A person analyzing business data with colorful graphs on a tablet screen.

Photo: Jakub Zerdzicki / Pexels

BillingJune 24, 2026via CIO Dive

AI Is Pushing SaaS Toward Usage-Based Billing — and Quietly Multiplying Your Payment Failures

AI is repricing SaaS in real time, and consumption billing is winning. GitHub switched to token-based billing on June 1. Zendesk and Workday followed with their own model changes. The shift is rational for capturing AI's value — but it quietly raises your involuntary churn risk with every new charge.

Jun 1

GitHub switches to token-based billing

>50%

of tech execs expect usage revenue growth by 2027

3 vendors

GitHub, Zendesk, Workday all repriced in June

40%+

of SaaS spend shifting to usage-based by 2030 (Gartner)

What happened

In "What CIOs should know about AI-driven SaaS pricing changes," CIO Dive's Jen A. Miller (June 18, 2026) laid out a clear shift: vendors are moving away from per-person pricing toward charging partly or totally on consumption.

GitHub

Shifted to token-based billing on June 1, charging for input, output, and cached tokens.

Zendesk

Rolled out pricing structure changes as its AI features expanded.

Workday

Introduced a Flex Credits pricing model.

Many people look at AI as a technological upgrade. It's not just a technological upgrade. It's really a fundamental shift in how SaaS companies will operate.
Marko Markov, RSM analyst, in CIO Dive

The driver is structural. If your software automates a customer-service team, the seat count shrinks over time and per-seat pricing stops making sense. Vendors move to consumption or outcome-based metrics instead. Analysts at Gartner expect a large share of enterprise SaaS spend to shift toward usage-, agent-, or outcome-based pricing by 2030.

Why it matters

Per-seat pricing has one underrated virtue: it is predictable. One charge, same amount, same date each month. Usage-based pricing trades that predictability for variable charges that move with consumption — bigger invoices in heavy months, surprise spikes, and in many setups, more frequent billing events overall.

Every one of those billing events is another moment a card can decline. More charges, larger and more variable amounts, and customers who did not expect this month's total all add up to a higher surface area for payment failure. Usage pricing does not just change what you charge — it changes how many chances you give a card to say no.

What this means for SaaS founders

If you are adopting usage- or AI-token-based pricing, treat payment recovery as part of the rollout, not an afterthought. Two things compound under consumption billing:

  • Larger, variable charges fail more. A card that sails through a $49 flat fee may bounce on a $480 usage spike. The recovery opportunity grows with the invoice.
  • More billing cadence means more decline events. If you are charging on consumption cycles instead of a single monthly seat fee, you have multiplied the number of transactions that need to clear.

The defense is the same as ever, just more valuable now: detect the failed charge immediately, retry on smart timing, and reach the customer with a branded card-update prompt before the account lapses. Under usage pricing, recovering a single failed invoice can be worth far more than a flat subscription.

The bottom line

AI is repricing SaaS in real time, and consumption billing is winning. It is the right call for capturing value — but it quietly raises your involuntary churn risk with every extra charge. The founders who move to usage pricing and lock down payment recovery will keep the revenue. The ones who only do the first half will watch a growing slice of it decline at checkout.