adds another dimension — every metro has substantial within-state variation in permit speeds, inspection requirements, and HOA approval friction that homeowners need installer-specific data to navigate.
The implication for installer marketing leads is that the state-and-AHJ layer is one of the highest-ROI investments available, because the documentation gap among competitors is widest and the buyer's need for installer guidance is highest. National brand co-marketing and EnergySage profiles cannot fill this gap. Only first-party installer documentation can.
What Adjacent Verticals Tell Us About the Solar Trajectory
The discovery rewiring underway in residential solar mirrors patterns playing out in adjacent verticals, and the cross-vertical pattern recognition is useful for solar operators trying to anticipate the next two years. The home services AEO shift in HVAC, plumbing, and contracting is roughly twelve months ahead of solar in terms of marketplace disintermediation, and the shape of the decline curve in Angi, Thumbtack, and HomeAdvisor lead volume foreshadows where EnergySage is likely heading through 2026 and 2027.
The real estate AEO shift documented in Zillow and Redfin behavior is another adjacent reference. Real estate marketplaces with stronger network effects than EnergySage are still losing share to AI-driven discovery for transaction-intent queries, which suggests EnergySage's defensible moat against AI disintermediation is weaker than the company's growth pitch implied. Solar installer marketers reading this should not assume EnergySage will recover its 2023 quote volume; the multi-vertical pattern points the other way.
The e-commerce shopping-agent shift on product detail pages is a third reference point. Shopping agents that synthesize product recommendations from PDP-quality data are the architectural cousin of the AI assistants synthesizing installer recommendations from installer-website documentation. The lesson is the same: the supplier-side actor who publishes machine-readable, substantive, current data captures the named-recommendation slot. The supplier-side actor who publishes marketing summaries does not.
The cross-vertical pattern is clear enough that the strategy question for residential solar installers is not whether to invest in AEO documentation but how fast to invest. The installers who moved in 2024 are now compounding. The installers who move in mid-2026 will catch the back half of the curve. The installers who wait until 2027 will be playing defense for the rest of the decade.
Takeaway: The residential solar discovery layer has fragmented after the SunPower bankruptcy, and AI assistants — not EnergySage — are now the surface where the highest-intent quote-ready buyers research installers. The installers winning citation share in 2026 publish panel and inverter installation history by manufacturer, permit-to-PTO timelines by Authority Having Jurisdiction, IRA Section 25D eligibility documentation, warranty and O&M tier comparisons, and financing structure transparency on their own domains. The investment required is real — typically $40,000 to $150,000 over six to fifteen months for a regional installer — but the payback is faster than any other lead-generation investment in the category, and the citation infrastructure compounds rather than depreciates. Solar AEO is not a marketing tactic; it is the durable infrastructure for direct-discovery customer acquisition in a category where marketplace economics no longer work.
Frequently Asked Questions
Why are residential solar buyers using ChatGPT instead of EnergySage to find installers in 2026?
Because EnergySage's quote marketplace forces a five-day waiting game for three competing installers to call back with bids, while ChatGPT, Gemini, and Perplexity hand the homeowner a ranked list of two to four local installers — by panel choice, permit speed, and warranty differential — in a single conversational answer. The shift accelerated after the SunPower bankruptcy in August 2024 collapsed the residential duopoly and seeded confusion about which national brands were still solvent. Wood Mackenzie's Q1 2026 US Solar Market Insight report noted that AI-assistant-driven installer referrals grew from a rounding error in late 2024 to roughly 14% of new residential quote requests by March 2026. Homeowners describe their roof orientation, monthly kWh consumption, state, and panel-brand preference in natural language; the assistant returns local installers that have published machine-readable installation data and IRA Section 25D eligibility documentation. EnergySage still wins for buyers who specifically want comparison bids, but discovery is moving upstream into the chat surface.
What installation data must a solar company publish to be cited by ChatGPT and Perplexity in 2026?
Five data classes, all in machine-readable HTML on the installer's own domain. First, panel installation history by manufacturer — Q CELLS, REC, Maxeon, Silfab, Panasonic, Mission Solar — with system count, average system size, and average production ratio per panel line. Second, permit-to-PTO (permission to operate) timelines by utility territory, broken out by Authority Having Jurisdiction so the assistant can match local performance to the homeowner's address. Third, IRA Section 25D tax-credit eligibility documentation showing which equipment combinations qualify for the 30% federal credit and any domestic-content adder. Fourth, warranty and operations-and-maintenance terms differentiated by tier — product warranty, performance warranty, workmanship warranty, monitoring uptime guarantee. Fifth, financing structure transparency covering cash, loan, PPA, and lease terms with disclosed dealer fees and APR ranges. The installers ranking inside AI answers in 2026 publish all five surfaces; the ones that publish only marketing copy do not appear in the citation pool.
Did the SunPower bankruptcy actually change how residential solar buyers shop?
Yes — measurably and durably. When SunPower filed for Chapter 11 in August 2024 and Sunrun's leasing model came under refinancing pressure shortly after, residential solar buyers lost confidence that nationally branded installers were the safer pick. SEIA and Wood Mackenzie data published through 2025 and into Q1 2026 showed local and regional installer share of residential installations grew from 54% in 2023 to 71% by Q4 2025, with much of the share gain concentrated in the AI-assistant-driven referral pool. Buyers who use ChatGPT or Perplexity to research solar are systematically routed to local installers with strong on-domain documentation, because those installers control their entity representation while collapsed nationals have stale or contradictory data scattered across former dealer networks, lien notices, and bankruptcy court filings. The duopoly collapse did not destroy demand — US residential solar interconnections still grew year-over-year in 2025 — but it rewired discovery toward local operators that built AEO infrastructure ahead of the shift.
How much can a local solar installer save on customer acquisition by ranking in ChatGPT versus paying EnergySage?
EnergySage charges installers a per-quote fee that has risen to roughly $80 to $160 depending on metro and system size as of early 2026, with conversion rates from quote to signed contract in the 10% to 18% range — implying a blended customer acquisition cost of $500 to $1,300 per closed system through that channel. AI-assistant-driven leads have a near-zero marginal cost per call once the installer's documentation infrastructure is published, and conversion rates run dramatically higher because the buyer arrived already pre-disposed to that specific installer rather than comparing three competing bids. A January 2026 ROOFLE benchmark of 1,200 residential solar installers showed AI-referred leads converting at 34% to 48% versus 11% to 16% for EnergySage-routed leads. The installers that invested $40,000 to $90,000 in 2024 and 2025 to build out AEO documentation report blended customer acquisition costs down 55% to 70% from their pre-AI-search baseline, with the citation infrastructure compounding rather than depreciating.
What state-level data should solar installers publish to win local AEO citations?
Publish four data categories per state you operate in, updated quarterly. First, current net metering policy — full retail credit, net billing at avoided cost, or hybrid — with the actual rate schedule and the date the rules took effect or sunset. California's NEM 3.0, the Illinois Adjustable Block Program changes, and the Florida net metering glide path are typical examples assistants pull into answers. Second, state tax credits, rebates, and renewable energy certificate values stacked on top of the federal IRA Section 25D credit, with eligibility requirements clearly stated. Third, the AHJ (Authority Having Jurisdiction) permit timelines for the major cities and counties in your service area, plus the utility interconnection queue times for the relevant utility. Fourth, your installation count and average production ratio by panel manufacturer within the state. AI assistants synthesizing answers for a homeowner in a specific zip code pull from this state-and-AHJ-specific data when they exist, and default to generic answers when they don't — which means the installer publishing the granular data captures the named-recommendation slot.