Support AI List Price
The public price of a support AI product per seat, outcome, automated interaction, or another vendor-defined billing unit.
AI cost per correct resolution = total AI subscription and usage charges / customer tasks verified as correctly resolved by AISupport AI list price is not one metric; it is a family of incompatible billing models. Copilots are often priced per agent seat, autonomous agents per outcome or automated interaction, and evaluation or quality products by conversation volume. The vendor also defines what counts: a handoff may count as an outcome in one contract while another bills only a conversation that does not escalate.
Before comparing prices, map each product to the same production workload and define the unit in plain language. Model eligible conversations, expected automation, human rework, minimum commitments, included allowances, overages, and the base helpdesk plan required to buy the add-on. Then calculate cost per correctly resolved customer job — not cost per vendor-labelled outcome — because an inexpensive answer that creates a repeat contact is not inexpensive.
What it hides: unit pricing transfers measurement power to the vendor. Classification rules, silent resolution windows, bundled allowances, retries, handoffs, and contract minimums determine the bill. Public prices are useful for a market snapshot, but procurement needs an auditable event definition, sample invoice, dispute process, and reconciliation between billed outcomes and your own task-success data.