By mid-2025, AI business cases were colliding with new platform pricing, model consumption and the ongoing cost of evaluation. A simple labor-savings estimate could not represent the economics of a living AI service.
A useful model compares the complete cost and value of a business outcome across current work, deterministic automation and agentic assistance.
Cloud Group point of view
Model the marginal outcome, then stress-test adoption and quality. Enterprise AI economics improve through better process design, reuse and learning—not through optimistic token assumptions.
A practical playbook
The strongest next step is narrow enough to govern and useful enough to produce evidence. We would structure the work around these moves:
- Baseline volume, cycle time, error, waiting and fully loaded labor.
- Estimate platform, model, integration, supervision and improvement costs.
- Separate capacity created from cash actually removed or revenue actually added.
- Model low, expected and high adoption with quality thresholds.
- Fund reusable data and action capabilities across the portfolio.
The architecture and operating implication
Instrument each use case so consumption and human effort can be tied to a Salesforce outcome. Shared identity, retrieval, action and evaluation services should be treated as portfolio assets, with their cost allocated transparently rather than charged repeatedly to each pilot.
Measure what changes
Model activity is not a business result. Track a small set of indicators that connect behavior to accountable work:
- Total cost per outcome before and after
- Value realized per unit of consumption
- Reuse contribution from shared capabilities
- Payback sensitivity to quality and adoption
The purpose of unit economics is not to make every initiative look inexpensive. It is to reveal which design creates durable operating leverage.
Primary sources
This field note is grounded in the product and market context available at the time of publication.
