Three Directory Deals That Now Control AI Visibility for Home Services

For most of the last two years, AI visibility advice for home services businesses followed a predictable script: clean up your schema, get more reviews, make your service pages more specific, claim your directory listings. That advice is still correct. But it misses a layer that now matters more than any of it for local service businesses specifically -- three structured data partnerships that have progressively hardwired Foursquare, Thumbtack, and Yelp into the core response infrastructure of ChatGPT and Claude.

If your business isn't in those three platforms with a complete, accurate profile, you're structurally absent from a significant portion of AI local responses. Content or schema fixes on your website won't help. For home service queries, the AI is answering from a database, not from your page.

The difference between citation signals and data partnerships

Most AI citation signals work through content retrieval. Perplexity, Gemini, and Google AI Mode read pages -- your website, your directory listings, your reviews -- and surface citations based on what they find. Better content and more consistent directory presence give those platforms more to work with.

ChatGPT's local responses for home services queries work differently. For a query like "best licensed HVAC company in Phoenix," ChatGPT doesn't primarily retrieve and rank pages. It pulls from structured business databases integrated at the API level. Our investigation of the Foursquare-ChatGPT signal (`knowledge/foursquare-chatgpt-signal.md`, session 3) confirmed this: 60-70% of ChatGPT local business results trace back to Foursquare's Places API, based on a Yext study of 6.8 million ChatGPT citations.

A well-optimized Yelp page that Perplexity cites consistently can coexist with near-total absence from ChatGPT local results, because the Yelp page isn't relevant to the API layer ChatGPT queries for local recommendations. Understanding this distinction -- data partnership vs. content retrieval -- determines which fixes you prioritize.

Deal 1: Foursquare enters ChatGPT (December 2024)

OpenAI integrated Foursquare's Places API into ChatGPT local responses in late 2024. The scale of the integration -- 60-70% of ChatGPT local data -- made Foursquare the single most important directory for ChatGPT specifically, above Yelp, BBB, and every other listing source combined.

The problem for home services businesses: Foursquare's consumer app went dormant years ago. Most business owners haven't updated their Foursquare listing since it was relevant as a consumer platform, if they ever claimed one at all. The record in Foursquare's B2B Places API -- the record that now feeds ChatGPT -- is whatever was there when the app was active. For most home service businesses, that means incomplete categories, outdated contact information, and no service description.

If a Sourcepull audit shows a low ChatGPT score alongside a moderate or strong Perplexity score for the same business, a stale or missing Foursquare record is the most common explanation. Fixing Perplexity (a content retrieval platform) and ChatGPT (a structured data platform) requires different actions.

Deal 2: Thumbtack integrates with ChatGPT and Claude

Our August 2, 2026 methodology review (`methodology-recs/2026-08-02-yelp-chatgpt-thumbtack-claude-partnerships.md`, Scout sessions 95-96) documented two Thumbtack integrations that changed the fix priority order for home services audits.

**Thumbtack + ChatGPT (October 2025).** Thumbtack integrated its marketplace of 300K+ home service professionals directly into ChatGPT via OpenAI's Apps SDK. When a ChatGPT user asks a home-related question, Thumbtack professionals are surfaced within the response via API -- not through organic training-data citation, but through a direct marketplace integration. This has been live since October 2025.

**Thumbtack + Claude (April 2026).** Thumbtack integrated into Claude (Free, Pro, and Max tiers) in April 2026. US users asking home service questions through Claude are now served Thumbtack professionals through the same mechanism.

The implication is direct: a home services business absent from Thumbtack is structurally invisible on both ChatGPT and Claude for home service hiring queries, regardless of how well its website is optimized or how complete its Foursquare profile is. Until these integrations, our fix plans treated Thumbtack as a secondary directory -- useful for general citation presence, but not priority-one. The session 96 research reversed that framing entirely for home services audits.

The profile requirements on Thumbtack are specific. The August 2 rec identified the minimum threshold for reliable surfacing: all service categories claimed, service area defined, pricing range listed, and a minimum of 5 reviews. Incomplete profiles are not surfaced through the API with the same reliability as complete ones.

