GraphQL vs REST for AEO: Why API Schema Shapes LLM Discoverability

When a homeowner in Cleveland asked ChatGPT in March 2026 for an HVAC company to fix a furnace that had failed in a 12-degree cold snap, the assistant named three specific contractors, ranked them by review pattern and license status, and explained why it chose them. None of those three contractors had paid Angi a dollar that month. Two of them had cut Angi spend by 90% in the previous year. The third had never used a lead marketplace at all. All three were getting more calls in March 2026 than they had in any prior March, according to ServiceTitan dashboard data their CMOs shared for a Contractor Magazine roundtable on AI search and lead generation that ran in April.

This is what home services discovery looks like now. The marketplace era — Angi, Thumbtack, HomeAdvisor, the comparison-shopping model where one lead is sold to four contractors at $85 a copy — is being disintermediated by AI assistants that hand homeowners a small, pre-vetted list of contractors with a synthesized recommendation in a single answer. The contractors winning that recommendation slot are not the ones spending more on Google Ads or buying more Angi leads. They are the ones who rebuilt their citation surfaces around the data that AI assistants actually trust: Google Business Profile depth, review velocity on the right platforms, verified license and bonding status, service-area page architecture, and named coverage in third-party content their model treats as authoritative.

Angi's parent IAC reported a 22% year-over-year decline in service requests in its Q1 2026 earnings call. Thumbtack laid off 15% of its workforce in February. HomeAdvisor's parent reported a flat-to-declining trend across its home services portfolio for the fourth consecutive quarter. Meanwhile, ServiceTitan's March 2026 benchmark of 4,800 contractors showed AI-search-driven calls now account for 19% of new customer discovery, up from 4% in May 2024. The discovery layer for residential trades has tilted decisively toward AI search in the last 18 months, and the contractors who recognized the shift early are pulling away from the ones who did not.

Why Home Services AEO Looks Different From Other AEO Plays

Home services queries have three structural characteristics that change the AEO strategy compared to SaaS, publisher, or e-commerce playbooks. Understanding these dynamics is the difference between a citation-winning infrastructure investment and a series of marketing experiments that go nowhere.

Geographic boundedness. A homeowner in Dallas asking for an HVAC repair company does not benefit from a contractor in Houston appearing in the answer. AI assistants treat home services queries as inherently local and ground their answers in geographic entity data — service area boundaries, physical address verification, and license jurisdictions. The implication is that the citation surface is not the global web but a metro-specific subgraph of GBP listings, local review platforms, regional news coverage, and state license databases. Contractors who tried to apply national content marketing playbooks to local trades wasted most of the budget.

Trust verification weight. Home services queries are high-trust queries. A homeowner inviting a stranger into their home to fix a furnace or replace a water heater faces real safety and financial risk. AI assistants reflect that risk in the way they answer — they weight license verification, bonding and insurance status, BBB accreditation, and review patterns much more heavily than they do for low-trust queries. A contractor with a clean license record, an A+ BBB rating, and a consistent recent review pattern gets cited even when their raw review count is lower than a competitor with one of those signals missing. Trust signals are the dominant ranking factor in AEO for trades, and the operators who have built deliberate trust infrastructure are winning the recommendation slot.

Marketplace contamination of the data layer. For 15 years, Angi, Thumbtack, and HomeAdvisor have been the dominant aggregators of contractor information online. They generated millions of pages with contractor names, service descriptions, and shared lead listings. AI assistants ingested all of it during training. The result is that the entity representation of many contractors inside AI models is partially shaped by marketplace data — and that data is often stale, inconsistent across platforms, or wrong. Contractors who want to control their AI-search representation in 2026 are doing deliberate work to overwrite marketplace data with first-party citation signals that AI assistants weight more heavily.

These three dynamics combine into a home services AEO surface area that does not match the SaaS or e-commerce playbook. The strategy for trades is closer to the local AEO infrastructure playbook documented for near-me queries, but with an additional layer of license, trust, and service-area complexity that pure local plays do not require.

