The Evidence: A Renewal Quote That No Longer Counts Seats
Picture a five-person HR operations team opening a renewal quote and finding a line item that doesn't say "per employee per month." Instead it gestures at something squishier: the work the software's AI agents are supposed to absorb. That's the shape of the story circulating as of October 16, 2026. According to Google News aggregation, The Register has reported that Workday is pursuing a pricing approach for its AI agents that reaches toward the customer's labour budget — the money a company currently spends on the humans doing that work — rather than the familiar seat count.
Here is the honest caveat, and it matters more than usual: as of October 16, 2026, the specific mechanics of that pricing — rate cards, percentage take, contract floors, how "replaced" labour would even be measured — were not independently verifiable from multiple outlets at the time of writing. The Register's coverage is the anchor; corroborating detail from other business-press outlets was not available to cross-check. So treat the direction as reported and the numbers as unpublished. Any article quoting you a precise per-agent dollar figure today is filling in blanks nobody has confirmed.
The demo is not the product, and the press cycle is not the contract. But the direction is the news, and it's a bigger deal than a pricing tweak.
The Job You're Actually Hiring an AI Agent To Do
Strip away the vendor language and ask the Christensen question: what job does a buyer hire an HR or finance AI agent to do? Not "be intelligent." The job is closing a queue. Someone has to reconcile the expense report, chase the missing timesheet approval, answer the 400th question about parental leave policy, re-key a req into the ATS. These are high-volume, low-judgment tasks that currently sit inside somebody's salary.
That framing is exactly why labour-budget pricing is coherent rather than cynical. Seat-based pricing was always a proxy for value — more humans logging in meant more value extracted. If the agent removes the human from the loop, the proxy inverts: the more successful the automation, the fewer seats the vendor can bill. No software company builds a growth plan on a meter that runs backwards. So the meter has to move from "who logs in" to "what got done."
The non-obvious part the surface coverage mostly skips: this is not primarily an AI story. It's a revenue-architecture story that AI forced into the open. Consumption pricing, credit bundles, and outcome fees have been creeping across enterprise business tools for years. Agents just made seat pricing visibly incompatible with the product's own value proposition.
Photo by Jakub Żerdzicki on Unsplash
Seat Pricing vs Labour-Budget Pricing: Who Wins Under Which Condition
This is the comparison no single news write-up gives you, because it requires sitting on the buyer's side of the table.
Under seat pricing, cost is predictable and decoupled from volume. A 200-person company pays roughly the same whether its HR queue handles 300 tickets a month or 3,000. Budgeting is easy; waste is invisible. The buyer captures 100% of any productivity gain, which is precisely the arrangement vendors are now trying to renegotiate.
Under labour-budget or outcome pricing, cost tracks work volume. Seasonal businesses, companies in hiring freezes, and teams whose ticket volume is genuinely shrinking do better — you stop paying for capacity you don't use. Fast-growing companies and high-churn workforces do worse, because the meter accelerates exactly when headcount pressure is already peaking.
Run the arithmetic in purely illustrative terms (these inputs are hypothetical, not a Workday quote, since no rate card has been published): suppose a shared-services pod costs a company $400,000 a year fully loaded, and agents credibly absorb a quarter of that workload. The buyer's gross saving is $100,000. If a vendor prices the agents at 30% of displaced labour, the software takes $30,000 and the buyer keeps $70,000 — a 70/30 split of a gain the buyer used to keep entirely. At 50%, the split is even, and the "ROI case" collapses into a wash once you add integration and oversight time. The percentage take, whenever it surfaces, is the whole story. Everything else is packaging.
The skeptic's pushback deserves airtime: outcome pricing can be fairer. Buyers have complained for a decade about paying full price for shelfware nobody opens. Pay-for-work-completed is, in principle, the model that punishes vendors whose agents don't perform. Our read is that the principle is sound and the measurement is where it goes wrong — because the vendor usually owns the meter.
The Switching Cost Nobody Prices
Here's the second-order consequence that rarely makes the headline. Labour-budget pricing requires a shared definition of "work the agent replaced." That definition lives in the vendor's telemetry. The buyer is then in the position of auditing an invoice computed from data it does not control and often cannot fully export — the data export reality that bites every HR and finance platform migration.
Core HR and payroll suites are already among the stickiest software a company buys; multi-year migrations, historical records, and compliance filings make leaving expensive before any AI is involved. Bolt on a pricing model whose inputs are vendor-measured, and the negotiating position at renewal narrows further. The moment you outgrow seat pricing is also the moment you lose the simplest tool you had for comparing quotes across suites.
Agents touching systems of record raise a governance question too, which AI Agents News Lens examined in the context of write-permission protocols for journal entries: an agent allowed to post, approve, or close records needs a different audit trail than one that only reads and drafts. Pricing those two capabilities identically would be a mistake in either direction.
Three things worth doing before the next renewal, whatever the final model looks like. First, instrument your own baseline now — ticket volumes, cycle times, hours per process — because if you don't bring your own numbers, you will negotiate against the vendor's. Second, insist that any outcome-linked fee comes with a measurement definition in the contract and a cap, not just a rate. Third, ask the unglamorous workflow automation question: what does the agent do when it's wrong, who reviews it, and is that reviewer's time counted anywhere in the savings math? It usually isn't.
Frequently Asked Questions
Is outcome-based AI pricing cheaper than per-seat pricing for a small business?
It depends entirely on your work volume relative to headcount. Teams with flat or shrinking transaction volume generally benefit from consumption-style pricing, while growing teams can see costs rise faster than under a fixed per-seat plan. As of October 16, 2026, no published Workday rate card was available to compare directly, so the only reliable method is modelling both structures against your own twelve-month volume data.
How do you verify that an AI agent actually replaced labour hours?
Through a baseline you collect yourself before deployment: tickets closed per week, average handling time, and the share of exceptions routed back to a human. Vendor dashboards typically count tasks the agent attempted, which is not the same as work a person no longer has to do. Review-and-correction time is the line item most often omitted from savings calculations.
Should small teams wait before signing an AI agent contract tied to labour spend?
Adopt now if your process volume is high, repetitive, and measurable, and if you can negotiate a hard spending cap. Wait if your workflows still change month to month, if your data lives across several disconnected business tools, or if the contract asks you to accept vendor-side metering without an audit clause. On balance, the pragmatic move in late 2026 is a scoped pilot on one process with a fixed fee, not a suite-wide outcome agreement.
- As of October 16, 2026, The Register's reporting — surfaced via Google News — indicates Workday is pushing AI agent pricing toward the customer's labour budget; the underlying rates and mechanics remain unpublished and were not cross-verifiable across outlets at the time of writing.
- Seat pricing breaks structurally when software removes seats. Expect every major HR, finance, and productivity software vendor to test some version of consumption or outcome pricing, not just one company.
- The negotiable variable is the vendor's share of the savings. A 30% take leaves most of the gain with the buyer; a 50% take makes the business case roughly neutral once oversight costs are included.
- Our analysis: the most likely near-term outcome is hybrid pricing — a platform fee plus metered agent usage with caps — because pure outcome pricing requires an agreed measurement standard that does not yet exist. Buyers who build their own baseline metrics this quarter will negotiate from evidence; everyone else will negotiate from the vendor's dashboard.
Disclaimer: This article is editorial commentary based on publicly reported information and does not involve independent product testing. Tool features, pricing models, and contract terms change frequently; verify current details with the vendor before making purchasing decisions. Research based on publicly available sources current as of October 16, 2026.