Per-Seat Pricing

How it works

Per-seat pricing was the dominant model for SaaS support tools through the 2010s (Zendesk, Intercom, Gorgias). It made sense when software cost scaled with server load — more users meant more compute. In the AI era, it is increasingly misaligned: an AI agent does the work of 10 human agents but consumes one seat's worth of compute cost.

For support teams, per-seat creates compounding problems. Hiring a seasonal agent for BFCM requires buying a new seat for three months, then downgrading. Onboarding a new team member mid-quarter triggers a prorated charge. Large teams pay enterprise rates while small teams subsidize the vendor's infrastructure at a premium per-seat cost.

Per-seat pricing also misaligns incentives: vendors benefit when you hire more agents, not when the AI deflects more tickets. This creates a structural tension between the vendor's revenue model and your goal of reducing support headcount through automation.

When you'll encounter it

Per-seat pricing is front-and-center on every Zendesk, Gorgias, and Intercom pricing page. You'll encounter it in budget planning, vendor comparisons, and annual contract renewals. Watch for 'agent seats,' 'user licenses,' or 'team member accounts' as synonyms.