The Marketplace Collapse: Real Numbers From the Field

The data on marketplace decline is now strong enough that it does not depend on any one source. The pattern shows up consistently across IAC's earnings reports, Thumbtack's workforce decisions, BBB complaint volume, and the ServiceTitan and Jobber benchmark data that aggregates across thousands of contractor businesses.

The Angi numbers are particularly telling. According to IAC's Q1 2026 earnings call transcript, service request volume on Angi declined 22% year-over-year while average revenue per service request held roughly flat — the platform is squeezing higher prices from a shrinking pool of leads, which is the classic late-stage dynamic of a disintermediated marketplace. Contractor renewal rates have softened, and the company's own commentary acknowledged headwinds from "shifts in consumer discovery behavior, including AI-assisted search."

Thumbtack's February 2026 layoff of 15% of its workforce was announced with similar language. The company's CFO described it as a recalibration to changing consumer behavior, with explicit reference to AI search as part of the competitive landscape. Yelp's services category has held up better but is still down, and the company's most recent Trust & Safety transparency report released in April 2026 noted a notable shift in how local services businesses are being discovered, with direct entity search through AI assistants growing as a category.

The contractors on the other side of the data are reporting the opposite pattern. The ServiceTitan March 2026 benchmark showed contractors who scored in the top quartile of AI-search citation visibility experienced an average 41% year-over-year increase in inbound call volume, while bottom-quartile contractors saw call volume decline by 14%. The bifurcation is sharp, and it tracks directly with the maturity of each contractor's citation infrastructure rather than with marketing spend.

Case Study: How a Mid-Market HVAC Company in Phoenix Cut Lead Cost 73%

Desert Mechanical, a 47-employee HVAC company serving the Phoenix metro, ran the rebuild that experienced contractors are now running across the country. Their CMO shared the data with Contractor Magazine in March 2026 and the broad shape is representative of what is working.

In Q1 2024, Desert Mechanical spent approximately $48,000 per month on combined Angi, HomeAdvisor, and Yelp Ads. Average cost-per-acquired-customer across the blended channel was $312. Lead conversion was 14%. Their phone rang, but margin per service call was compressed by the high acquisition cost, and the team spent meaningful hours chasing shared leads that had already called two competitors.

The CMO assembled a six-month rebuild plan in April 2024. The investments were:

1. A full Google Business Profile rebuild across all four locations, with weekly posts, monthly photo updates, and active Q&A management.
2. A review generation pipeline using Podium and ServiceTitan integration that automated review requests after every service call and tracked review velocity across Google, Yelp, BBB, and Angi.
3. Twenty-three new service-area pages covering the specific Phoenix metro neighborhoods they serve, each with 800 to 1,200 words of substantive prose, real technician photos, and accurate service descriptions.
4. License verification page updates on their site exposing their Arizona ROC license numbers in clean, machine-readable HTML, with direct links to the state license board verification pages.
5. A BBB accreditation upgrade from accredited to A+ rated with active dispute resolution.
6. A monthly local PR push targeting Phoenix-area home and lifestyle publications, generating three to five named mentions per month.

By Q4 2024, AI-search-driven calls were 8% of new customers. By Q2 2025, 16%. By Q4 2025, 31%. By Q1 2026, 44% of new customers were arriving through AI-search-driven direct calls. Combined marketplace spend was down to $13,000 per month — a 73% reduction — and overall call volume was up 38% from the 2024 baseline. Blended cost-per-acquired-customer dropped from $312 to $89.

The investment was real. The six-month rebuild cost approximately $140,000 in agency fees, software subscriptions, and internal time. The payback period was eight months. The compounding benefit is durable in ways that paid lead spend is not — the citation surfaces they built keep working without recurring per-lead fees, and they are the assets the next acquirer will pay a premium to inherit.

This pattern is repeating across the contractor base. The CMOs running similar rebuilds in plumbing, electrical, roofing, and HVAC report comparable arc — six to nine months of infrastructure investment, then a step change in inbound direct calls and a sustained reduction in marketplace dependence.

