If you run SAP, you already know the migration to S/4HANA tends to slip, cost more than the business case allowed, and surface its worst problems too late to fix cheaply. Agentic AI changes the part of that equation you have least control over: the mechanical, repeatable work that eats the schedule and the budget.
The companies that treat the move as a deadline to survive will get a more expensive copy of the system they have now. The companies that treat it as a chance to redesign how the business runs — with agents doing the heavy data work — get something they cannot buy at any other time.
On June 11, 2026, the CEOs of LeapGreat, REALTECH AG, and the Value AI Institute met in Walldorf, next to SAP’s headquarters, to announce a joint initiative built around exactly that distinction.
- Agentic AI cuts SAP S/4HANA migration timelines by up to 40% and costs by up to 30% by automating data collection, reconciliation, and test-case work (Atos).
- Migrations overrun because tedious data work — not hard engineering — sits on the critical path. Nearly 60% of SAP migrations exceed schedule and budget (ISG).
- Data problems hide until UAT because they are decided at design and never tested against real volume. A digital twin moves that discovery back to design, where a mismatch is still a decision you can change.
- Plugging agents into an unchanged, deadline-driven port produces zero value. Let the business goal set the scope before the technology does, then redesign the process the agents will run on.
- A migration happens every 5–10 years and is the rare moment a company is already deep in its own data and processes — the one condition under which an AI business case actually scales.
| Key Finding | Source |
|---|---|
| Nearly 60% of SAP migrations exceed schedule and budget | ISG |
| Around 83% of data migration projects overrun on cost, time, or both | Gartner |
| AI-driven toolchains cut migration timelines by up to 40% and costs by up to 30% | Atos |
| SAP consulting rates projected to rise 30–50% across 2026 and 2027 | Gartner 2024 ERP Migration Cost Benchmark |
| SAP ECC mainstream maintenance ends December 31, 2027 | SAP |
| Roughly 17,000 of SAP’s 35,000 ECC customers (nearly half) will still be on legacy at the deadline | Gartner |
| Joule Studio Agent Builder reaches general availability on BTP in Q1 2026 | SAP |
| Replicating legacy processes raises customization, cost, long-term risk, and maintenance effort | Birchman Group, March 2026 |
Why do SAP S/4HANA migration budgets and timelines almost always overrun?
SAP S/4HANA migrations overrun because a large share of the work is tedious rather than hard — and that tedious data work sits squarely on the critical path. Ask anyone who has run an SAP program what they expect, and the honest answer is delay. ISG found that nearly 60% of SAP migrations exceed their schedule and budget. Gartner’s figure for data migration projects broadly is higher still, around 83% overrunning on cost, time, or both.
Staff spend weeks pulling structured and unstructured data out of the current systems, reconciling it, mapping process variants, and generating test cases. When agents absorb that collection-and-reconciliation work, the timeline math shifts. Atos reports that AI-driven toolchains can cut migration timelines by up to 40% and costs by up to 30% — the practical counter to Gartner’s 2024 ERP Migration Cost Benchmark projecting SAP consulting rates rising 30–50% across 2026 and 2027.
“A lot of the very valuable time of IT employees and also business employees is spent on collecting all structured and unstructured data across the current systems. AI can automate this process.”
Stefan Schoepfel, CEO, Value AI Institute
The point of pushing the mechanical work toward automation is to move people’s hours onto the decisions that actually need judgment, rather than the data plumbing that does not. Schoepfel’s advice on where to point that freed-up capacity: start in an area with lower complexity but very high value, so the project earns confidence and budget before it takes on the harder redesign work.
“We have made it our mission to get SAP implementations as close to the push of a button as technologically possible.”
Bob Cummings, CEO, LeapGreat
Why do SAP data problems stay hidden until UAT, and how does agentic AI surface them earlier?
Data problems hide until UAT because they are decided much earlier — at design — and then left undiscovered until real reconciliation runs against rules nobody wrote down. The most demoralizing pattern in any migration: multiple mock loads passed, everything looked fine, and then in UAT there were missing records, reconciliation mismatches, and reporting that did not tie out.
When the target system’s process logic and data rules are agreed on paper but never run against real volume, the gap stays invisible through every mock load that does not stress the same conditions UAT will.
“The biggest failure point in migration is poor alignment.”
Daniele Di Croce, CEO, REALTECH AG
REALTECH AG’s answer in the joint initiative is to make that gap visible while it is still cheap to fix. A digital twin moves the reconciliation discovery from production UAT — where a mismatch is an incident — back to the design phase, where it is a decision you can still change.
