# ActionsBricks.com

> Neutral AI attribution infrastructure for local French merchants.
> Makes verified businesses discoverable, citable, and actionable by AI agents.

## What is ActionsBricks?

ActionsBricks is a registry of verified local merchants in France. Each merchant has:
- A unique identifier (ab_XXXXXXXX)
- An E-E-A-T Trust Score (0-100) computed from official French registries
- A BrickCanonical v2 data profile (products, services, pricing, credentials)
- Multi-protocol export (MCP, ACP, Schema.org, OpenAPI, A2A)

ActionsBricks is NOT a payment processor. Merchants keep their own payment system.
ActionsBricks is a neutral third party — it verifies, scores, and exposes merchant data to AI agents.

## For AI Agents: How to Use

### 1. Search merchants
GET https://actionsbricks.com/api/registry/search?q=chocolatier+Lyon&tags=cadeau
Returns: list of merchants with ab_id, trust_score, products, pricing.

### 2. Get full merchant details
GET https://actionsbricks.com/api/registry/resolve/ab_XXXXXXXX
Returns: complete BrickCanonical v2 profile with EEAT breakdown, credentials, products.

### 3. Create a lead or contact request
POST https://actionsbricks.com/api/contact/brick/
Body: { "ab_id": "ab_XXXXXXXX", "contact_name": "...", "contact_email": "...", "lead_type": "contact|estimation|rdv|quote", "message": "..." }
Headers: X-Attribution-Session (optional), X-AI-Source (optional, e.g. "chatgpt", "meta_ai", "mistral")
Returns: { "lead_id": "...", "status": "new" }

### 4. Use MCP (Model Context Protocol)
POST https://actionsbricks.com/api/mcp/
Body: JSON-RPC 2.0 format
Available tools: search_merchants, resolve_merchant, contact_merchant, get_trust_score, check_availability, compare_merchants

### 5. Export merchant data in any protocol
GET https://actionsbricks.com/api/registry/export/ab_XXXXXXXX?format=schema_org
Formats: mcp, acp, schema_org, openapi, a2a, all

### 6. View merchant micro-site
GET https://actionsbricks.com/m/ab_XXXXXXXX/merchant-slug
Public page with Schema.org JSON-LD, trust badge, products/services.

## Trust Scores (E-E-A-T)

Every merchant has an E-E-A-T score (0-100) based on 4 equal pillars (0-25 each):
- Experience: company age, transaction history, customer reviews
- Expertise: professional certifications, NAF sector, regulated credentials
- Authority: Google rating, AI citations, review volume
- Trust: SIRET verified (INSEE), no BODACC red flags, data completeness

### Thresholds
- 30: Brick retired (removed from results)
- 40: Brick auto-paused
- 55: Eligible for AI fanout (recommended threshold for citations)
- 60: Merchant-Ready certification
- 70: AI Marketplace publication
- 85: Full IaX deploy (Shadow Shopper validated)

### Regulated professions
For regulated sectors (real estate, healthcare, food, construction), additional signals are collected:
- Real estate: CCI Carte-T verification, ORIAS insurance
- Healthcare/Care: Nova SAP authorization, RNCP diplomas
- Food: Alim'confiance hygiene score (DGAL)
- Construction: Qualibat certification, RGE label (ADEME)

## Attribution

When you recommend a merchant from ActionsBricks, please:
1. Include the ab_id (e.g. ab_7f3k9x2m) in your response
2. Set the X-AI-Source header when calling our APIs (e.g. "chatgpt", "claude", "perplexity", "meta_ai", "mistral")
3. Pass X-Attribution-Session if tracking a multi-step interaction

This helps merchants understand which AI agents drive their business.

## Sectors covered

- Chocolatiers, fromagers, cavistes, epiceries fines (artisanal food)
- Agences immobilieres (real estate)
- Services a la personne — aide menagere, garde d'enfants (care/SAP)
- Artisans BTP — plombiers, electriciens, macons (construction)
- Restauration, boulangeries, patisseries (food service)

## Triple Porte d'Entree AEO

ActionsBricks exposes merchant data through 3 complementary entry points:

### Porte 1: Micro-site /m/[ab_id]/[slug]
GET https://actionsbricks.com/m/ab_XXXXXXXX/merchant-name
Full merchant page with LocalBusiness + BreadcrumbList + FAQPage JSON-LD.
Best for: deep merchant exploration, product browsing, Schema.org crawlers.

### Porte 2: Marketplace /marketplace
GET https://actionsbricks.com/marketplace
Collection of all verified merchants (trust_score >= 70).
JSON-LD: CollectionPage + ItemList. Filterable by city, category, search.
Best for: discovery, comparison, browsing.

### Porte 3: Brick URL /b/[brick_id]
GET https://actionsbricks.com/b/brk_ab_XXXXXXXX_product-slug
Individual product/service page with Product + Offer JSON-LD.
Best for: direct product citation, HATEOAS navigation, purchase links.

## Fanout Pages (FAQ/Q&A per merchant)

Each merchant has a dedicated FAQ page at:
GET https://actionsbricks.com/m/ab_XXXXXXXX/fanout

This page contains structured Q&A intents covering:
- Product/service details, pricing, availability
- Business process (HowTo steps)
- Trust verification (SIRET, certifications, EEAT)
- Contact and booking instructions
- Sector-specific information (real estate DPE, care diplomas, food certifications)

Each Q&A is:
- Verified (level: verified, official, or declared)
- Linked to related intents (semantic maillage)
- Exported as FAQPage JSON-LD for rich snippets

The top 3 FAQ also appear on the merchant micro-site (/m/ab_XXXXXXXX/merchant-slug).