The Five Citation Surfaces That Actually Matter for Trades

The citation surfaces that drive home services AEO are different from the SaaS or publisher playbook. Across the 4,800-contractor ServiceTitan benchmark and our own analysis of AI citation patterns for home services queries, five surfaces account for nearly all of the variance in citation rate.

1. Google Business Profile. This is the dominant local entity signal, and it is not optional. Across the home services queries we tracked, AI assistants pull from GBP for ground-truth data on contractor name, address, service area, hours, photos, and recent posts approximately 84% of the time. A complete, claimed, actively maintained GBP is the table stakes of home services AEO. The contractors who win citation share treat GBP as a daily editorial product, not a one-time setup. Weekly posts, monthly photo updates, active Q&A responses, and rapid response to every review are the operational baseline.

2. Review velocity on the right platforms. Raw review count matters less than review velocity, platform diversity, and recency. AI assistants weight a contractor with 240 Google reviews and a steady recent cadence higher than a contractor with 800 Google reviews where the most recent is 14 months old. They also cross-reference across platforms — a contractor with strong Google reviews but no Yelp presence at all is treated as partially verified, while a contractor with reviews on Google, Yelp, BBB, and Angi is treated as well-grounded. The platforms that AI assistants treat as authoritative for trades are Google, Yelp, BBB, and Nextdoor, in roughly that order. Newer platforms like NiceJob and Birdeye contribute review data into the broader graph but are weighted less directly.

3. License and bonding verification. This is the surface that home services AEO requires that SaaS AEO does not. Every state has a contractor license board with a public database where homeowners can verify a contractor's license number, status, bond, and complaint history. AI assistants check those databases — directly when they can, and indirectly through state-level license aggregators when they cannot. A contractor with a clean, verifiable license record gets cited. A contractor whose license cannot be verified or has unresolved complaints is systematically downgraded in AI answers, often without the contractor knowing why their visibility dropped.

4. Service-area pages on the contractor's own website. This is the highest-ROI surface that contractors directly control. A serious service-area page architecture covers each city and substantial neighborhood the contractor serves, with each page containing substantive, locally specific content — real photos from jobs in that neighborhood, accurate descriptions of common service issues in that area, named references to local landmarks and conditions. The contractors winning citation share have 15 to 40 service-area pages, not three. The pages get cited in AI answers for hyperlocal queries because they are the closest match to the user's geographic intent.

5. Named third-party mentions. AI assistants build the entity representation of a contractor partly from how that contractor is mentioned in third-party content. Local news coverage, BBB profiles, trade association directories, Reddit threads in metro subreddits, and named mentions in home services publications all contribute. A contractor mentioned by name in five to ten third-party sources is treated as a more substantial entity than a contractor whose only online presence is their own website. This is where the local PR investment pays off — not for direct traffic but for entity reinforcement that compounds across AI citation decisions.

The Service-Area Page Playbook for Trades

Service-area pages are the single highest-ROI editorial investment a contractor can make in 2026, and they are systematically under-built across the industry. Most contractor websites have one Service Areas page that lists 15 cities with no substantive content. That page is functionally invisible to AI assistants. The contractors winning citation share have built 15 to 40 individual pages, each one targeted at a specific geographic intent.

The architecture that works has five elements.

A page per city or substantive neighborhood, not a list. Each city in your service area gets its own URL — typically /service-area/[city-name] or /[city]-hvac-repair. The page is treated as a first-class editorial product, not a SEO doorway page.

Substantive locally specific content. Each page is 800 to 1,500 words of real prose about serving that specific city. What kinds of homes are common, what the most frequent service calls are, what local conditions affect HVAC or plumbing work, named references to the neighborhoods within the city. The content should not be templated — search engines and AI models can detect templated city pages and discount them.

Real photos from jobs in that city. Photos with EXIF data showing the date and approximate location, captioned with descriptions of the work done. This is the single highest-impact differentiator between service-area pages that get cited and pages that do not.

