AI and SAP careers: what changes, what doesn't

AI in SAP work is the use of generative and predictive tools to draft, look up, classify and summarise the routine output of a project — specifications, code, test scripts, tickets — while the decisions those outputs feed stay with an accountable person. The short version: AI compresses tasks, not accountability — and SAP work is unusually heavy on the part that doesn't compress.

What compresses

• First drafts — specs, test scripts, training material, documentation: the writing-down half of consulting accelerates dramatically.
• Code scaffolding and lookup — boilerplate, unfamiliar-API navigation, "how do I read this table" questions.
• Ticket triage — classification and first-response in AMS pipelines.
• Explaining the standard — what a transaction does is now cheap knowledge; why your client's variant behaves oddly is not.

What appreciates

• Judgment under ambiguity — the fit-gap call, the "should we customise" trade-off: the core of the consulting job is deciding, and deciding needs someone accountable.
• System-of-record stakes — no company lets generated changes flow unreviewed into the machine that books its revenue; review skill becomes more valuable as generation gets cheaper.
• Cross-area seeing — knowing that the SD symptom is an FI cause (adjacency knowledge) is exactly what pattern-completion tools lack context for.
• The human side of change — workshops, resistance, trust: unchanged.

How does each SAP role change?

Every role keeps its accountable core and sheds part of its routine output; what differs is which of the two fills the day. Developers: more review, architecture and integration judgment; less typing. Functional consultants: the documentation burden shrinks, the advisory core grows. Business analysts: drafting compresses hardest — the differentiator shifts to stakeholder craft. Juniors: the classic learning tasks are automatable, so deliberate access to real problems matters even more than before.

Will AI replace SAP consultants?

It replaces parts of the work, not the seat, and the reason is structural rather than hopeful. An ERP project exists to change the system a company books its revenue and pays its people through; every change to it is signed off by a named person, tested against real data and audited afterwards (control as a system). Tools that draft faster do not remove the sign-off, the test or the audit — they move more of the consultant's day onto those three. What does change is the ratio: fewer hands may produce the same specification, while the scope a single person can be accountable for grows. The consultants exposed are the ones whose whole value was producing the draft; the ones insulated are the ones a client would still call when the draft is wrong. Position for the second kind, whichever role you hold.

A grounded way to position

• Use the tools daily — fluency is table stakes, not a specialism.
• Invest where AI is weakest: one area deep, its neighbours understood, real project scars.
• Stay visible — in a market where routine output is abundant, provenance (a real person, findable, with their facts stated in public and consistent everywhere) quietly gains value; that is precisely the fact-based profile the map exists for.