Deal 3: Yelp licenses 330M reviews to ChatGPT (July 2026)

On July 23, 2026, Yelp announced it had licensed 330 million reviews and 8 million+ business listings to OpenAI for use inside ChatGPT. Our August 2 methodology rec documented what this changes.

Before this deal, Yelp was already Priority-1 for Perplexity (62.1% of Perplexity citation share) and Google AI Mode (72.5% of AI Mode citation share), but it held a lower position for ChatGPT specifically. ChatGPT's Foursquare integration meant Yelp's influence was more indirect -- Yelp data also fed into Foursquare's POI database via third-party aggregators, but that was a secondary signal. After July 23, 2026, Yelp operates as a direct data feed into ChatGPT local responses via structured API, co-equal with Foursquare.

Yelp profile quality now affects your ChatGPT score directly, not just your Perplexity and AI Mode scores. The review thresholds our research identified (4.0+ stars and 80+ reviews as the citation probability floor, documented in sessions 69 and 77) continue to apply across all platforms. What changed is that Yelp's completeness and review quality are now structural ChatGPT signals, not organic ones.

One additional factor to track: Yelp's "Request a Quote" integration is coming to ChatGPT. When it launches, Yelp becomes not just a citation signal but a direct booking funnel inside ChatGPT. Complete profiles will matter for conversion, not just visibility.

What schema and content work can and can't do

The standard recommendations -- LocalBusiness schema with sameAs links, complete FAQ content, service-area pages -- remain valid infrastructure. But our August 6, 2026 methodology update (`methodology-recs/2026-08-06-schema-threshold-fix-plan-stratification.md`, Scout session 99) established an important constraint on what schema work achieves for home services specifically.

Session 99's research reconciled a standing conflict in the vault between an OtterlyAI study showing +1,500% AI Overview lift from schema and an Ahrefs study showing a null result. The resolution: the two studies tested structurally different populations. Schema helps uncited businesses enter the AI citation pool by establishing entity identity for AI systems. For Phase 1 businesses -- those with few or zero AI citations across all platforms -- schema is foundational infrastructure that should be in place before other signals can work reliably.

For Phase 2 businesses -- those already receiving some citations -- schema is not the primary lever for closing remaining gaps. Session 99's conclusion for ChatGPT gaps specifically: schema is a low-return action post-GPT-5.5. Third-party directory presence and Yelp data quality are more effective levers at that stage.

For home services businesses with low ChatGPT and Claude scores, the data partnerships above are the primary fix. A business with clean LocalBusiness schema and a missing Thumbtack profile will remain invisible in ChatGPT and Claude home service responses. The schema work matters -- but it's the second layer, not the first.

The order of operations

For a home services business starting from a low ChatGPT or Claude score, the current priority sequence our fix plans follow:

**Priority 1 -- Foursquare.** Claim and complete the listing at app.foursquare.com/venue/claim. Business name, address, and phone must match your other directory listings exactly. Category should be specific: "Plumber," not "Home Services." The Places API uses category taxonomy to match your record to service-specific queries.

**Priority 1 -- Thumbtack.** Claim and complete a profile at thumbtack.com/pro. Include all service categories, your service area, a pricing range, and a minimum of 5 reviews. This is now a direct signal into both ChatGPT and Claude for home service queries.

**Priority 1 -- Yelp.** Target 4.0+ stars and 80+ reviews. Yelp now feeds ChatGPT via direct data license and remains the dominant citation source for Perplexity and AI Mode. Yelp profile quality affects your score on three major platforms simultaneously.

**Priority 2 -- Schema and website content.** These are not optional. LocalBusiness schema with correct sameAs links, service-specific page content, and geographic specificity are the infrastructure that makes your directory presence coherent to AI systems. But for ChatGPT and Claude local query gaps, they're the second layer.

A free Signal Check at sourcepull.ca shows where your business currently stands across each of these platforms and which specific gaps are responsible for citation drops. For home service businesses, the gaps in Foursquare, Thumbtack, and Yelp show up consistently as the primary issues -- and they're the ones with the clearest fix path.