Customer reviews specific to that area. Embedded or quoted reviews from customers in that city, with the city name visible. Pulling from the Google Business Profile API or your CRM to surface relevant local reviews is one of the higher-ROI integrations.

Clear service catalog and pricing transparency for that area. The specific services offered in that city, with pricing ranges where appropriate. Pricing transparency is increasingly weighted by AI assistants as a trust signal, and contractors who expose typical price ranges on their service-area pages are getting cited more often than competitors who hide everything behind a quote request.

The build cost for a full service-area page program is real — typically $300 to $800 per page if outsourced to a specialist agency, or 6 to 12 hours of internal editor time per page if built in house. For a contractor serving 25 cities, the total build is in the range of $7,500 to $20,000 of agency cost or 150 to 300 hours of internal time. The payback period from our benchmark data is six to twelve months for contractors serving metros where AI search adoption is high.

The CRM Integration Layer: ServiceTitan, Jobber, and Housecall Pro

The home services AEO playbook is most powerful when it is integrated with the operational CRM the contractor already uses. Three platforms dominate — ServiceTitan in the enterprise and mid-market, Jobber in the small-business segment, and Housecall Pro in the very small business segment. All three have released native AEO and AI search integration features in the last 18 months.

ServiceTitan's Marketing Pro suite released in 2025 includes automated review request workflows, GBP post syndication, and an AI citation tracking dashboard that monitors how the contractor appears across ChatGPT, Gemini, and Perplexity for the head-term queries in their service categories. The integration with the CRM means that review requests are sent within minutes of job completion, GBP posts are auto-generated from completed jobs with customer permission, and the citation dashboard cross-references AI search appearances with actual call volume to attribute revenue.

Jobber's review and reputation toolkit, deeply integrated with Podium, automates similar workflows for smaller contractors. The platform also exposes a service area management feature that syncs GBP service area boundaries with the contractor's website service-area pages and the local listing aggregators, ensuring consistency across the entity data layer that AI assistants ingest.

Housecall Pro added a Local Pro feature in late 2025 that handles GBP optimization, review automation, and license verification updates for contractors who do not have the bandwidth to manage these surfaces manually.

The integration that matters most for AEO is the connection between job completion and review velocity. Reviews requested within 30 minutes of job completion convert to actual posted reviews at roughly 3x the rate of reviews requested 24 hours later. The contractors with the highest review velocity have automated this workflow through their CRM, removing the human friction of the front-desk team remembering to ask. This single workflow change, properly implemented, has driven 4x and 5x increases in review posting rate for contractors who previously relied on manual asks.

Trust Signals: License, Bonding, BBB, and Insurance Verification

Home services AEO weights trust signals more heavily than any other AEO domain. A homeowner inviting a contractor into their home faces real safety, financial, and property risk, and AI assistants reflect that risk in their citation decisions. The trust surfaces that matter are concrete and verifiable.

State contractor license verification. Every state with a licensure requirement has a public lookup database. AI assistants check these databases directly or through aggregators when assessing whether to cite a contractor. The contractor's license number should be prominently displayed on the website in machine-readable HTML, with a direct link to the state verification page. A contractor whose license is expired, suspended, or under investigation will be systematically downranked or excluded from AI citation lists.

Bonding and insurance documentation. A contractor who exposes proof of general liability insurance, workers' compensation coverage, and surety bonding on their website is treated as a more substantial entity than a contractor who does not. The format that AI assistants extract well is a clean Trust or Credentials page with the policy carriers named, the bond number listed, and the documentation downloadable as PDF.

BBB accreditation and rating. The Better Business Bureau remains one of the most trusted third-party verification sources for home services in AI search. AI assistants weight BBB accreditation, rating grade, and unresolved complaint history heavily. According to the BBB's 2025 trust survey, 76% of homeowners check BBB before hiring a contractor for jobs over $1,000, and AI assistants mirror that behavior in their citation logic. Contractors should be accredited, maintain an A or A+ rating, and respond rapidly to any complaints filed.

