Is Your SAP Core AI-Ready?
What 100+ Readiness Assessments Reveal
Everyone left SAP Sapphire 2026 talking about the Autonomous Enterprise — hundreds of AI agents, Joule assistants across the suite, a finance close that increasingly runs itself. It was a compelling picture of where enterprise software is going.
The part that didn't make the highlight reel: that autonomous future was built for SAP S/4HANA with a clean core, not for the legacy ECC systems most of the market still runs. SAP said as much itself. Alongside the agent announcements, it launched a formal Clean Core Certification that certifies BTP extensions as upgrade-compatible across three S/4HANA Cloud release cycles — quietly turning “clean core” from a best-practice suggestion into a certified prerequisite. And CEO Christian Klein was blunt about the ceiling: simply plugging AI agents into your landscape, he told the keynote audience, will drive zero value without the process and data work underneath.
That gap — between the AI on stage and the core most customers actually run — is now the single most important thing for SAP leaders to understand.
The market backdrop: a slow-moving install base against a hard deadline
To see why “AI-ready” matters, start with the shape of the market.
SAP's ECC install base is enormous and, until recently, slow to move. Industry analysis pegs the base at roughly 35,000 ECC customers, of whom only around 39% had migrated to S/4HANA by the end of 2024 . Heading into 2026, multiple assessments put the share of customers who still haven't started a serious migration at 60% or more. In other words, the majority of the market is still standing at the starting line — even as the value narrative has shifted decisively toward AI.
Against that inertia sits a fixed deadline — but read the deadline carefully, because it is not one date. SAP has confirmed that mainstream maintenance for SAP Business Suite 7 (including SAP ERP 6.0) ends at the end of 2027 . What most coverage omits is that the deadline you actually face depends on your enhancement package. Landscapes on EHP 0–5 already passed end of mainstream maintenance at the close of 2025; EHP 6–8 run to the end of 2027. An optional extended-maintenance window then runs from 2028 through the end of 2030 at a two percentage-point premium — buying security and legal updates only, with no innovation attached. SAP has publicly stated there is no further extension, while committing to maintain S/4HANA through 2040.
The practical consequence: a meaningful slice of the base is already living on borrowed time and doesn't know it. If you're on EHP 5 or below, your “2027 problem” was a 2025 problem — you're running unsupported now.
This is the tension defining the SAP market in 2026: a large, under-migrated base; a support cliff that arrives earlier than the headline suggests; and a vendor pouring its innovation — especially AI — exclusively into the destination platform.
The three forces squeezing the window
- The migration timeline math. A typical S/4HANA migration runs 12 to 24 months, and longer for complex or heavily customized landscapes. Count backward from your real EHP deadline and the comfortable window has, for many organizations, already closed. Readiness work that looks like a 2027 problem is a 2026 decision — and for the EHP 0–5 cohort, an overdue one.
- The capacity crunch. Deadlines create demand spikes, and demand spikes tighten supply. As more of the 60% who haven't started finally move — many inside the same 2026–2027 window — the market for experienced S/4HANA talent, partner capacity, and delivery slots tightens. Early movers lock in capacity; late movers inherit whatever's left, typically at higher cost.
- The AI value gap. This is the newest and most strategic force. SAP's roadmap now treats AI as the primary reason to be on S/4HANA at all — Joule, SAP Business AI, and agentic automation are destination-platform capabilities by design. Every quarter spent on an AI-incapable core is a quarter of that value left unrealized, and a widening gap against competitors already compounding AI-driven efficiency.
Why the core — not the AI — is the real blocker
Here's the counterintuitive part. For most SAP customers, the thing standing between them and SAP's AI future isn't the maturity of the AI. It's the state of their core.
SAP Business AI, Joule, and agents depend on standardized data models and upgrade-stable extensions. A heavily modified ECC system breaks those assumptions before an agent ever runs — which is precisely why SAP moved to certify clean core rather than merely recommend it. An SAP core is AI-ready when it runs on S/4HANA with a clean core: minimal in-core custom code, extensions on SAP Business Technology Platform (BTP), de-duplicated and governed master data, and business logic that lives in the system rather than in spreadsheets.
If you're in a regulated industry, the stakes flip in both directions
For banking, insurance, and life sciences customers, clean core is simultaneously harder and more urgent. Harder, because every change collides with GxP validation, SOX-style change control, and audit-trail requirements — you cannot “just” retire custom code or de-duplicate master data without documented, validated evidence. More urgent, because AI agents executing mission-critical processes in a regulated environment must be governable, auditable, and explainable by design; a messy core doesn't just produce unreliable automation, it produces automation you can't defend to an examiner. In these landscapes, the readiness assessment isn't a technical nicety — it's the document trail your compliance function will eventually demand anyway.
