Clean Core Guardian

Overview

The Clean Core Guardian embeds clean-core thinking into daily transformation decisions rather than treating it as a late architecture review.

Lifecycle: End to end
Default autonomy: Assess and block pending review
Human owner: Design Authority or Clean Core Board
Typical consulting roles: Clean-core lead, SAP technical architect, design-authority reviewer, quality-assurance lead

The implementation challenge

Custom complexity rarely arrives as one dramatic decision. It accumulates through individually reasonable exceptions across processes, code, integrations, data and operations.

Mission and boundaries

The Guardian assesses proposals across the clean-core dimensions, identifies alternatives, maintains exceptions and prepares evidence for quality gates.

It does not accept exceptions, determine business value or block work permanently without human review.

Role across SAP Activate

• Discover: establish clean-core ambition and baseline.
• Prepare: define principles, metrics, governance and exception workflow.
• Explore: challenge process and solution deviations.
• Realize: monitor configuration, code, integrations and data decisions.
• Deploy: assess readiness and unresolved exceptions.
• Run: detect drift and support continuous “keep clean” governance.

Key use cases

• classify an extension using approved extensibility patterns.
• challenge a custom process when standard functionality is available.
• assess use of APIs, events and integration monitoring.
• identify data-quality risks affecting clean-core maturity.
• track time-bound exceptions and remediation commitments.

Required skills

SAP clean-core principles, SAP Activate, application extensibility, integration, process standardization, data governance, operations and architecture assurance.

How it works

The agent consumes process and design decisions, code-analysis results, integration inventories, data-quality measures and operational controls. It produces assessments, alternative patterns, exception records, metrics and gate evidence.

Human–agent operating model

The agent performs continuous policy checks and evidence assembly. Human authorities interpret ambiguous cases, balance business value and complexity, approve exceptions and accept residual risk.

Governance and risks

A simplistic guardian can become a mechanical “no” function. Policies need context, proportionality and clear escalation. The agent’s own rules and SAP content must be versioned and reviewable.

Success and maturity

Measures include standard-process adoption, approved extension patterns, exception aging, integration compliance, data quality and upgrade effort.

Example

A team proposes modifying standard code to support a local approval. The Guardian identifies process and side-by-side alternatives, estimates lifecycle consequences and routes an exception decision to the Design Authority.

Related agents

All domain agents. Quality-Gate Assessor and Evidence Validator independently assess its conclusions.