Industry certifications. NATE certification for HVAC technicians, master plumber licenses, electrical journeyman cards, and trade association memberships (ACCA, PHCC, NRCA, IEC) all contribute to the trust surface. The contractor websites that surface these certifications cleanly — with the issuing body named and verification information exposed — get cited more often than contractors who treat certifications as ornamental.

Insurance carrier transparency. Naming the specific insurance carriers (State Farm, The Hartford, Travelers) the contractor uses for general liability is a small but measurable trust signal. AI assistants extract carrier names and use them as additional grounding data.

The total infrastructure investment to get trust signals to the level AI assistants reward is modest — typically a few thousand dollars in agency time plus the cost of any certification renewals. The ROI is significant because trust signals are weighted as a multiplier on top of other citation surfaces. A contractor with strong reviews but weak trust verification gets cited less than a contractor with adequate reviews and complete trust documentation.

Review Velocity, Authenticity, and the Yelp Trust & Safety Effect

Reviews are the second-most-weighted citation factor for home services after GBP completeness, but the dynamics in 2026 are different from the era when raw review count was the metric that mattered. AI assistants now weight three sub-factors: recency (how recent the most recent review is), velocity (how often new reviews are posted), and authenticity (whether the review pattern looks organic or manufactured).

Yelp's most recent Trust & Safety report released in 2026 showed that the platform's filtering system flagged 28% of submitted reviews as potentially fake or manufactured, with the home services category among the highest filter rates. AI assistants ingesting Yelp data weight the unfiltered reviews more heavily than filtered ones, and they cross-reference Yelp review patterns against the same contractor's Google and BBB review patterns to detect manufactured reviews. Contractors who buy reviews or run aggressive incentive programs typically show review pattern anomalies that AI assistants flag and discount.

The contractors who win on review velocity are doing five things consistently:

Automated review requests within 30 minutes of job completion. The review request goes out while the customer's satisfaction with the work is fresh. Manual requests sent the next day or later convert at much lower rates.

Multi-platform review requests. Different customers prefer different platforms. The request should include options for Google, Yelp, and BBB, with the customer choosing where to leave the review. Forcing all reviews to one platform is increasingly suboptimal because AI assistants cross-reference across platforms.

Real names and photos. Reviews from accounts with full names, profile photos, and review history on other businesses are weighted higher than reviews from anonymous-looking accounts. Contractors cannot directly control this, but the review request UX can encourage customers to use complete profiles.

Active management of response to reviews. Responses to both positive and negative reviews within 24 to 48 hours signal that the business is active and engaged. AI assistants extract response presence as a quality signal.

Specific work descriptions in review content. Reviews that describe the specific service performed — "Joe installed a new 80% efficiency furnace and was clear about the cost up front" — are weighted higher than generic praise. The review request workflow can subtly encourage specific work descriptions by including the job type in the email or SMS prompt.

The contractors with weak review velocity are typically running a manual ask process that depends on individual technicians remembering to request reviews. The contractors with strong velocity have automated the workflow through their CRM and consistently generate 8 to 25 new reviews per month per location, across multiple platforms.

What Killed Lead Marketplaces and What Comes Next

The Angi, Thumbtack, HomeAdvisor model was built on a specific consumer behavior — homeowners going to a marketplace, comparing multiple contractors, and choosing the cheapest or most responsive bid. AI assistants disintermediated that flow by offering the homeowner a pre-synthesized recommendation with two to four named contractors, eliminating the comparison shopping step entirely. The marketplaces did not lose because their economics broke. They lost because the user behavior they were built around stopped happening.

The pattern is consistent with what is happening across B2B services where consulting agencies and brokers are seeing similar AI-driven disintermediation. When the AI assistant can synthesize a recommendation directly, the comparison-aggregator middle layer collapses.

