Workday

What it is

Workday is the HR system of record at most large enterprises, and its talent acquisition suite is what happens when that record decides to run hiring too. Three pieces carry it. Workday Recruiting and Candidate Engagement is the ATS. The Recruiting Agent, powered by HiredScore and acquired in 2024, does sourcing, screening and candidate matching. The Candidate Experience Agent, powered by Paradox and closed October 1, 2025, runs conversational apply and self-scheduling. Around all three sits Workday Illuminate, the agent layer, and the Agent System of Record, which registers and governs agents — Workday’s own and third-party ones.

The company closed fiscal 2026 on January 31, 2026 at $9.552B total revenue, up 13.1%, with subscription revenue at $8.833B, up 14.5%. Fiscal 2027 subscription guidance is $9.925–9.950B, 12–13% growth. That deceleration is the commercial context for everything below.

The argument is the suite, not the ATS

Workday Recruiting does not beat Greenhouse at structured hiring, and no honest reference call will claim it does. The argument is adjacency: the requisition, the worker record, the org chart, the comp band and the position budget already live in one place, so a screening agent reads job architecture without an integration and internal mobility is a query rather than a project.

That advantage is real and it is also the trap — it only pays if you already run Workday HCM. Bought standalone, Workday Recruiting loses to most of the ATS category on recruiter speed and configuration turnaround.

Paradox is the exception worth knowing. Workday’s own announcement says the Candidate Experience Agent is “available for purchase now for both existing and new customers through Workday or Paradox,” so you can buy the frontline conversational layer without buying the suite. Workday’s published outcomes — over 30 million interviews scheduled with AI, time to hire as low as 3.5 days for frontline roles, 95% candidate satisfaction — are vendor-selected ceilings from finished deployments, not a forecast for yours.

Flex Credits and metered APIs

Workday moved agent consumption onto Flex Credits: a credit pool included in every subscription, refreshed annually, spendable across any agent rather than locked to a per-agent SKU, with more available for purchase. It is live today. The quieter change matters more to an ops team — APIs that external platforms call are now metered per call. If your stack pulls Workday data on a schedule, whether that is a Radancy career site, a CRM sync or a nightly BI job, that traffic carries a price it did not carry before.

On MCP, be precise. Workday has no shipped first-party MCP server. Agent-Ready Tools, announced at DevCon on June 2, 2026, expose HR and finance data to agents over MCP with Workday’s security model, delegation controls and audit trail inherited; they are in early access with general availability projected for H2 2026, alongside Developer Agent and Agent Passport. Until that lands, MCP access to Workday runs through third-party connectors. The mcp_available: false here means shipped, first-party and GA — recheck it after H2 2026.

Pricing reality

Workday publishes no price, and the third-party benchmarks disagree enough that any single figure is a negotiating anchor rather than a fact. Base HCM estimates run $20–32 PEPM at list against 30–40% enterprise discounting in one set of benchmarks and $34–42 PEPM at scale in another. The full suite with financials and recruiting is quoted at $80–150 PEPM. Annual software lands near $150–300K under 500 employees, $300–500K at 500–2,500, $500K–1.5M at 2,500–10,000, and past $1.5M above that.

What breaks budgets is not the license. Implementation runs $300–800K in year one — 100–200% of the annual subscription — on an average 8.2-month deployment. The initial term is three years with auto-renewal and a 60–90 day opt-out, escalators of 3–5% a year, and documented cases reaching 9%. Headcount bills on a full-time-equivalent model where part-timers count 25% and contingent workers 15–65%. Negotiate that contingent percentage line by line if you run a large contractor population; it is the clause nobody reads and the one that moves the number most.

Best for

Enterprise TA leaders at 5,000+ employee companies already running Workday HCM, hiring at volume across many job families, where the question is what to add rather than what to replace. Frontline-heavy employers — retail, health systems, logistics — extract the most, because the Paradox agent’s economics come from conversational apply and self-scheduling across thousands of requisitions.

It is the wrong pick if Workday is not already your HCM. Buying the suite to get its ATS is a $300K implementation aimed at a $60K problem.

Versus the alternatives

SAP SuccessFactors and Oracle HCM are the other two enterprise poles. The logic for picking either is identical to the Workday logic and it runs the same direction: the answer follows your existing system of record, not your ATS shortlist. Do not run a three-way TA bake-off if the HCM decision is already made.

Greenhouse wins on interviewer discipline and structured hiring. Pick it when hiring-manager behavior is the actual problem, and integrate to Workday for the worker record. iCIMS wins when you need enterprise TA breadth — career sites, CRM, high volume — without committing the HCM.

Ashby is the fast-growing entrant and the sharpest counter-pick up to roughly 2,000 employees: recruiter speed and native analytics that Workday Recruiting makes you build in a reporting tool.

Paradox standalone is the right buy when frontline conversational apply is the only thing you actually want. You keep your ATS and skip the suite.

If none of these fit, the ATS is not your constraint. A 14-step application and unstaffed hiring managers produce the same funnel on every platform here, and each one reports it faithfully as a low apply-completion rate.

Watch-outs

• Workday is the named defendant in the AI hiring case everyone else is watching. Mobley v. Workday is an ADEA collective action in the Northern District of California over algorithmic screening; the court authorized notice to potential members with a March 7, 2026 opt-in deadline, denied dismissal on March 6, 2026, and allowed core claims to proceed on June 22, 2026. Discovery is open and no trial date is set. Your exposure as the employer is separate from Workday’s. Guard: keep the current bias audit artifact for the Recruiting Agent on file, hold a human decision step on every rejection the agent scores, and put a contractual duty on Workday to supply audit evidence in the format NYC Local Law 144 requires rather than a compliance attestation.
• Per-call API metering repriced integrations you already built. Anything syncing on a schedule now has a variable line, and nobody costed it at signature. Guard: pull 30 days of API call volume by consuming system before your next true-up, then move the high-frequency polling jobs to event-driven or batched pulls. Get the metering rate and any included allowance written into the order form, not the rate card.
• The FSE model and the escalator compound quietly. A 3–5% escalator on a three-year term with 60–90 day auto-renewal means the renewal conversation starts before you have year-two adoption data. Guard: diary the opt-out date at signature, cap the escalator at CPI with a hard ceiling, and fix the contingent-worker percentage for the full term instead of letting it reprice.
• Half the agent story is early access, not shipped. Agent-Ready Tools and Developer Agent are projected GA in H2 2026 and Agent Passport before the end of 2026 — projections, from a vendor with a three-year contract in front of you. Guard: buy on what is GA today and tie any agent-dependent scope to a written availability date with a credit if it slips. If the business case needs H2 2026 functionality, the business case needs a signed date.