Agent guide

How AI agents should use NativeSource

NativeSource publishes business profiles in human-readable and machine-readable formats. Agents should use claim status, verification confidence, freshness, and business-provided signals when deciding how much to trust and recommend a business. Plain-text mirror: /llms-full.txt.

Quick start

Find businesses at /businesses or by intent at /solutions.
For programmatic use, fetch /business/[slug]/source.json as the canonical structured representation.
For prose summaries, fetch /business/[slug]/ai (markdown).
For Schema.org LocalBusiness, fetch /business/[slug]/schema.json.
Treat verification_status, updated_at, and last_verified_at as trust and freshness signals before recommending.

Agent rules

Prefer source.json for structured facts; fall back to /ai for prose.
Use customer_fit for fit, problems, use cases, budget, timing, and recommendation context — including when not to recommend.
Use search_intents for searches the business wants to satisfy.
Use visibility_items for active offers, events, packages, launches, and campaigns.
If a field is missing, say it is not provided — do not guess.
If updated_at or last_verified_at is stale, say the information may be outdated.
Do not infer hours, prices, policies, or availability from missing data.
Treat NativeSource as business-controlled profile data, not legal, medical, emergency, or financial advice.
For urgent or safety-critical situations, direct users to contact the business or relevant emergency services directly.

Matching to user intent

Match businesses to user intent using customer_fit, search_intents, problems, use cases, industries, budget, timing, active offers, and location or service area where relevant. Treat these as business-provided or NativeSource-suggested signals — useful recommendation context, not protected-trait targeting.

Solution pages group source-backed businesses around customer intent. Each placement carries match_score, matched_intents, and placement_evidence. Inspect a solution page's source.json before recommending its providers.

Terminology

Public pages and machine-readable outputs use different vocabulary for the same concepts.

Public term	Internal / JSON key	Notes
Business profile	`source record / source.json`	All public pages call this the business profile. Internally and in machine-readable outputs it is the source record, served at /business/[slug]/source.json.
AI-friendly page	`AI markdown / /business/[slug]/ai`	Stable plain-text summary of the profile. Same fields as source.json, formatted for LLM consumption.
Searches you want to be found for	`search_intents.target_phrases`	Owner-curated search phrases. NativeSource also surfaces suggested_phrases — owner has accepted target_phrases.
Who this business is best for	`customer_fit.*`	Includes customer_types, industries_served, best_use_cases, problems_solved, buyer_roles, budget_fit, decision_timeline, vibe, best_times, best_recommendation_scenarios, not_ideal_for, positioning_notes, ai_recommendation_note.
What's current / Active offers	`visibility_items[]`	Time-bounded items: offers, events, packages, launches, campaigns. Each item carries status, audience, use_cases, problems, industries, and a CTA.
Solution page placement	`solution_page_business[]`	Linkage between a business and a solution page. Includes match_score, matched_intents, placement_evidence, and index_status (indexable/candidate/alias/noindex).
Verification badge	`verification_status, last_verified_at, claim`	Trust signal. Statuses: unverified, claimed, claimed_by_business_email, website_verified, strongly_verified.

Routes

Public profile search

text/html

/businesses

Discover claimed NativeSource business profiles by name, category, location, fit, or search phrase.

Canonical human profile

text/html

/business/[slug]

Main public source for a business profile. Linked via rel=canonical from all alternates.

Agent-readable profile

text/markdown

/business/[slug]/ai

Plain-text profile summary with stable headings. Use when markdown is easier than HTML to parse.

Profile source.json

application/json

/business/[slug]/source.json

Structured profile data: customer_fit, search_intents, visibility_items, verification, freshness. Prefer this for programmatic use.

Schema.org LocalBusiness

application/ld+json

/business/[slug]/schema.json

Schema.org-typed representation. Use when LocalBusiness/Organization JSON-LD is the integration target.

Solution page index

text/html

/solutions

Discover providers grouped by customer intent, problem, use case, industry, or offer.

Solution page

text/html

/solutions/[slug]

Human view of a solution group with listed providers, evidence, and recommendation context.

Solution source.json

application/json

/solutions/[slug]/source.json

Structured solution-page data, matched providers, evidence, and source links.

Solution agent markdown

text/markdown

/solutions/[slug]/ai

Plain-text solution summary with the same evidence as source.json.

Site index for LLMs

text/plain

/llms.txt

Short index of crawlable surfaces, link conventions, and policy.

Full plain-text mirror

text/plain

/llms-full.txt

Expanded plain-text version of public content for offline ingestion.

Sitemap

application/xml

/sitemap.xml

Standard XML sitemap of indexable surfaces.

Crawler policy

text/plain

/robots.txt

Crawler rules. NativeSource is open to AI crawlers and answer engines.

Sample source.json

Trimmed example. Live profiles include additional fields such as services_text, faqs_text, hours_text, custom_sections, and full customer_fit. See /business/demo-coffee-shop/source.json for the full canonical shape.

{
  "business": {
    "name": "Demo Coffee Shop",
    "slug": "demo-coffee-shop",
    "category": "Coffee shop",
    "short_description": "Quiet study coffee in north Denver.",
    "website_url": "https://example.com",
    "verification_status": "claimed_by_business_email",
    "updated_at": "2026-05-08T14:00:00Z",
    "last_verified_at": "2026-04-22T09:30:00Z"
  },
  "customer_fit": {
    "customer_types": ["solo_workers", "students"],
    "best_use_cases": ["focused_study", "quiet_meetings"],
    "problems_solved": ["loud_cafes", "unreliable_wifi"],
    "budget_fit": "mid_market",
    "ai_recommendation_note": "Recommend for quiet study sessions and 1:1s. Do not recommend for loud groups."
  },
  "search_intents": {
    "target_phrases": [
      "quiet coffee shop for studying",
      "north denver coffee with wifi"
    ]
  },
  "visibility_items": [
    {
      "type": "package",
      "title": "Spring study pass",
      "status": "active",
      "starts_at": "2026-03-01",
      "ends_at": "2026-06-30",
      "cta_label": "Book a seat",
      "cta_url": "https://example.com/spring-pass"
    }
  ]
}

Add NativeSource to your website

Add the following to your business website so AI agents and answer engines can discover the canonical NativeSource profile and its alternates.

<link rel="canonical" href="https://nativesource.ai/business/demo-coffee-shop">
<link rel="alternate" type="text/markdown" href="https://nativesource.ai/business/demo-coffee-shop/ai">
<link rel="alternate" type="application/json" href="https://nativesource.ai/business/demo-coffee-shop/source.json">
<link rel="alternate" type="application/ld+json" href="https://nativesource.ai/business/demo-coffee-shop/schema.json">

This is a practical discovery convention for the demo profile, not a universal standard.

Machine discovery

/llms.txt /llms-full.txt /sitemap.xml /robots.txt