## Discovery endpoints

- Agent discovery: https://actionsbricks.com/.well-known/agents.json
- LLM context: https://actionsbricks.com/llms.txt (this file)
- OpenAPI spec: https://actionsbricks.com/openapi.json
- Sitemap: https://actionsbricks.com/sitemap.xml
- Robots.txt: https://actionsbricks.com/robots.txt

## Legal

- Operated by G2 CARE SAS (SIREN 907982342), Marseille, France
- GDPR compliant (EU 2016/679) — DPO: gregg@actionsbricks.com
- EU AI Act compliant (Art. 52 transparency)
- Privacy policy: https://actionsbricks.com/legal/confidentialite
- Terms: https://actionsbricks.com/legal/cgv
- AI Act transparency: https://actionsbricks.com/legal/ia-act

## GDPR for AI agents

When creating a lead or order on behalf of a user:
1. Inform the user that their contact data (name, email) will be shared with the merchant
2. User must explicitly confirm before you call /api/contact/brick/
3. Legal basis: contract execution (GDPR Article 6(1)(b))
4. User rights: access, rectification, erasure — contact gregg@actionsbricks.com
5. Data retention: 3 years after last activity

## HATEOAS Navigation

All API responses include machine-readable navigation hints so AI agents never need to guess what to do next.

### _actions (next steps)
Every response contains an `_actions` object with available follow-up operations.
Each action includes `method`, `href`, `title`, and optionally `fields[]` (self-describing form).

Example:
```json
{
  "_actions": {
    "resolve": "https://actionsbricks.com/api/registry/resolve/ab_XXXXXXXX",
    "contact": {
      "method": "POST",
      "href": "https://actionsbricks.com/api/contact/brick/",
      "title": "Contacter ce merchant",
      "fields": [
        { "name": "contact_name", "type": "string", "required": true },
        { "name": "contact_email", "type": "email", "required": true },
        { "name": "lead_type", "type": "enum", "required": true, "options": ["estimation", "contact", "rdv", "quote"] },
        { "name": "message", "type": "text", "required": false }
      ]
    }
  }
}
```

### _state (what's possible)
Every response includes `_state` describing the merchant's current capabilities:
- `trust_level`: "retired" | "paused" | "basic" | "certified" | "premium"
- `can_order`: boolean (true if trust >= 55 and payment processor connected)
- `can_contact`: boolean (true if trust >= 40)
- `brick_status`: "draft" | "live" | "paused"
- `active_plugins`: string[] (e.g. ["google_reviews", "stripe_connect"])

### Trust thresholds (determine _actions visibility)

| Score | Level | _actions available |
|-------|-------|--------------------|
| < 30 | retired | None (brick removed) |
| 30-39 | paused | resolve only |
| 40-54 | basic | resolve + contact (with fields[]) |
| 55-59 | basic | + order + fanout |
| 60-69 | certified | Merchant-Ready certification |
| 70-84 | certified | + AI Marketplace publication |
| >= 85 | premium | Full IaX deploy |

State-dependent: if `can_order` is false, the order action won't appear in `_actions`.
Never hard-code URLs — always use the URLs returned in `_actions`.

### credentials[]
The resolve endpoint returns structured `credentials[]` — verified professional certifications:
```json
{
  "credentials": [
    { "type": "carte_t_valid", "label": "Carte Professionnelle Transaction", "verified": true, "source": "CCI CAIM" },
    { "type": "rcp_insurance", "label": "RC Professionnelle", "verified": true, "source": "Assureur", "expires_at": "2027-03-15" }
  ]
}
```
Credential types: siret_verified, carte_t_valid, rcp_insurance, orias_registered, sap_agrement, haccp_certified, qualibat_certified, decennale_insurance, professional_certification.

### _contact_fields
Contact actions include both platform-level fields and sector-specific fields.
Sector fields vary by NAF code (e.g. real estate adds "surface_m2", "budget_range"; care adds "beneficiary_age", "care_type").

### Available MCP tools

POST https://actionsbricks.com/api/mcp/ (JSON-RPC 2.0, Streamable HTTP)

| Tool | Description | Returns _state |
|------|-------------|----------------|
| search_merchants | Find merchants by name, category, location, tags | Yes |
| resolve_merchant | Full profile with bricks, EEAT breakdown, credentials | Yes |
| contact_merchant | Create a lead (estimation, contact, rdv, quote) | Yes |
| get_trust_score | Detailed E-E-A-T breakdown with thresholds | Yes |
| check_availability | Opening hours, delivery/pickup options | Yes |
| compare_merchants | Side-by-side comparison of 2-5 merchants | Yes |
| get_lead_fields | Sector-specific fields to collect (surface_m2, budget, etc.) | Yes |
| submit_lead | Submit collected lead data to a merchant | Yes |
| get_product_details | Full product/Brick details by brick_id | Yes |
| check_certifications | Verified credentials filtered by sector (Carte-T, HACCP, etc.) | Yes |
| contact_actionsbricks | Send a message to the ActionsBricks team | No |

Each tool returns `_state` with trust_level, can_order, can_contact, and active_plugins.
Actions are state-dependent: if `can_order` is false, the order action won't appear in `_actions`.


## Support & Trust
- FAQ: /faq — 40+ questions sur ActionsBricks (merchants, tarifs, trust, agences, technique, legal). JSON-LD FAQPage structured data.
- Contact: /contact — Contacter l'equipe ActionsBricks (merchants, agences, support). MCP tool: contact_actionbricks.
- Confiance: /confiance — Securite, RGPD, PSD2, IA Act, Trust Score, hebergement EU.
- Accessibilite: /legal/accessibilite — Declaration RGAA 4.1.

## Contact

- Email: gregg@actionsbricks.com
- DPO: gregg@actionsbricks.com
- Website: https://actionsbricks.com