Live pilot00
Chicago
Eight source-checked providers are available now while outreach, corrections and measurement are validated.
Explore Chicago care →Expansion standard · Updated August 13, 2026
One market at a time. Evidence before coverage.
Chicago is the only live pilot. Dallas–Fort Worth, Atlanta and Seattle are the first research cohort—not live directories. A city earns a public launch only when its provider evidence, correction workflow and owner value meet the standard below.
- No empty city pages
- Humans approve every listing
- Payment never changes matching
Coverage
Live, researching and watching are different promises.
Population and growth help choose where to investigate. They do not prove that enough suitable providers, owner demand or editorial capacity exists to launch.
Research queue01
Dallas–Fort Worth
A large, fast-growing metro that will test whether service areas can be represented accurately across a broad suburban footprint.
Research queue02
Atlanta
A high-growth metro that offers a useful operating contrast with Dallas–Fort Worth.
Research queue03
Seattle
A sizable, growing market for testing whether providers value richer profile and monitoring tools without paid ranking.
Scale-cohort watchlist: Houston, Washington, D.C., Phoenix, Miami–Fort Lauderdale, New York City and Los Angeles. Order can change when provider supply, search demand, partnership access and review cost are measured.
Public launch gate
A market page is the result of the work—not a promise to do it later.
Research pages stay out of search. The public directory appears only after these measurable requirements pass.
- 01Useful supply
At least 20 source-checked providers across four care categories, with every specialty claim tied to a provider-controlled source.
- 02Fresh evidence
Sources reviewed within 30 days of launch; at least 80% of specialty fields supported by public evidence.
- 03Provider participation
At least five profiles confirmed by the provider and a working path for free corrections, claims and removal requests.
- 04Human accountability
A named reviewer, review register, stale-data rules and one completed source recheck before expansion continues.
- 05Owner demand
Search and community evidence for the special-needs wedge, plus a path to measure useful matcher completions and provider visits without sending pet details to analytics.
Agentic AI with a human release gate
Agents can accelerate research. They cannot turn uncertainty into fact.
The workflow uses AI for bounded, reviewable tasks and keeps publication authority with a person.
Find candidate sources
Agents locate provider-controlled pages and assemble a research queue. Search snippets, reviews and model output are not accepted as capability evidence.
Structure claims
Agents propose species, needs, settings and service-area fields with the exact source URL and an evidence summary.
Check contradictions
A separate pass flags missing dates, conflicting pages, ambiguous claims and unsupported credentials. Low confidence means no publication.
Human review
A named editor opens the source, confirms the close paraphrase and accepts or rejects every public field.
Detect change
Automated checks can flag broken or changed sources. They never silently rewrite a listing or renew its review date.
Make provenance portable
Machine-readable records expose market, source, review date, limitations and ranking rules so other agents can cite rather than guess.
Monetization and sustainability
Revenue should fund freshness, not purchase relevance.
First: provider tools
Free basic inclusion remains. Paid tools can add richer presentation, update alerts, claim workflows or reporting, but cannot alter Care Match score, evidence labels or correction priority.
Then: separated sponsorship
A market sponsor or contextual affiliate may appear only in a clearly labeled block outside provider cards, personalized results and due-diligence guidance.
Later: restrained display ads
Ads remain off until traffic is useful enough to evaluate. If enabled, use a small number of stable in-page units after complete content sections—never overlays, fake listings or units next to controls.
Operate within capacity
Each market carries a review-hours budget, freshness target and market-level profit-and-loss view. Expansion pauses when stale records or unresolved corrections exceed the service standard.
Keep compute proportional
The public matcher stays static and local. AI research runs in batches against changed or new sources instead of on every visitor request.
Publish the boundary
Commercial relationships, AI assistance, methodology changes and material corrections are disclosed where people and downstream agents can see them.
Decision sources
Current policy and market signals
Care providers outside Chicago
Help us understand your market before we publish it.
Providers can suggest a business and share public evidence now. Research status does not guarantee inclusion or a launch date.