A newcomer to the SAP world could be forgiven for thinking AI has already replaced the migration story that has dominated the past three years.

But recruiter data shows that rates for experienced S/4HANA architects keep climbing, vacancies for ABAP-to-Clean-Core remediation specialists sit open for months, and consultancies still build their bench around delivery track record.

The real story for 2026 comes down to a single gap, visible from two different angles. SAP customers are behind on migrating to S/4HANA, and they are further behind on using SAP’s own AI in production.

Both delays share one cause. Most of the client base has not finished the move to S/4HANA, and Business AI only works properly once that move is done.

The Distance Between the Keynote and Adoption

SAP’s messaging at the Sapphire keynote in May described their vision of an “autonomous enterprise,” and progress toward that already including more than fifty Joule Assistants and upward of two hundred specialized agents, while most customers were still thinking about S/4HANA. 

SAP’s current cloud backlog for the first quarter had climbed twenty-five percent at constant currencies to nearly twenty-two billion euros, with Cloud ERP Suite revenue up around thirty percent.

That figure needs a caveat: current cloud backlog measures contracted revenue expected over the next twelve months, not systems already live. A strong backlog and a slow completion rate describe the same pipeline at different stages. Deals get signed well before the work behind them gets delivered, so both numbers can be true at once without contradicting each other.

The German-speaking user group DSAG published a more grounded picture in its investment report, released in late February. Forty-three percent of member companies report running some form of AI in production today. More than three-quarters of those rely on Microsoft Copilot, ChatGPT, or another outside tool, rather than anything built by SAP.

A third of respondents named Business AI part of their long-term target picture, according to DSAG chairman Jens Hungershausen, but that figure describes an intention, not a deployment.

Delivery timelines have the same credibility problem, and this one predates the AI conversation by years.

A Horváth study of two hundred mid-sized and large companies, fielded in early 2025, found that only eight percent of S/4HANA transformations finished on schedule. The average project ran thirty percent too long, and well over half went over budget.

Line the findings up and it is clear that the roadmap describes what the software can do, the backlog describes what’s been sold, and neither describes what’s actually running.

The Deadline Ladder

The often-repeated “2027 deadline” is actually a longer and more forgiving timeline. Mainstream maintenance for SAP Business Suite 7 (the ECC generation) runs out at the end of 2027.

Customers not ready by then can buy extended maintenance through the end of 2030, at a premium of two percentage points on their maintenance fee.

Anyone still uncommitted after that point doesn’t lose support outright. SAP shifts them onto a customer-specific maintenance arrangement that keeps fixing known issues without adding anything new.

S/4HANA itself carries an innovation commitment running to 2040.

A generous runway like that would normally ease hiring pressure rather than intensify it. It doesn’t, for a specific reason. According to the SAPinsider benchmark published in March only thirty-four percent of organizations have fully completed their move to S/4HANA, against fifty-five percent who describe some form of deployment as done, and a further thirty-six percent are still building the business case or mid-implementation.

A program of this size runs twelve to thirty-six months from kickoff to cutover, and on the Horváth numbers above, often longer.

Anyone still building a business case today, aiming to avoid the 2028 to 2030 maintenance premium, needs to be staffed and moving now, not in 2027. That’s what’s driving the hiring market, not the calendar date itself.

The same research recorded a shift for delivery planning specifically: S/4HANA Cloud now sits at twenty-six percent of the installed base, nearly level with the twenty-nine percent still on the traditional, more customizable deployment, the closest the two figures have been in SAPinsider’s tracking.

That changes what a migration team looks like as much as the deadline does.

Why the Software Comes First

There’s a structural reason SAP customers can’t buy their way into Business AI without dealing with the migration first, separate from anything to do with marketing.

Joule and the wider Business AI set draw their business context from the SAP Knowledge Graph, a layer built around the S/4HANA data model, its best-practice process content, and industry-specific metadata that ECC never held.

An infrastructure layer called AI Foundation sits beneath every agent, handling model access and governance. Remove that foundation, and the assistant has nothing dependable to reason over.

SAP softened this position at Sapphire in May. Nearly two-thirds of its customers still run ECC or an on-premise version of S/4HANA, and SAP announced that a subset of Joule agents would reach those customers from the third quarter of 2026.

But access isn’t unconditional. Customers qualify by combining a RISE with SAP contract with a meaningful shift of on-premise maintenance spend toward the cloud, with the Max Success Plan alongside as a separate paid service built on top of that commitment.

So cloud customers keep the full feature set; everyone else gets a subset, in exchange for a signature promising to leave ECC eventually.

Read that way, the concession isn’t evidence that AI now works independently of migration. It’s SAP using partial AI access as a lever to pull migration forward.

A UBS analyst had already tested the edges of this on SAP’s first-quarter earnings call back in April, asking how SAP intends to protect the ontology and knowledge graph behind its AI without giving that intellectual property away through open interfaces. Management’s answer, that interfaces stay open but throttled, only makes sense if the Knowledge Graph itself is the asset being protected.

An April 2026 report added a second point: SAP’s own partner ecosystem has started raising lock-in concerns over a new AI-related API policy.

