European software giant SAP is going to radically overhaul its commercial model. Instead of billing based on the number of users, the foundation of the SaaS industry for the past two decades, the company wants to switch to a model where the actual consumption of AI services determines the bill. It is one of the most far-reaching strategic shifts the Walldorf-based company has ever announced.

Why SAP is changing course

The logic behind the decision is clear. SAP executive Christian Klein states that the traditional subscription model is losing its viability now that artificial intelligence plays an increasingly large role in business processes. AI agents are taking over tasks that were previously performed by employees. This eliminates the basis for a pricing model that relies on counting active users.

Klein formulated it succinctly in an interview with Bloomberg in Berlin: it would be illogical to stick to fixed subscription costs now that AI can automate so many tasks. The CEO views the switch not as an adjustment, but as a fundamental reinvention of how SAP delivers value, rewards employees, and generates revenue.

An additional factor is that the adoption of SAP's own AI tools has so far lagged behind expectations. The commercial reorientation therefore goes hand in hand with a new organizational approach to accelerate that adoption.

Not a sudden switch, but a phased transition

Anyone expecting SAP to abandon the subscription model immediately will be disappointed. The transition is phased and has already been set in motion. In the summer of 2025, SAP already extensively restructured its cloud ERP packages: tools such as the Joule AI Assistant and the SAP Datasphere analytics platform were removed from standard bundles and transformed into individually purchaseable services.

What CEO Klein is announcing now is the next chapter: as customers adopt more of the new AI functionalities, billing will shift increasingly towards consumption-based rates. The classic subscription structure will not disappear overnight, but will gradually lose its dominant position.

The new pricing system in practice

SAP uses so-called AI Units as a unit of payment for its AI services. Customers purchase a package of Units in advance, which is subsequently accessed as they use AI functions. The number of Units a specific action costs varies by service and application.

According to information on the SAP website, AI Units are typically calculated on a per-user-per-month basis. Joule Premium packages, available for domains such as supply chain, finance, and HR, feature tiered volume-based pricing. The minimum entry level is 100 AI Units at a list price of approximately 700 euros per year.

For applications where a per-user model does not apply, such as Document Grounding, consumption-based billing applies per processed record. Customers can track their consumption and remaining balance in real time via the SAP for Me portal.

The basic version of Joule is included in all standard SAP applications. Premium scenarios with demonstrable added value, such as automatic processing of payment notices or scanning of freight documents, are billed separately.

Forward-deployed engineering: SAP on-site at the customer

In addition to the price change, Klein announces a new organizational structure. Starting in July 2026, SAP will establish so-called 'forward-deployed engineering' teams. These groups consist of consultants and developers with industry-specific expertise and will be deployed on-site at customers to build custom solutions.

Klein illustrates the idea with a concrete example: anyone wanting to improve delivery times must literally accompany the truck drivers to understand their daily working methods and translate those insights into a solution. This approach breaks with the traditional consultancy mode, where consultants work remotely based on generic best practices.

The fact that the company is now embracing this approach on a large scale underscores the urgency with which Klein wants to accelerate the AI ​​transition.

Joule: the AI ​​copilot as the centerpiece of the strategy

SAP's strategic shift revolves largely around a single product: Joule, the AI ​​assistant that is being increasingly integrated into SAP's application landscape. At SAP Connect 2025, the company presented fourteen new Joule Agents for domains such as finance, HR, procurement, and supply chain.

For example, a Cash Management Agent can independently process daily bank statements and perform reconciliations, while a Bid Analysis Agent compares supplier bids based on price, delivery time, and payment terms. Joule Studio, the development environment that allows organizations to build their own Joule agents, became generally available in the first quarter of 2026. According to SAP, there are now more than four hundred AI-driven use cases spread across its application portfolio.

Critical remarks: transparency remains a challenge

Not all reactions to the new model are positive. Independent consultants and procurement experts point out a structural pain point: the complexity of SAP’s pricing structure makes it difficult for customers to oversee and budget their total costs.

An additional point of attention is that unused AI Units expire annually without being carried over to the next year. This forces organizations to make accurate consumption forecasts, a discipline that is new to many procurement teams. Analysts advise explicitly requesting transparency regarding conversion methods, price caps, and overrun costs during contract negotiations.

Usage-based pricing requires a new internal discipline: consumption-based licenses only work well if an organization has a clear understanding of its own usage. Without that foundation, budgeting becomes guesswork.

The broader context: a sector in motion

SAP's change of course is not an isolated phenomenon. Several major SaaS providers are struggling with the question of how to monetize AI value within a pricing model that was originally designed for a world without autonomous agents. The fear that AI is structurally undermining the per-user model is weighing on the stock prices of several software companies.

SAP itself saw its share price come under pressure in recent months. Nevertheless, the vast majority of analysts following the company remain optimistic. SAP’s cloud revenue for fiscal 2025 amounted to approximately 21 billion euros, a growth of 23 percent year-on-year. The Cloud ERP Suite grew by 28 percent.

What does this mean for your organization?

For organizations working with SAP or considering it, this announcement marks a turning point. The direction is clear: those who use more AI pay more. This calls for a different approach to budgeting, contract formation, and usage monitoring.

  • Map out your expected AI usage before entering into or renewing contracts.
  • During negotiations, request clear conversion tables per AI function and establish price caps contractually.
  • Consider bundling AI Units in a broader RISE with SAP or S/4HANA negotiation.
  • Activate monitoring via SAP for Me and assign an internal person responsible for tracking AI usage.
  • Take into account the annual expiration of unused units; align your purchasing volume with realistic expected usage.

The transition SAP is now initiating is not a technical curiosity. It is a fundamental shift in the relationship between software vendor and customer, and preparation for it begins now.