What replaces the marketplace is not a single new platform. It is a distributed infrastructure where contractors own their direct discovery pipeline through GBP, reviews, license verification, and service-area content, with AI assistants acting as the recommendation layer that connects homeowners to that infrastructure. The contractors winning this transition are the ones who built the infrastructure deliberately rather than waiting for the next marketplace to emerge.

There are second-order implications for the categories adjacent to home services. Lead aggregators in legal, financial, and medical services are running the same playbook as Angi and HomeAdvisor, and the same disintermediation is starting to affect them. The category that runs the AI-search rebuild fastest will be the one that captures the next decade of inbound discovery for the categories that follow.

This is also consistent with the broader shift documented in brand mentions becoming the new currency as backlinks decline. The links-based SEO regime that supported the marketplace model is giving way to an entity-and-mentions regime that rewards direct first-party citation surfaces. Home services is one of the categories where that shift is happening fastest and most visibly.

The 90-Day Home Services AEO Rollout

For a contractor running a $2M to $20M business in HVAC, plumbing, electrical, or general contracting, the 90-day rollout that delivers the highest leverage:

1. Audit your current AI citation rate. Run 30 head-term queries (furnace repair [city], plumber near me, HVAC company [neighborhood]) across ChatGPT, Gemini, and Perplexity. Document where you appear, where competitors appear, what is being cited. This baseline is the foundation for everything else.

2. Rebuild your Google Business Profile. Complete every field, add 30 to 50 recent photos, set up weekly posts, enable and respond to Q&A, and verify all service areas. This is the highest-leverage single investment in week one.

3. Implement automated review requests. Connect your CRM (ServiceTitan, Jobber, Housecall Pro) to a review platform (Podium, NiceJob, Birdeye) and configure within-30-minute review requests after every job. Target a 15% review request to posted review conversion rate.

4. Verify and expose your license status. Add a Credentials or Licensing page with your state license number, bond information, insurance carriers, and BBB accreditation. Link directly to the state license verification page.

5. Build the first ten service-area pages. Cover the ten cities or neighborhoods that drive the most revenue. Each page gets 800 to 1,200 words of substantive locally specific content, real photos from jobs in that area, and embedded local customer reviews.

6. Achieve BBB A+ rating if you do not already have it. Resolve any open complaints, respond to all historical complaints, and complete the accreditation upgrade if necessary.

7. Launch a local PR push. Three to five named mentions per month in local home and lifestyle publications, neighborhood Facebook groups, and metro subreddit threads. Use a fractional PR specialist if you do not have an internal team.

8. Instrument AI citation tracking. Subscribe to one of the AI citation monitoring tools (Profound, AthenaHQ, ContractorIntel) and set up weekly tracking of share-of-citation for your top 30 queries. Use the data to prioritize the next quarter of investment.

The first 90 days do not finish the work. They establish the foundation. The compounding gains happen in months four through twelve as the citation surfaces accumulate authority and review velocity builds momentum. Contractors who run the 90-day rollout consistently report meaningful inbound call shifts by month four and the major step change between months seven and ten.

For the contractors still relying on Angi, Thumbtack, and HomeAdvisor as the primary lead pipeline, the window to make the transition without losing meaningful revenue is closing. By Q4 2026, the contractors who built citation infrastructure in 2025 will have a defensive moat that newer competitors cannot easily breach. The trades that wait will spend the back half of the decade trying to catch up to defaults that have already hardened.

Takeaway: Home services AEO is an infrastructure rebuild, not a marketing campaign. The contractors winning in 2026 — and the ones positioned to keep winning through 2028 — are the ones who treated GBP, reviews, license verification, service-area pages, and named third-party mentions as five coordinated surfaces requiring deliberate, sustained investment. The marketplace model that defined the 2010s and early 2020s is being disintermediated by AI assistants that hand homeowners a pre-vetted shortlist instead of a comparison page. The economics favor the contractors who own their citation infrastructure. The contractors still buying $85 shared leads from Angi while AI search routes direct calls to their competitors are watching the pipeline shift in real time. The 90-day rollout pays back inside a year. The cost of waiting is measured in market share that does not come back.