Seat-based SaaS has a stable cost to serve. AI products don't — cost moves with model prices, token volume, and how much work an agent does per request. That means metering every request and repricing without a re-implementation.

Most AI companies charge on one of four units: tokens consumed, agent runs, outcomes delivered, or a credit that abstracts all three. Orb supports all four, plus hybrids with seats and platform fees. Learn more here.

Prepaid credits let customers buy a balance up front and draw it down as they use your product. AI companies use them to collect cash before compute costs land, cap exposure to runaway usage, and give customers one simple unit instead of token math.

Threshold billing invoices a customer as soon as their usage crosses a set dollar amount, instead of waiting for month end. It's the standard defense for self-serve AI products, where a stolen card can generate five figures of inference cost overnight.

Yes. Orb ingests raw usage events once and applies pricing as a layer on top, so changing a rate, adding a tier, or redefining a billable metric is a configuration change rather than an engineering project.

Protecting margin takes per-customer visibility into where cost is outrunning price, and a pricing system fast enough to act on it. Orb provides real-time consumption and revenue data, plus price changes that can be simulated against historical usage before shipping.

Yes. Most AI companies land on a hybrid model: a platform fee or seat license for predictable revenue, plus usage or credits for consumption. Orb handles seats, usage, credits, minimums, and commitments on one subscription and one invoice. Learn more here.