# noisy-kiwi Project Context

Source: Loomi MCP discovery on 2026-05-31.

## Project Identity

- Project name: noisy-kiwi
- Project ID: `97b71c7a-5469-11f1-aaeb-9ee3018bf36e`
- Category: sandbox
- Workspace: Hackathon Workspace
- Organization: Hackathon Org
- URL: `https://uqa.app.exponea.dev/p/noisy-kiwi`

## Current State

- Existing scenarios: none returned by Loomi MCP.
- Existing customer filters: none returned by Loomi MCP.
- Email campaigns: none returned by Loomi MCP.
- SMS/MMS campaigns: none returned by Loomi MCP.
- Initiatives: none returned by Loomi MCP.
- Project variables: none returned by Loomi MCP.

## Customer Identity

Customer identifiers configured in the project:

- `registered` - hard identifier
- `cookie` - soft identifier
- `google_analytics` - soft identifier

## Project Shape

The tracked event schema is ecommerce-oriented and includes browsing, cart, checkout, purchase, returns, loyalty, support, campaign, banner, session, consent, retargeting, survey, call-center, and store-visit events.

The project currently looks like a sandbox data project with schema coverage but little or no configured CRM execution layer. Agents should treat generation work as greenfield unless a later Loomi MCP discovery finds assets.

## Important Gaps

- Standard ecommerce mapping is not configured in Bloomreach mapping: purchase, purchase item, cart, view item, category view, checkout, and related standard fields are all null in `get_mapping`.
- Customer standard fields such as first name, last name, gender, email, and phone are also unmapped in `get_mapping`, even though similarly named customer properties exist.
- Consent category mappings are null.
- Catalog mappings are null.
- No existing scenarios, campaigns, customer filters, initiatives, or project variables were found.

## Agent Guidance

- Use exact event and property names from [schema.md](./schema.md) or the JSON schema files.
- Confirm mapping configuration before relying on Bloomreach standard ecommerce helpers.
- Prefer explicit event/property references over standard mapped fields until mapping is configured.
- For revenue reporting, start with the `purchase` event and `total_price` property, then confirm whether `purchase_status`, currency, returns, and tax rules should filter or adjust the metric.
- For lifecycle generation, use `view_item`, `cart_update`, `checkout`, `purchase`, `purchase_item`, `return`, `campaign`, `banner`, and `consent` as the primary available signals.
- Because no existing CRM assets were found, recommend creating reusable filters, scenarios, and project variables before building complex journeys.