Usage-based billing for AI companies: what your billing system must do

AI products rarely sell a flat seat price. Their costs follow consumption (model calls, tokens, compute) and so does their pricing: credits, usage tiers, outcome-based fees, often on top of a platform subscription. That puts billing at the center of the business model, and it exposes the limits of billing systems designed for seats.

How AI companies price

The common models, often combined:

• Pure usage: a price per unit (token, API call, minute, document).
• Prepaid credits: the customer buys a credit balance, drawn down by usage, with top-ups and expiry.
• Hybrid: a platform fee or minimum commitment, plus usage above an included allowance.
• Outcome-based: a fee per resolved ticket, qualified lead or completed task.
• Seats plus usage: per-user pricing with a usage component for heavy users.

Each model changes in months, not years. The billing system has to let finance change pricing without an engineering project.

The capabilities that matter

Usage metering

Ingesting usage events from the product, deduplicating them and aggregating them by customer and period. Volume matters: an AI product can send millions of events a day. 38 of the 44 tools we score document usage metering.

Usage rating

Applying the price to the metered usage: per unit, tiered, volume, with included allowances. 42 tools document it.

Prepaid credits

Credit balances, drawdown, top-ups, expiry and the revenue treatment of unused credits. 33 tools document prepaid credits.

Hybrid pricing

A subscription and usage on the same contract and the same invoice, with minimum commitments and overages. 39 tools document hybrid pricing.

Real-time visibility

Customers want to see their consumption before the invoice; sales wants to see who is about to exceed their commitment. Look for a customer portal with live usage and alerts on thresholds.

Revenue recognition for usage

Usage revenue is recognized as consumed; prepaid credits are deferred until used or expired. If the billing system does not produce these schedules, finance rebuilds them every month. See ASC 606 revenue recognition for SaaS.

The tools that document all four

Usage metering, usage rating, prepaid credits and hybrid pricing together, in ranking order:

Usage-first tools (Metronome, Orb, m3ter, Lago) are strong on metering and rating; the difference with full quote-to-cash platforms is usually upstream (quoting a usage contract, ramps, approvals) and downstream (collections, revenue recognition). Hyperline, for example, documents usage metering, rating, prepaid credits and hybrid pricing as core strengths, along with native quoting and ASC 606 revenue recognition.

Questions to ask a vendor

1. How many usage events per second can you ingest, and how late can an event arrive and still be billed?
2. Can finance launch a new usage price without engineering?
3. Show a customer with a minimum commitment, prepaid credits and overage on one invoice.
4. How are unused credits treated at expiry, in billing and in revenue?
5. Can a salesperson quote a usage contract, with ramps, from the CRM?

FAQ

What is usage-based billing?

Usage-based billing charges customers for what they consume (API calls, tokens, minutes, transactions) instead of, or on top of, a fixed subscription. The billing system meters the usage, rates it against the price and invoices the period.

What is the best billing software for AI companies?

It depends on the pricing model and on how much of quote-to-cash you want in one tool. The table above lists the tools that document usage metering, rating, prepaid credits and hybrid pricing, in the order of our quote-to-cash ranking.

Should AI companies use credits or pay-as-you-go?

Credits give revenue predictability and upfront cash; pay-as-you-go lowers the barrier to adoption. Many AI companies combine them: a prepaid credit commitment with overage billed in arrears. The billing system has to support both on the same contract.