The 5 Dimensions of SAP AI Readiness
SAP AI readiness breaks into 5 weighted dimensions: Clean Core, Data Quality, Infrastructure/Licensing, Process, and Skills. Here's what each one measures.
Only ~17% of organizations have AI embedded in core workflows, despite ~91% using it at some level.
SAP AI readiness, scored using Kernstein's proprietary SAP AI Readiness Model, breaks into five dimensions:
- Clean Core & Technical Foundation
- Data Quality & Governance
- Infrastructure, Licensing & AI Enablement
- Process & Organizational Readiness
- Skills, Change & Governance Readiness
Clean Core and Data Quality tend to matter most: agents reasoning over messy code or dirty master data fail quietly, giving wrong answers with full confidence. How Kernstein's proprietary SAP AI Readiness Model weighs and combines the five dimensions isn't published.
Why five dimensions instead of one score?
Failure modes differ by dimension, and a single score hides which one is blocking a rollout:
- Excellent data governance + Level A ABAP modifications → still fails
- Clean, modern S/4HANA core + unchecked Joule contract entitlement → still fails
Five independently measurable dimensions point at the actual gap. This is the exact framework behind the site's free interactive Readiness Check, applied to specific answers.
Dimension 1: Clean Core & Technical Foundation
Measures how much of an S/4HANA system runs on SAP's standard, supported code path vs. custom modifications. SAP's extensibility levels, worst to best:
| Level | Description | Debt |
|---|---|---|
| A | Classic ABAP modifications, user exits | Highest |
| B | Mix of classic modifications and key-user extensions | Medium-high |
| C | Key-user extensibility (Custom Fields & Logic, UI adaptation) | SAP's recommended baseline |
| D | ABAP Cloud and released APIs only | Target state |
- Agents reason over standard business objects and released APIs.
- Custom code between an agent and its data source produces results that look correct and aren't.
- Measurement tools: ABAP Test Cockpit clean-core checks, RISE Methodology dashboard.
- Few IT teams have run these checks recently enough to know their current number.
Dimension 2: Data Quality & Governance
- Question: is master data governed well enough that an agent's decision won't be silently wrong?
- Duplicate vendor records, inconsistent customer data, ungoverned reference data don't stop an agent — they make its output unreliable in ways hard to catch after the fact.
- Master data governance maturity (how formally MDG runs) matters as much as current data state — it's what prevents re-decay after a cleanup project.
- Only ~6% of organizations report data environments ready for production AI use.
Dimension 3: Infrastructure, Licensing & AI Enablement
- BTP services provisioned and in active use vs. contracted and idle.
- Joule contract entitlement: RISE with SAP customers typically get ~2,500 Joule messages per Full User Equivalent per year — many Order Forms just say "available in relevant cloud subscriptions" with no figure.
- GROW with SAP and non-RISE private cloud/on-premise carry different, less-publicized entitlement structures.
- Deployment model: Joule and agentic features require S/4HANA.
- ECC 6.0 EHP0-5 mainstream maintenance ended 31 December 2025.
- ECC 6.0 EHP6-8 mainstream maintenance ends 31 December 2027.
- ECC customers get no access regardless of score elsewhere.
Dimension 4: Process & Organizational Readiness
- Measures how standardized core business processes are before automation.
- An agent built for a defined process breaks down against one that varies by region, business unit, or individual judgment.
- Process standardization is the most commonly skipped step under time pressure.
- Only ~1 in 4 organizations has moved beyond AI pilots into scaled production use; process variability is a recurring cause.
Dimension 5: Skills, Change & Governance Readiness
- Basis/ABAP team skills to build and maintain agents on BTP.
- Business-user familiarity with AI tools.
- Formal, approved (not draft) AI governance framework.
- EU AI Act high-risk-system obligations reach full enforcement 2 August 2026 — applies to HR decisions, financial decisioning, public-sector use cases many SAP customers are already piloting against.
- Gartner-style CIO benchmarks: ~16% rate delivery processes AI-ready, ~14% workforce, ~12% architecture.
Why do Clean Core and Data Quality gaps count more?
- Weak fundamentals in either dimension are difficult for the other three to fully offset — exactly how Kernstein's model weighs and combines them isn't published.
- SAPinsider: ~91% of organizations use AI at some level; ~17% have it embedded in core workflows. The gap reflects AI enthusiasm running ahead of technical foundation.
Frequently Asked Questions
What are the 5 dimensions of SAP AI readiness?
Clean Core & Technical Foundation, Data Quality & Governance, Infrastructure, Licensing & AI Enablement, Process & Organizational Readiness, and Skills, Change & Governance Readiness. Together they determine whether Joule Assistants and Agents can run reliably in a given SAP environment.
Which SAP AI readiness dimension matters most?
Clean Core & Technical Foundation and Data Quality & Governance tend to matter most — weak fundamentals in either one are hard to fully offset elsewhere. Exactly how Kernstein's SAP AI Readiness Model weighs and combines the five dimensions is proprietary.
Can you be AI ready without migrating to S/4HANA?
No. Joule and SAP's agentic AI capabilities require S/4HANA; ECC customers don't have access to these features under any of the five dimensions, no matter how strong their data or process maturity is otherwise.
Key Takeaways
- SAP AI readiness splits into five weighted dimensions rather than a single score, so a diagnosis can point at the specific gap instead of a vague verdict.
- Clean Core and Data Quality tend to matter most — weak fundamentals in either one are hard to fully offset elsewhere.
- Infrastructure & Licensing covers both BTP provisioning and actual contract entitlement, not assumed entitlement.
- Process and Skills/Governance round out the framework, and the EU AI Act's 2 August 2026 enforcement date makes governance maturity a harder deadline than it used to be.
- Only around 17% of organizations have AI embedded in core workflows despite 91% using it at some level, a gap this framework is built to explain.
See exactly how your own environment scores across all five dimensions. Take the free SAP AI Readiness Check, a 10-minute questionnaire that produces an instant, scored PDF report with a prioritized upgrade list.