Four companies. Four different software categories. One month. Salesforce declared their API is now the UI. ServiceNow confirmed that over half of their new business is already non-seat-based. GitHub announced Copilot moves to token-based credit billing on June 1st. And OpenAI’s Nick Turley said publicly that an unlimited AI plan is like an unlimited electricity plan — it simply doesn’t make sense.
This didn’t happen in a strategy deck or an analyst report. It happened on main stages, earnings calls, and product blogs in April 2026. The per-seat era didn’t die quietly. It announced its own retirement.
Why Per-Seat Is Breaking — And Why It Won’t Come Back
Agents Don’t Sit in Seats — Your Pricing Model Shouldn’t Either
Per-seat pricing was never really about seats. It was about a reliable proxy — one human, one login, one predictable unit of consumption. For thirty years that logic held.
Then AI agents arrived. An agent doesn’t sit in a seat. It calls an API, runs a multi-step workflow, burns a million tokens in a session that used to take a human three hours. GitHub’s own Chief Product Officer put it plainly: “Copilot is not the same product it was a year ago. It has evolved from an in-editor assistant into an agentic platform. Agentic usage is becoming the default.”
When the product changes that fundamentally, the pricing model built around the old product cannot survive. And this is structural, not cyclical. When AI agents replace human roles — as ServiceNow’s new Prime tier explicitly enables — the number of seats in an organisation shrinks while value delivered by the software goes up. Charging less for more value delivered is not a sustainable business model.
The direction of travel is clear: per-seat gives way to hybrid, hybrid gives way to usage and outcome-based models, and the long-run destination is utility pricing. Software billed like electricity. You pay for what flows through the pipe. No public company will get there overnight — the revenue optics are too visible, analysts rebuild their models, NRR changes definition. But the transition is underway and it is not reversing.
What SaaS Companies Actually Need Right Now
The companies navigating this transition well are not the ones who picked the right pricing model. They are the ones running experiments before making announcements. In a period of genuine pricing model uncertainty, data-driven price discovery is the only responsible approach.
That means:
- A/B testing pricing structures across customer cohorts before committing publicly
- Segmenting customers by actual usage patterns rather than assumed personas — consumption signals that SaaS companies never had to track in a per-seat world
- Measuring customer sentiment toward different pricing options before they become policy
- Bundling products flexibly to smooth the transition for existing customers while opening new monetisation lanes
- Running multiple pricing models simultaneously across different segments — per-seat for established enterprise accounts, consumption-based for new PLG-driven cohorts, outcome-based for customers who came in expecting to pay for results
- Supporting PLG and SLG motions concurrently — new AI-native companies use product-led growth as price discovery. Usage data tells them what to charge before they formalise a pricing structure. That requires billing infrastructure that can meter, capture, and report on consumption from day one
The billing platform underneath all of this is no longer a back-office detail. In a period of pricing model transition, it becomes a strategic asset.
OneBill: Built for This
Usage-based billing is not a new concept for us. Our telecom and utility customers have operated on consumption-based models for decades — B2B reliability at B2C scale, complex rating engines, real-time metering, revenue recognition that doesn’t break when pricing models change mid-contract. We built OneBill around that reality.
Our Q2C2C framework — Quote to Cash to Care — means every layer of the revenue lifecycle plays a role in pricing experimentation:
- Quote (CPQ): Where pricing experimentation begins. Configure and price bundles dynamically, A/B test structures at the quote stage before they reach billing
- Cash (Billing, Usage Metering, RevRec): Runs per-seat, usage-based, outcome-based, token-based, tiered, and hybrid models simultaneously without breaking revenue recognition
- Care (Help Desk & Ticketing): The feedback loop most companies overlook. Support patterns and friction signals around pricing changes are real-time inputs for what is working and what is not
We work with new-age software companies and neo datacenters — including stealth startups that came into existence already thinking in consumption tiers and outcomes rather than seats. And increasingly, established SaaS companies are coming to us to run pricing experiments, manage model migrations, and give finance the data to decide before they have to announce anything publicly.
If you are a SaaS company figuring out what comes after the seat, let’s talk.