On SAP’s third-quarter call, chief financial officer Dominik Asam made a remark aimed at the consulting industry. In knowledge work, he said, business models that rely only on selling more hours will increasingly be challenged by AI. SAP’s own services revenue fell 5 percent at constant currency in the quarter, and he said further portfolio adaptation in services is a necessity.
For customers who buy implementation and migration work, this is not just a comment about SAP’s accounts. It signals that the way SAP and its partners price and deliver projects is going to change, and that buyers can use the shift to their advantage.
What SAP is doing
Asam said SAP wants to use services more intentionally to promote the adoption of its AI solutions and build critical mass. The approach is producing a positive return, he said, so SAP will continue to invest in adoption support for cloud and AI customers, which will probably put further downward pressure on billable hours and services revenue next year. He described the logic as converting a low-margin hour into recurring, higher-margin AI and data business.
Klein described a related move. SAP is evolving its consulting services into what he called an AI factory to accelerate adoption, and he said systems integrators are helping SAP disrupt their own business model by speeding the migration of customers to the cloud. He also said SAP’s migration agents and Joule for Consultants are getting strong feedback.
What this means for what you buy
The change is already being discussed with customers. A recap of the DSAG congress reports that SAP representatives said hourly billing should not be assumed to remain a central element, because AI tools can reduce the hours needed. If the work takes fewer hours, the question is who benefits from the saving.
The answer is a matter of contract. If a partner is paid by the hour and uses AI to work faster, the customer pays less only if the rate card or the estimate reflects it. If the project is fixed price, the partner keeps the saving unless the price was set with the tools in mind. Neither outcome is automatic, and it is worth deciding in advance which you want.
Ask how the work will be done
When you request a proposal, ask which AI tools the partner will use, for which tasks, and what difference they expect them to make to effort and timeline. Ask what evidence they have from similar projects. Ask whether the estimate has been reduced to reflect those tools, and if not, why not.
Ask who pays for the tools. Some, such as SAP’s assistants and agents, may be metered or priced separately. SAP’s pricing guidance describes a per-user plan for Joule for Consultants and consumption pricing in AI Units for agents. A statement of work that does not say who bears those costs can produce a surprise during delivery.
Fixed price and outcome-based options
As hours become a weaker measure of value, fixed-price and outcome-based structures become more attractive. A fixed price shifts delivery risk to the partner and gives you budget certainty. An outcome-based structure, where part of the fee depends on a result such as a go-live date or a data quality target, aligns incentives more closely. Both need clear definitions of scope, acceptance criteria and change control, or they become disputes.
Be careful with any model where the partner takes the risk but you take the dependency. If a partner’s fixed price relies on a particular SAP tool, check that you are not also committing to buy that tool, or related products, as a hidden condition.
Prerequisites and bundling
DSAG raised this at the congress. It says SAP’s Migration and Modernization Assistants should be offered cost-effectively and without mandatory prerequisites such as Signavio or Tricentis. If a proposal assumes you will use those products, ask whether they are optional, what they cost and whether the project plan still works without them.
Check also that the proposal does not assume you will buy more AI capacity than you need. Tools that speed up delivery can consume AI Units, and a project plan that depends on them needs a budget line for the units.
Contract wording worth considering
A few clauses help. State which tools and accelerators the partner will use, and who holds the licence for each. Say whether any productivity gain from AI is reflected in the price, and how. Define acceptance criteria and what happens if the work finishes early. Allow for the partner to propose a cheaper method as the tools improve, so that you can benefit during a long project. Make sure the statement of work lists any metered usage, such as AI Units, and who pays for it.
Finally, keep the knowledge. If AI tools do much of the work, the learning about your landscape may sit in prompts, configuration and logs rather than in people. Ask for those to be handed over at the end, so that you are not dependent on the partner for every future change.
Conclusion
SAP’s finance chief has said openly that selling hours is a business model under pressure, and SAP’s own services revenue is falling as it shifts towards adoption and AI. Customers who buy implementation work should expect proposals to change, and should shape the change rather than wait for it.
Ask partners how AI will change effort, require the savings to show in the price, make tool costs explicit, and look closely at fixed-price and outcome-based structures. The shift away from hours gives buyers new levers, provided they use them.
