SAP reshuffles spending and research to chase agentic AI returns

SAP is concentrating new hiring exclusively on core AI roles and suspending non-AI business travel, according to The Register's reporting on an internal company memo. The move, confirmed by an SAP spokesperson, is a direct bid to sustain AI investment as competition in the enterprise application market intensifies.

The company's statement, shared with The Register, described the approach as 'applying greater discipline to hiring, external spending, and internal travel' while keeping customer-facing activities and critical AI initiatives fully funded. The reallocation signals how seriously SAP is treating AI not just as a product line but as its primary cost-of-competitiveness expense.

ROI numbers are climbing, but agentic AI is the real bet

The internal spending shift arrives alongside new data that supports the underlying logic. The Value of AI Report 2026, produced by SAP and Oxford Economics from a survey of 2,600 business leaders across 13 countries, found that average global AI spending rose modestly to $28 million this year, up from $26.7 million in 2025, according to SAP News. The more striking change is on the returns side.

Enterprise AI ROI climbed to 21% in 2026 from 16% the prior year, translating to roughly $6.3 million on that average investment base. That figure is expected to reach 38%, or approximately $15.9 million, within two years, according to the same report.

Agentic AI is where expectations are most aggressive. The report found that average ROI from agentic AI is projected to reach $17.6 million within two years, more than quadrupling from last year's estimate of $4.3 million, per SAP News. That growth rate helps explain both SAP's product moves and its budget reorientation.

On the product side, SAP introduced the Autonomous Enterprise concept in May, backed by a new SAP Business AI Platform designed to ground AI in 'real business context' drawn from its ERP, CRM, and HCM systems. The Register reported that Joule Studio 2.0, part of that launch, gives developers tools to build and manage AI agents with native support for Model Context Protocol and Agent2Agent, enabling interoperability with third-party tools and data sources.

Broad optimism, narrow readiness

Despite the improving ROI picture, the Oxford Economics survey reveals a readiness gap that procurement and IT operations leaders should take seriously. More than eight in ten respondents (83%) told researchers that agentic AI has moderate to very high potential to change their organization, yet only 3% described themselves as fully prepared for it, according to SAP News.

The broader AI adoption pattern also shows room to mature. About 30% of tasks in a typical business are currently AI-supported, a share the report expects to rise to 48% within two years. Yet 41% of organizations still rely on piecemeal AI approaches rather than strategic deployment, and fewer than half have a dedicated AI leader, clear development frameworks, or formal AI training programs.

Data quality and governance are the operational blockers

Data quality stands out as the biggest single barrier. The report found that 73% of companies report challenges with incomplete data, down from last year's figure, and 79% say they have experienced rework, delays, or backlogs caused by low-quality AI outputs, according to SAP News. Those are not abstract risks: they translate directly to productivity losses and project delays that operations and IT teams absorb daily.

Workforce readiness is similarly strained. Nearly 78% of respondents said their organization's upskilling efforts are either not keeping pace with AI tool evolution or they are unsure. Shadow AI use is also growing, with 69% reporting it happens at least occasionally, raising compliance and security exposure for governance teams.

Governance infrastructure remains thin across the board. Only 12% of businesses say their skills or frameworks are fully ready to govern AI, and 38% have no human-in-the-loop process in place for agentic workflows, per the SAP and Oxford Economics report. As agents begin executing multi-step business processes autonomously, the absence of access controls and oversight structures becomes a material operational risk, not a future concern.

Context on SAP's broader position

SAP's push into AI comes as the company navigates a slower-than-expected cloud migration among its customer base. The Register noted that on-premises software support revenue came in at €10.5 billion for full-year 2025, down 7% from 2024's €11.29 billion but still roughly €2 billion above the target SAP set in 2022 for that year. The persistent on-prem revenue base gives SAP both the financial runway and the installed base incentive to bring AI capabilities to ERP and S/4HANA environments, including on-premises deployments, rather than waiting for a full cloud transition.

What this means for your team

• Audit your data infrastructure now. With 79% of enterprises experiencing AI-related rework or delays due to poor data quality, IT and operations leaders should assess data completeness and pipeline reliability before expanding AI workloads.
• Map governance gaps before deploying agents. Only 12% of organizations have fully ready governance frameworks, and 38% lack human-in-the-loop controls for agentic workflows. Establishing access controls, approval gates, and audit trails is a prerequisite, not an afterthought.
• Evaluate SAP's Joule Studio 2.0 and Business AI Platform for ERP-embedded AI use cases, particularly if your organization runs S/4HANA or ECC on-premises. SAP has signaled it will bring AI features to those environments, reducing the urgency of a cloud migration for AI access.
• Pressure-test your upskilling roadmap. With 78% of companies saying internal training is lagging behind AI tool evolution, procurement and HR leaders should review whether current programs cover agentic AI concepts, shadow AI risks, and governance accountability.