Capability
How to approach this class of problem.
Each discipline brings its own methods, artefacts, standards, and evaluation criteria. The platform provides the operating model they run on.
Platform
One standard. Every kind of work.
One standard. Every kind of work.
Run every discipline on one connected foundation for evidence, validation, and approval.
Beyond the instruction box
A longer prompt does not produce governance, repeatability, or an audit trail.
Most AI tools give you an instruction box and an output. That works until someone asks who approved the result, which version of the standard it was measured against, or whether the same input would produce the same assessment tomorrow.
Those questions need infrastructure. Kryterea provides it as a platform so that every capability, every discipline, and every team operates on the same foundation.
Read the methodThe common layer
What every capability inherits from the platform.
The platform separates what is universal from what is domain-specific. These services run beneath every capability so a new discipline does not rebuild the product.
Plugs into every service
Common platform servicesInherited by every capability
Context
The information the system needs to reason well.
Your material, the domain, the constraints, the audience, and the project history enter as stated inputs with their provenance recorded.
Knowledge
Organisational and domain understanding that persists.
Research, reference material, and organisational knowledge live in a structured knowledge base. A claim can point at the exact material that supports it, and the link survives export.
Intelligence
AI that works within the problem, not beside it.
Research, analysis, transformation, and creation are grounded in the capability, context, and state of the work. The system understands which stage the preparation is in and what remains unresolved.
Validation
Rubrics versioned, scores comparable, findings actionable.
The validation engine applies the published rubric for each deliverable type. A score records which rubric version produced it, so results stay comparable across runs, reviewers, and time.
Governance
Approval, accountability, and audit in one chain.
The system drafts, measures, and reports. It does not approve on your behalf. Approval scope is an explicit permission, and every export records what was approved, by whom, and against which standard.
Why it scales across disciplines
One workspace. Every discipline that needs a standard.
Kryterea separates the workspace from the capability. A new discipline brings its own methods, standards, and evaluation criteria without rebuilding what it runs on.
Capability Packs
A structured professional operating model for a defined class of work: its purpose, scope, roles, methods, activities, artefacts, and the standards that apply.
The approval is the record.
The system drafts, measures, and reports. It does not approve on your behalf, it does not spend, and it does not take an irreversible action while you are not looking. Approval scope is an explicit permission, not a side effect of access.
Separation of duties is the architecture, not the policy.
Sample runEvidence trail
Architecture approved for production deployment, pending operations review of recovery path.
- Approved byLead Architect, 2026-08-14, against Architecture description v1.1.
- ConditionOperations stakeholder concern (failure recovery) must be addressed before release.
- Findings at approvalScore 79/100. One fail finding outstanding (no runtime view). Two caution findings accepted.
- Audit refPRJ-041-ARCH-003. Export includes rubric version, score, findings, and approval chain.
Intelligence within the capability.
AI operates through the capability context, not beside it. The system understands which capability is active, which stage the preparation is in, which standards apply, and what remains unresolved. The capability provides the operating context. AI provides intelligence within it.
This means the user does not need to reconstruct the state of their work in every prompt. The platform holds the context, and the intelligence layer works inside it.
The system beneath the standard.
Walk the governance model on your own workspace. See how the validation engine, the approval chain, and the audit trail work together before you commit a team.
Coming soon. Start with one project.