Original OrbTrail editorial, written in our own language from the reference publication.
AI agents are challenging one of enterprise software's oldest rules: charging by user count. When technology moves from assisting a person to completing work, seat licenses and consumption metrics reveal little about the value produced.
Sierra, founded by former Salesforce co-CEO Bret Taylor, has placed outcome pricing at the center of its offer: customers pay when an agent completes a task. Salesforce is moving in a similar direction across parts of its portfolio. Flex Credits connect cost to actions, while Help Agent charges for issues resolved without human intervention.
The model makes ROI easier to discuss, but shifts complexity into the definition of success. A closed conversation does not necessarily mean a satisfied customer, while a task that reaches 90% before escalation may still create value. Contracts, telemetry and governance must separate activity, resolution and quality.
A mix of models is the most likely destination. Outcome pricing fits processes with verifiable endings, such as service requests and returns. Sales, marketing and productivity use cases may continue to rely on credits or licenses. Buyers should ask not only what AI consumes, but which results can be measured and audited.
