Agentic AI’s 95% Problem: Why Infor’s Honesty Should Reset SAP Customer Expectations

Hiroshi Ozaki connects SAP’s operating signals to executive decisions

In over three decades of enterprise technology work, I have rarely heard a software vendor publicly admit that its headline innovation represents just a sliver of the total effort required to deliver value. Infor’s recent candor—that the AI agent itself amounts to only about 5% of an agentic AI solution—is precisely such a moment. For board members, CFOs, and SAP architects advising leadership, this single data point should recalibrate every agentic AI business case, implementation budget, and vendor negotiation playing out across the SAP ecosystem.

The underlying dynamics are not new. What has changed is the commercial pressure to treat generative and agentic AI as an easy-layer-on-top, when in reality it shines a harsh light on the decades-old challenge of enterprise data readiness and organizational change. If SAP customers and the broader market absorb the lesson Infor is teaching, the next wave of AI adoption could be far more sustainable. If they don’t, we risk repeating the same costly pattern of under-scoped transformations we saw with cloud migrations and early S/4HANA programmes.

The Business Signal: Commoditized AI, Heavy Lifting

The core revelation is straightforward: while the AI agent—the conversational interface or autonomous decision point—grabs attention, it is surrounded by a 95% iceberg of integration, data preparation, process redesign, and organizational change management. Infor’s CTO publicly acknowledged that basic AI features are becoming commoditized across ERP vendors, and that true differentiation will come from deep industry-specific workflows and high-quality data context, not from the agent technology itself. (See ERP Today’s report for the original framing.)

For SAP, this has immediate strategic consequences. The company has heavily marketed Joule, its generative AI copilot, and increasingly agentic AI capabilities embedded in S/4HANA and the Business Technology Platform. If AI agents truly are only 5% of the value equation, then SAP’s ability to monetize these features through premium add-ons or consumption-based pricing depends almost entirely on the other 95%: the maturity of a customer’s data foundation, the quality of process documentation, and the speed with which business users adapt to new ways of working.

Investors should note that SAP’s AI revenue narrative—often presented as an expansion lever on top of cloud subscription growth—may be more fragile than it seems. Unless SAP can clearly demonstrate that its industry-specific content, pre-integrated data models, and ecosystem of partners can dramatically lower the 95% burden for customers, the incremental AI revenue could stall. Basic generative AI features are already being embedded at no extra cost by competitors; premium pricing will only stick if the surrounding solution reduces total implementation effort in measurable, industry-proven ways.

What It Means for SAP Customers

The Infor admission should serve as a planning baseline for every SAP customer currently evaluating agentic AI use cases. If the agent is only 5% of the work, then project scoping that focuses exclusively on AI model selection or conversational flows is dangerously incomplete. A realistic roadmap must allocate the dominant share of budget and timeline to:

• Data foundation and master data governance. Agentic AI’s effectiveness is directly proportional to the trustworthiness and completeness of underlying data. For many SAP landscapes, years of customizations, inconsistent master data, and fragmented analytics environments make this the single most expensive and time-consuming component.
• Process redesign. Agents do not simply automate existing tasks; they require rethinking how decisions flow across functions. This is the same kind of cross-functional, political work that made ERP implementations in the 1990s and 2000s so challenging—only now with the added complexity of probabilistic outputs.
• Organizational change management. In my experience, the greatest risk factor in any technology adoption is cultural. If employees distrust agent recommendations, the solution fails regardless of technical elegance.

For S/4HANA migration business cases, agentic AI should not be sold as a short-term productivity silver bullet. Instead, forward-looking leaders are treating

References

• When the Vendor Says the Agent Is Only 5%: What Infor’s Hospitality Candor Reveals About the Agentic AI Gap
• SAP Integration Suite Help Portal

References

• SAP Integration Suite Help Portal
• When the Vendor Says the Agent Is Only 5%: What Infor’s Hospitality Candor Reveals About the Agentic AI Gap
• When the Vendor Says the Agent Is Only 5%: What Infor’s Hospitality Candor Reveals About the Agentic AI Gap