The three blockers we find in 100+ assessments
Across more than 100 SAP readiness assessments spanning very different industries and landscapes, the same three problems recur. If two or more sound familiar, that's the signal to start with the core, not the agents.
- Custom-code sprawl. Years of custom ABAP and Z-programs no one wants to touch, own, or decommission — each one complicating upgrades and undermining the standardized extension model SAP Business AI needs. Fix: analyze with SAP Readiness Check and the ABAP Test Cockpit, retire the unused, and move necessary extensions to BTP — out of the core.
- Duplicate master data. The same customer, vendor, or material living in three places, in three slightly different forms. AI is only as trustworthy as the data beneath it. Fix: profile, de-duplicate, and govern before migration — cleaning during conversion is far cheaper than remediating after go-live.
- Undocumented workarounds. Critical business logic hiding in spreadsheets, shadow systems, and a few long-tenured employees' heads. Automation can't see any of it, which makes it the most common source of migration surprises. Fix: document it and bring it back into the system so processes become transparent, testable, and automatable.
The strategic choice: how you migrate shapes how AI-ready you become
Your assessment findings — not a vendor preference — should drive the route.
Three Routes. One Question: Are You AI-Ready?
Your assessment findings should pick the route — not a vendor preference.
Processes are sound and you want to preserve history.
Carries code and data debt forward unless remediated.
Only as clean as the remediation you do en route.
You want a clean, best-practice redesign.
Higher change effort; extensions rebuilt.
Highest clean-core ceiling.
Keep what's worth keeping, drop the rest.
Needs strong data-selection tooling and governance.
Balanced — clean where it counts.
| Approach | Best when | Trade-off | AI-readiness impact |
|---|---|---|---|
| Brownfield (system conversion) | Processes are sound; you want to preserve history | Carries existing code/data debt forward unless remediated | Only as clean as the remediation you do en route |
| Greenfield (new implementation) | You want a best-practice redesign | Higher change effort; re-implement extensions | Highest clean-core ceiling |
| Bluefield / selective | Keep what's worth keeping, drop the rest | Needs strong data-selection tooling and governance | Balanced — clean where it counts |
The pattern we see: teams that treat migration as a technical “lift” reach S/4HANA but not AI-readiness, because they carried the debt across. Teams that treat it as a clean-core program arrive ready to switch on Business AI.
The alternative path — and its ceiling
Not every organization will migrate on time. Some will opt for extended maintenance; others explore third-party support providers such as Rimini Street to stretch ECC further. These are legitimate ways to manage cost and timing — but they share one ceiling: they keep the lights on; they do not make the core AI-ready. The SAP AI roadmap does not run backward onto ECC. Choosing to stay is, functionally, choosing to defer AI value.
What SAP leaders should do next
Action Plan for SAP Executives
Run a readiness assessment to baseline custom code, data quality, and process gaps — this is the input every other decision depends on, and the first thing you'll need if you're already past your EHP deadline.
Build the clean-core roadmap and choose brownfield, greenfield, or bluefield from that baseline.
Count backward from your real deadline — not the headline one — and secure delivery capacity before the crunch peaks.
Sequence AI after the core — line up Joule and Business AI for the moment the clean core is in place.
The first AI question isn't “which agents do we want?” It's the quieter, more consequential one: is our core actually ready for what's coming?
FAQ
When does SAP ECC support end?
Mainstream maintenance for SAP Business Suite 7 (ECC), including SAP ERP 6.0, ends at the end of 2027. Optional extended maintenance runs from 2028 through the end of 2030 at a +2 percentage-point premium, covering security and legal updates only. SAP has confirmed there is no further extension.
What is a clean core in SAP?
An SAP S/4HANA system kept close to standard: minimal in-core custom code, extensions on SAP BTP, clean governed master data, and upgrade-stable configuration. It is the foundation required for SAP Business AI and Joule.
Why does SAP Business AI need a clean core?
Business AI, Joule, and agents rely on standardized data models and extension-safe upgrades. Heavy customization and duplicate data break those assumptions and produce unreliable automation.
How long does an SAP S/4HANA migration take?
Typically 12 to 24 months, and longer for complex landscapes, which is why planning backward from December 2027 matters.
Is ECC AI-ready?
No. SAP AI capabilities are built for S/4HANA with a clean core. A legacy ECC landscape is not AI-ready until it is migrated and cleaned.
Not sure where your core stands?
Uncover hidden code debt and data gaps before your EHP deadline hits. Get a customized readiness scorecard based on 100+ successful SAP diagnostics.
Book a 30-minute readiness review