Skip to main content

Frameworks

A Framework is a structured assessment that lives inside a practice. It encodes a methodology — for example an "Operating Model Assessment" or a cybersecurity maturity model — as a tree you run against a client's evidence in Diagnose.

The structure of a framework

  • Dimensions — the top-level areas of the methodology.
  • Assessment Areas (subsections) — groups of related questions within a dimension.
  • Questions — the individual things assessed, each with an answer type.

Where frameworks come from

Frameworks reach your firm two ways:

  • Published by Celeredge — a curated library available to all firms or a subset (you'll see a global or shared source badge).
  • Authored by your firm — on the Enterprise plan, your firm can author its own frameworks, encoding your proprietary methodology (these carry an org badge). See Framework Authoring.

Every firm user can use published frameworks (consultants assess against them; Org Admins control their visibility via Practices).

How it works for your firm

  1. A framework is published — either by Celeredge, or by your firm via Framework Authoring (Enterprise).
  2. The framework appears under its practice, with a source badge (org / shared / global).
  3. Consultants select it in Diagnose and run a versioned assessment against the engagement's evidence.
  4. The assessment produces per-dimension answers with confidence and citations, which become the engagement's gaps.

Framework status

A framework moves through states before your firm can run it: Draft → Compiled → Published. A framework must be fully prepared (compiled) before its assessments can execute. If a framework appears available but assessments fail to start, it may not be fully compiled — contact your Celeredge administrator.

Best practices

  • Match framework to objective. Pick the framework whose methodology fits the engagement's goal, not just the one you know best.
  • Run with evidence loaded. Frameworks assess against retrieved evidence — connect documents and data first (see Evidence Collection).