AI pricing is moving from seats to usage and outcomes

SaaS vendors are starting to rethink AI pricing around usage, outcomes, and agent activity instead of simple seat counts. CIO Dive reports that vendors including GitHub, Zendesk, and Workday are moving in that direction, while GitHub's internal analytics-agent case study shows why the shift matters: AI value is increasingly tied to work performed, not merely access granted. For buyers, the story is a budget-control warning. AI can raise software value, but it can also make software bills less predictable if contracts stay vague about consumption, automation volume, and outcome measurement.