SAP is not simply building AI on top of S/4HANA. It’s using AI access as one of the tools pulling the remaining ECC base toward migration, and a consultancy pitching Business AI as a stand-alone service to a client still on ECC is going to find the client can’t actually buy what’s being pitched.

Where the Squeeze Happens

The shortage shows up in specific job titles, not a general shortage of “SAP people.”

Architects who have carried an S/4HANA program from blueprint to cutover are scarce, causing very high rates for senior candidates. The top end of that market (architects and program directors with a proven migration record) runs higher still.

Firms describe a particular scarcity in people who combine SAP knowledge with integration, data, or SAP Business Technology Platform experience, since a modern S/4HANA landscape rarely stands alone anymore.

There is another more basic problem. A generation of consultants who learned SAP on ECC is moving into leadership or retiring faster than firms can train replacements.

Also, the newer public-cloud implementations under GROW with SAP need a narrower band of configuration skill than the private-cloud and on-premise work under RISE.

That pulls staffing in two directions at once, and it follows from the twenty-six/twenty-nine percent cloud split covered above: a firm bidding for both kinds of work now needs two different benches, not one bench stretched thinner.

DSAG’s own report described member budgets growing more slowly and spending becoming more selective overall, yet the specific skills tied to migration delivery keep getting more expensive regardless.

Selective spending still has to go somewhere, and right now it goes toward the people who can get a cutover done on schedule.

Why the Job Advert Still Asks for Modules

The technical explanation above, that Business AI depends on the S/4HANA data model, is only half the reason S/4 skills still dominate hiring. The other half has to do with how consultancies screen candidates.

Ask a hiring manager why an S/4HANA finance role still lists SAP Certified Associate in FICO as a requirement, ahead of anything to do with AI, and the answer is partly practical and partly a way of putting off a harder decision.

On the practical side, RFPs from clients still name specific module certifications as eligibility criteria, and some of SAP’s certifications have moved toward open-book, task-based assessment that mimics real project work, giving employers a credential they can trust as proof of hands-on ability.

A résumé claiming “AI skills” carries no equivalent proof yet. An experienced hiring manager discounts the claim on sight, similar to the way they’d discount an unverifiable reference.

The bench-utilization economics behind most consultancies reinforce the same habit. A manager staffing a fixed-price migration prefers someone billable and provably competent from day one, over someone whose value might be realized eighteen months into an AI program.

Nobody wants to commit a training budget to a skill set that could look different again by the time it’s needed, so it’s organizationally convenient to say the AI question can wait until after go-live. That’s a rational response to uncertainty, though it also means the industry keeps screening for what it already knows how to measure, and candidates who overclaim AI familiarity without project evidence tend to get filtered out faster than candidates who say nothing about AI at all.

Two separate mechanisms, then, push in the same direction. SAP’s own architecture ties AI to the migration, and the hiring market independently rewards what it can verify over what it can’t.

The Dual-Track Approach

For an individual consultant, keep the certification path weighted toward the named S/4HANA modules that clients and RFPs still screen for. That’s the credential the hiring market above actually trusts.

Treat time spent with Joule for Consultants or Joule Studio as a documented add-on alongside that. That’s the piece tied to SAP’s own architecture, worth building familiarity with regardless of how fast the market for it moves.

If your current engagement gives you any access to these tools, write up what you did with them: specific numbers, named tasks. A documented case beats a general claim every time a hiring manager reads it.

ABAP Cloud deserves a specific mention. Recruiters report rising demand for it ahead of classic ABAP, tied to Clean Core adoption, though a precise industry-wide growth figure isn’t available. It’s a genuine S/4HANA delivery skill in its own right, and the same extension layer will be used to build custom Joule agents later, which makes it one of the few areas where one action addresses both tracks.

For a consultancy, the same logic applies at firm level. Given that only eight percent of S/4HANA transformations finish on schedule, a firm’s most sellable point of difference right now is delivery discipline over AI, but don’t ignore it.

Build a relevant AI capability into existing statements of work as one workstream among several, rather than a separate service line. SAP’s move to consumption-based AI pricing matters here too: a workstream billed on outcomes is easier to quote than one billed on unpredictable Joule usage, which is another reason to scope AI as an addition to known deliverables rather than price it on its own.

After the Deadline Passes

Once the bulk of the ECC install base clears the 2027 line, whether by migrating or by paying for extended maintenance, the work on offer shifts from getting onto S/4HANA to getting value out of it.

That next phase will be about value-realization rather than a migration, and the optimization and process redesign work migration programs rarely have time for while cutover is still ahead of them.

The gap this piece opened with doesn’t close on that date. It becomes less about whether a client has migrated, and more about whether they can use what they migrated to, with AI capability becoming available.

The firms and consultants who spend this migration wave building their delivery track record, while treating AI exposure as something to document rather than market, will be the ones with something real to offer once that door opens.

If you are an SAP professional looking for a new role in the SAP ecosystem our team of dedicated recruitment consultants can match you with your ideal employer and negotiate a competitive compensation package for your extremely valuable skills, so join our exclusive community at IgniteSAP.