A catalogue of everything {{product.name}} does. Every entry follows the same structure —
what it does, how it works, how it's built and
how it's linked — so you can read across the product consistently. Filter by domain or search,
and click any entry to expand it.
⚠
Entries below are examples. They illustrate the format for a generic product. Replace them with
{{product.name}}'s real features (one entry per user-visible capability), adjust the filter categories to suit, and
keep the four-part structure. Remove this callout when done.
Core capabilities
The primary value of {{product.name}}. [Replace: one-sentence framing.]
What it does
[Replace: the user outcome in one or two sentences. Lead with the benefit, then the mechanism.]
How it works
[Replace: the main flow. For example: the user submits input, the service validates it, processes it asynchronously and notifies on completion.]
How it's built
Implemented in the core service; key code under [replace/path/]; configuration keys [REPLACE_KEY].
How it's linked
Depends on [Replace: component]; feeds [Replace: feature]. Switch: features.[replace].
What it does
Brings the customer's existing data into {{product.name}} from files or connected systems, with validation feedback.
How it works
Upload or connect a source, map fields, preview errors, then run. Re-runs are idempotent.
Prerequisite for most other features. Source connectors are swappable — see Integrations.
What it does
Lets users find items by keyword or natural language, with filters and saved searches.
How it works
Queries hit the search index; results are permission-trimmed per user before return.
How it's built
Search index; ranking configuration in [replace/path].
How it's linked
Reads content written by the core service; respects tenant boundaries.
Insight & AI
Analytics and AI-assisted features. [Replace: one-sentence framing, and state any human-review requirements.]
What it does
Answers questions and drafts content grounded in the tenant's own data, with citations.
How it works
A request is enriched with retrieved context, sent to the model provider through the AI service, filtered by guardrails and returned with sources. Users can accept, edit or reject.