Agents / AI Centre of Excellence
Use-Case Sizing Agent
Sizes a use case for effort and value from the evidence supplied: data readiness, build effort, reuse breadth, value case and model risk tier.
Runs, last 30 days
9
Tools
4
Data scopes
3
Human checkpoints
1
About this agent
Sizes a use case for effort and value from the evidence supplied: data readiness, build effort, reuse breadth, value case and model risk tier. Each factor carries its reasoning, and missing evidence is flagged rather than assumed. The funding decision stays with the Group AI Steering Committee.
Tools4
- Sizing rubric (five factors)
- Evidence bundle reader
- Comparable use-case index
- Scorecard writer
Data access scope3
- Read-only: use-case submissions and their evidence
- Read-only: scorecards of previously sized use cases and their actual delivery
- Write: draft scorecards marked as agent-prepared
Guardrails3
- Every factor score cites the evidence it rests on; a factor without evidence is scored as missing, never estimated
- Cannot change a size once the steering committee has confirmed it
- Draft sizes are never shown as final to the requesting business unit
Human-in-the-loop1
- Group AI Steering Committee confirms the size and commits budget
An agent gathers the evidence, applies the rule, scores the case and recommends an outcome. A named human decides and the decision is recorded against them.
- Acts for
- Marketplace Operator
- Action class
- advisory
- Decisions / 30d
- 9
- Overturned by a human
- 11%
Memory
Retains scorecards and actual delivery effort for comparables; purged when the use case is closed.
Safe state
Stops scoring and returns the use case with the factors it could not evidence when more than two factors lack evidence.
Reviewer time per case: 15 min with this agent, 90 without.
Orchestration
Every tool call is scoped by the declared data access above; the orchestrator cannot reach systems outside it. Owning business unit: Group Data & AI.
Ownership and sensitivity
Who approves access
- Owner, AI Centre of Excellence, Farah Iskandar
Business-sensitive. Models and data products scoped to named business units.
Entitlement per business unit, approved by the owner; conditions attach.
- Aggregated only
- No row-level customer data; aggregates with small-cell suppression.
Community · 0 threads
Questions, findings and requests from the business units that use this asset. Owners reply here; threads with upvotes surface to the owning team's inbox.
Live demo run
Watch the agent execute a real scenario step by step, every tool call, validation, and human checkpoint is traced and auditable. Typical run: ~5.6s.
Used in workflows
Retiring this asset would require these chains to be re-pointed first.
Audit
Updated 2026-07-21. Run traces retained 24 months for audit under the platform governance policy.