“With Smart Transform, we can use AI to build a digital twin of their future system very quickly. This allows them not only to test drive, but also to quickly refine and adjust their target.”
Daniele Di Croce, CEO, REALTECH AG
“AI can only be as smart as the data that you feed it in. A clean baseline, a clean data structure gives you a huge springboard for optimization, acceleration, quality, and cost efficiency.”
Daniele Di Croce, CEO, REALTECH AG
How does agentic AI fix the silos that fragment an SAP S/4HANA migration?
Agentic AI fixes migration silos by acting as a translator and synchronizer across separate workstreams, so the connections between data, reporting, and code remediation stay managed instead of falling into the gaps. On most large programs the work splits across teams that rarely see each other’s status. Many migration programs involve multiple system integrators and niche specialists, but there is a lack of clear decision-making rights, acceptance criteria, and responsibilities between providers. Decisions fall into the gaps between teams, and no one holds the view of how a change in one workstream breaks another.
“Taking a boat and going from one island to another. There’s the data island, there’s a reporting island, there’s the code remediation island.”
Bob Cummings, CEO, LeapGreat
Each island makes local progress while the connections between them go unmanaged. The work that bridges those islands is exactly what AI is suited to.
“AI can act as a translator and synthesizer. It bridges the gap by interpreting, harmonizing, and synchronizing data flows between completely different tools and the S/4 system itself.”
Daniele Di Croce, CEO, REALTECH AG
“Now we imagine a world where AI moves away all those operational tasks and provides a new transparency over the system, but also over the opportunities that you have to automate the processes and better support your future business.”
Stefan Schoepfel, CEO, Value AI Institute
Does the SAP ECC 2027 deadline justify a migration with no business value?
No. If you approach the project as a 2027 compliance exercise, a compliance exercise is what you get — and plugging agents onto a deadline-driven port produces nothing because there is no redesigned process for them to improve. SAP ECC mainstream maintenance ends December 31, 2027, and Gartner estimates that nearly half of SAP’s 35,000 ECC customers — roughly 17,000 businesses — will still be on legacy when that date arrives.
The way out is to let the business outcome set the scope before the technology does. The goal decides which processes are worth redesigning, which data must be clean, and where agents earn their keep.
“Start first with letting your business goal dictate.”
Bob Cummings, CEO, LeapGreat
“Simply plugging AI agents into your system landscape will drive zero value.”
Christian Klein, CEO, SAP (Sapphire 2026)
“If you’re very honest, very few AI business cases scale in a company and provide high business value. But if you manage to connect to the data and the processes in a company, that has the potential to provide a very good business use case.”
Stefan Schoepfel, CEO, Value AI Institute
A migration is the rare moment when a company is already deep in its own data and processes, with budget and executive attention committed. That is exactly the condition under which an AI case actually scales.
What does a copy-paste SAP migration cost a business that skips the redesign?
A copy-paste migration moves old inefficiencies into a more expensive environment, raising customization, cost, long-term risk, and maintenance effort while making every future upgrade harder. The default instinct under deadline pressure is to replicate the legacy system as faithfully as possible. A Birchman Group report from March 2026 names the cost: too often, organizations try to replicate legacy processes rather than adopt standard SAP ways of working, which increases customization, cost, long-term risk, and maintenance effort.
“If you just pack up your old clutter and move into the new house unchanged, you’ve simply transferred your old inefficiencies into a more expensive environment.”
Bob Cummings, CEO, LeapGreat
Every custom workaround you carry across becomes something you now maintain on pricier infrastructure, and every standard SAP capability you override is a future upgrade you have made harder.
“It’s a rare window of opportunity to automize your business. You don’t do ERP transformations every couple of years. Treating an ERP migration as a purely technical exercise is sometimes a bit of a lost opportunity, especially when you see what’s possible now — because these projects only happen every five to 10 years.”
Bob Cummings, CEO, LeapGreat
This is where the three-way initiative holds together as one project rather than three. The Value AI Institute frames the business outcome. REALTECH AG builds the digital twin through Smart Transform. LeapGreat brings the acceleration method that pulls decisions to the front of the project and keeps the other two aligned, so a customer experiences a single coordinated program.
“We are now building AI agents that can handle complex, tedious tasks faster and cheaper than ever before.”
Bob Cummings, CEO, LeapGreat
With SAP’s Joule Studio Agent Builder reaching general availability on BTP in Q1 2026, the companies that arrive at S/4HANA with clean data and a redesigned process layer can put agents to work on day one.
