AI Pricing Models: Enterprise ROI vs. Vendor Revenue Misalignment
Summary
- • Enterprise IT leaders argue no current pricing model — per-token or per-task — aligns AI cost with business value
- • SAP is pushing per-AI-task pricing as a token alternative; observers say it fails the same ROI alignment test
- • Both sides face a timing mismatch: pricing is locked in before projects start but AI value only emerges after deployment
- • No AI vendor will accept outcome-based commission-style pricing because it transfers financial risk to the vendor
Details
Current AI pricing models are structurally incompatible with enterprise ROI requirements
No existing model — per-token, per-task, or otherwise — satisfies the enterprise need to pay for realized business value rather than resource consumption or activity.
SAP is pushing a per-AI-task pricing model as an alternative to per-token billing
SAP's model charges for each completed AI task. Industry observers argue it does not solve the core misalignment because it still ties payment to activity rather than verified business outcomes.
SAP executive: pricing and value realization are sequentially misaligned in AI projects
Irfan Khan, president of SAP Data & Analytics: 'the day one cost is not necessarily the day one value.' Enterprises must commit to pricing before they have reliable data on what the AI will actually deliver.
Consulting CEO: AI is being priced like infrastructure when it behaves more like labor augmentation
Justin Greis, CEO of Acceligence: token pricing is like paying a worker by the keystroke — measurable for the vendor, but disconnected from value for the buyer.
Enterprise buyers want outcome-based pricing; no AI vendor is willing to offer it
Vendors reject commission-style pricing because it transfers financial risk to their side, making the model commercially unviable under current conditions.
Line-of-business workers experimenting independently complicate enterprise pricing governance
AI adoption is no longer driven solely by IT departments — business unit employees and partners run their own experiments, making it harder for IT leaders to project usage volumes or plan pricing negotiations centrally.
Insight = attributed argument or analysis from a named source, Industry Update = concrete business or product move, Strategy = positioning or commercial decision, Context = background condition shaping the story
What This Means
The fundamental challenge described is a structural pricing mismatch: enterprise buyers want to pay for realized business value, while AI vendors can only price on resource consumption and platform utilization. As agentic AI moves into production, neither side has a reliable mechanism to price against outcomes before they are known. The article argues this tension will persist until AI deployments generate enough performance data to enable outcome-oriented models, or until vendors find ways to share risk with customers.