SAP recently announced it is moving away from its traditional per-user and subscription-based pricing model as AI agents begin to automate core enterprise workflows. In a March 18 interview with Bloomberg, CEO Christian Klein said the company will begin charging customers based on AI consumption, while also deploying new “forward deployed engineering” teams to build AI applications directly with clients.
The shift marks one of the most significant changes to SAP’s business model since its transition to cloud subscriptions. It also reflects broader pressure across the software industry as agentic AI reduces the number of human users interacting with enterprise systems.
Agentic AI Undermining Per-User Pricing
SAP’s pricing model has historically been tied to users—licenses scaled with the number of employees accessing the system. That logic weakens as AI agents begin to execute tasks autonomously.
“It would be foolish to still charge subscription base, because AI is so powerful that it will automate a lot of tasks,” Klein said.
As AI agents take over workflows in finance, supply chain, and procurement, fewer human users are required to complete the same volume of work. That directly undermines the economic foundation of per-seat pricing.
The move toward consumption-based pricing aligns SAP more closely with cloud infrastructure models, where usage, not access, determines cost.
Analysis
What this means: Per-user pricing is structurally incompatible with agent-driven ERP. As AI agents take over workflows, the number of human users declines, breaking the link between system access and pricing. SAP’s move to consumption-based pricing reflects a broader shift toward measuring software by the work it performs, not the number of people using it. ERP customers should model how agent adoption will change usage patterns and align licensing strategy with execution volume.
SAP’s AI Pivot Extends Beyond Pricing
The pricing shift is part of a wider transformation inside SAP. Klein described the AI pivot as a reinvention that will reshape how the company:
- allocates and rewards employees
- engages with customers
- generates revenue.
To support this, SAP is building a new unit with hundreds of employees focused on AI adoption and deploying consulting-style teams to work directly with customers on implementation.
These “forward deployed engineering” teams signal a shift toward a more services-heavy model, where SAP is not just delivering software, but co-developing AI-driven workflows within customer environments.
Consumption Pricing Creates New ROI Problem
The transition introduces new challenges for customers, particularly around predictability and cost control.
Bloomberg reports that customers are already struggling to forecast spending under consumption-based pricing, which lacks the predictability of subscription models.
Separate market coverage has also pointed to concerns around transparency in SAP’s “AI Units” pricing framework and the difficulty of linking usage-based costs to business outcomes.
That tension reflects a broader issue in enterprise AI commercialization.
As one LinkedIn commentary on the shift argues, pricing AI based on compute usage rather than business outcomes forces customers to solve the ROI equation themselves. In that model, organizations receive bills tied to technical metrics such as usage units or processing cycles without a clear connection to the value generated.
The critique highlights a growing gap between how AI is built and how it is sold: Vendors understand infrastructure costs, while customers evaluate outcomes.
Analysis
What this means: Aligning AI pricing with business value and ROI is a challenge. Moving to consumption-based pricing shifts accountability to measurable outcomes, not just system usage. If AI costs scale with activity, then value must scale with results, whether in efficiency gains, cost reduction, or throughput improvements. ERP leaders should define clear KPIs for each AI use case and track whether increased consumption is translating into measurable business impact.
Sponsor Industry‑Grade Research
Investor Pressure Speeding Up Shift
SAP’s pivot is also occurring under external pressure. The company has lost roughly a fifth of its market value this year, as investors reassess the long-term viability of SaaS models in an AI-driven environment.
At the same time, competition from generative AI providers such as Anthropic and OpenAI is raising questions about whether traditional enterprise software vendors can maintain their position as AI capabilities expand.
SAP is positioning its advantage around access to enterprise data and deep integration into customer workflows. Plus, it is now investing directly in master data infrastructure to operationalize that advantage. Klein emphasized that future differentiation will come from building agents tailored to specific business processes, not generic AI capabilities.
Data Architecture at the Center of SAP’s AI Model
SAP’s shift to consumption-based AI pricing is only viable if the underlying data foundation can support agent-driven execution at scale. To address this, the company is set to acquire Reltio, a cloud-native master data management (MDM) provider, to strengthen its AI-first data strategy and expand the capabilities of SAP Business Data Cloud.
The move targets one of the most persistent barriers to enterprise AI adoption: fragmented data across systems. Reltio’s platform unifies structured and unstructured data from SAP and non-SAP environments into a single, consistent “golden record,” using AI-based entity resolution to reconcile inconsistencies.
Once integrated, Reltio will become a core component of SAP’s Business Data Cloud, providing the context layer required for Joule and SAP’s emerging agent ecosystem. The goal is to reduce the lag between AI queries and actionable outputs, while enabling real-time, multi-agent workflows across heterogeneous environments.
The implication is architectural. If pricing is shifting toward consumption, and agents are executing workflows autonomously, then the system must ensure that every action is grounded in consistent, trusted data. Without that foundation, consumption-based pricing becomes difficult to justify, as outputs remain unreliable or disconnected from business reality.
Analysis
What this means: Consumption pricing only works if data is consistent and actionable. Usage-based models assume that AI outputs are reliable and directly tied to business processes. Without a unified data foundation, increased consumption simply scales inconsistent or low-quality outputs, making costs harder to justify. ERP customers should prioritize master data management and data integration across SAP and non-SAP systems to ensure AI-driven actions are grounded in trusted information.
Editor’s note: A version of this article was originally published by SAPinsider on April 9, 2026.
