CLARA

What it does

An AI-enabled product intelligence tool that produces research artefacts across the Research, Design, and Test phases of the ProductOps pipeline. CLARA reads your programme's knowledge base, drafts structured artefacts, and files them back — so each team starts richer than the last.

CLARA connects to more than one knowledge-base backend: Confluence (via a direct connector, no manual setup) and Plane (via an MCP server). She owns the research artefacts that live in the Knowledge Base across Research, Design, and Test (personas, journeys, PRDs, capability storyboards, test plans, and the rest). Prototype scaffolding still goes through Claude Code, not CLARA.

How CLARA behaves

She is disciplined about evidence — every finding she surfaces cites the source Confluence page that supports it. If the corpus is silent on something, she says so plainly rather than extrapolating.

She is cautious about fabrication — when inputs are thin, she flags what's missing before drafting rather than producing something that looks complete but rests on guesses.

She is strict about filing — she verifies every level of the target hierarchy exists before creating a page. If she can't create the full path, she stops and reports rather than silently filing somewhere else.

She is track-aware — she always knows whether she's working at programme-wide or track-level scope, and her downstream prompts cascade automatically (reading track-level material first, falling back to programme-wide when no track-level version exists).

Where outputs land

CLARA files every artefact under a top-level Knowledge Base page in the programme's knowledge base:

Knowledge Base /
  Programme-wide /          cross-track artefacts
    Personas (Programme-wide) /
    PRDs (Programme-wide) /
    ...
  <your track> /            track-specific artefacts
    Personas (<your track>) /
    Journeys (<your track>) /
    PRDs (<your track>) /
    ...

Track is your programme's own organising slice — workstream, capability area, feature line, sub-system — whichever applies. Artefacts that span tracks file under Programme-wide.

Provisioning the Knowledge Base

CLARA creates any missing parts of the hierarchy automatically on first use, so teams can start drafting without setup. Two optional invocations are recommended when you want clean structure upfront:

• New programme — use clara's setup-kb for [programme]. CLARA walks through programme type and current track list, previews the full hierarchy, and creates it in one pass. Safe to re-run.
• New track on an existing programme — use clara's add-track [track] to [programme]. Bolts one track's folders onto an already-provisioned KB without re-walking the full setup.