Share with your CIO
SAP is making a direct pitch that the path to the autonomous enterprise runs through S/4HANA, and the clock is ticking. Nazia Pillay, MD of SAP Southern Africa, argues that agentic AI inside ERP represents a genuine operating model shift, not a feature upgrade. AI agents, in SAP’s framing, are co-workers executing end-to-end processes with human oversight, not chatbots surfacing reports. Two enablers matter most: context (the structured business data already inside SAP systems) and governance. With mainstream ECC support ending in 2027, South African enterprises are losing the option to watch from the sidelines.
What this means for your business
The 2027 ECC deadline is doing real work in this conversation, and CIOs should be honest about whether their current migration posture reflects a considered architecture decision or simple inertia. Companies still running experiments rather than production deployments of AI aren’t being cautious; they’re accumulating a governance debt (the gap between where AI is actually operating in the business and where formal controls exist) that compounds with every new proof of concept that never graduates to scale.
Pillay’s core claim, delivered by an MD whose incentive is to accelerate S/4HANA migrations, is that data context and governance are the actual differentiators, not the AI models themselves. That framing conveniently positions SAP’s existing knowledge graphs and compliance infrastructure as a moat, which nudges the argument toward “stay in the ecosystem” faster than the evidence alone would take you. But the underlying logic holds independently of who’s selling it. Agentic AI, meaning systems that plan, execute, and iterate across multi-step business processes without hand-holding, genuinely does fail when it lacks reliable data structures and clear authorization rules. The ERP layer is where both of those live for most large enterprises.
The real exposure for CIOs isn’t missing the agent wave; it’s arriving at agent deployment with a data foundation that was never designed to support autonomous decision-making. If your S/4HANA migration has been treated as a technical lift-and-shift rather than a process redesign, you’ll hit the governance wall the moment agents need to act on ambiguous data or cross process boundaries. That’s the budget conversation worth having now, before the 2027 pressure forces a rushed answer.
Concept deep-dive: Agentic AI
Agentic AI describes systems that don’t just respond to a prompt but pursue a goal across multiple steps, making decisions and calling tools along the way, much like delegating a task to a junior analyst who figures out the intermediate steps themselves. In an ERP context, an agent might detect a supply shortfall, identify alternative vendors, draft a purchase order, and flag exceptions for human review, all without a human initiating each step. The governance question is who’s accountable when the agent gets a step wrong.
Based on reporting from ITWeb TV: Agentic AI fuelling rise of the autonomous enterprise, originally published 2026-08-07 05:45:00.
