What Happened to Workday Illuminate Agents? Meet Sana

Workday has been shipping AI agents faster than most HR and Finance teams can track them, and in early 2026 it changed what they're called. What were once known as Workday Illuminate agents, as of March 2026 they live under a new umbrella called Sana. Same core idea, new name.

In this guide, we'll cover what Workday Illuminate was, what replaced it, and what each current agent actually does for HR and Finance teams, grouped by function. We’ll also provide an answer to the question "what does this automate?". And we'll finish on what must be true inside your organization before any of these agents deliver value.

Workday Illuminate vs. Sana

The question "what is Workday Illuminate?" now has a past-tense answer. Workday introduced Illuminate in September 2024 as the umbrella brand for its AI meant to add automation and intelligence across HR, Finance, and industry workflows. For about eighteen months, "Illuminate" was the name on everything from the platform's AI features to its growing roster of agents.

That changed after Workday acquired the AI company Sana in 2025 for $1.1 billion. On March 17, 2026, Workday announced that Sana from Workday was available to customers worldwide, rebranding its AI around Sana. The Illuminate name has largely retired. The agents it introduced didn't disappear; they were renamed, reorganized, and in several cases folded together under Sana. That's Workday Illuminate explained in a sentence: the 2024 AI brand that became Sana from Workday in 2026.

So what is Sana? Think of it as three layers working together.

1. Sana itself is top layer: it reads what a user asks in plain language, then coordinates the multi-step work across Workday's individual agents.
2. The agents are the specialists that do the actual tasks.
3. And underneath, the Workday Agent System of Record (ASOR) is the governance layer that tracks every agent, enforces permissions, and keeps an audit trail.

Sana runs on your trusted Workday data and inherits your existing security and configuration, which is the whole point. This is embedded AI in the truest sense: the agents already know your job architecture, approval chains, and history because they operate inside Workday rather than beside it.

That last part matters most. the architecture underneath, trusted data plus governance through ASOR, is what determines whether they actually work.

Workday AI Agents: Agentic HR

Workday now groups its HR AI agents into an "agentic team" coordinated by Sana. Here's the current list, with old names noted where an agent you may have formerly known under Illuminate has changed.

• ‍HR Service Agent (formerly the Case Agent) answers routine HR questions from your policies and Workday data, deflects tickets, drafts responses, and routes the cases that need a human to the right specialist.
• ‍Talent Acquisition Agent sources and grades candidates from internal and external pools, schedules interviews, preps interviewers, and rediscovers past and contingent talent, which absorbs the old Contingent Sourcing capability. The hiring decision stays with recruiters.
• ‍Talent Management Agent surfaces engagement signals early to flag retention risk, runs the performance cycle, and drafts reviews and goals from connected data, with a human reviewing every output.
• ‍Workforce Management Agent tracks demand to recommend schedule changes before coverage gaps form, handles bulk time actions and shift swaps, and flags cost and compliance impact. This covers much of the frontline scheduling the earlier Frontline Agent handled.
• ‍Total Rewards Agent monitors pay equity, models compensation scenarios, and guides employees through benefits and life-event decisions grounded in their real options, while flagging market and minimum-wage shifts.
• ‍Payroll Agent delivers accurate, timely, compliant payroll from inputs to pay, catching issues before payday and surfacing the insights that move a team from reactive processing to proactive oversight.
• ‍Organizational Development Agent models complex org changes in plain language inside a secure sandbox and cleanses job data, building job profiles, families, and grades and filling in missing skills and descriptions.

Taken together, these agents are less about replacing HR judgment than clearing the administrative noise that keeps HR teams from using it where it counts.

Workday AI Agents: Agentic Finance

The finance AI agents follow the same pattern: automate the document-heavy, repetitive work that slows down contracting, procurement, and the closing process.fo

• ‍Accounting Agent (the evolution of the Document-Driven Accounting Agent) reconciles, tests controls, and catches variances continuously, automating audit evidence and closing the books in days rather than weeks, leaving only the exceptions to review.
• ‍Procurement Agent reviews every transaction against contract terms and SLAs to block overcharges, unlocks trapped contract data, and steers spend toward the best-fit, policy-approved suppliers. It took on much of what the Supplier Contracts Agent used to do.

Two more agents sit adjacent to Finance and are worth knowing.

1. The Legal Agent consolidates the former Contract Intelligence and Contract Negotiation agents into one: it drafts and redlines contracts against your playbook, answers plain-language questions with links to the source, and tracks obligations, risks, and renewals across every agreement.
2. The Travel and Expense Agent books in-policy trips, drafts expense reports from receipts, and settles reimbursement through Workday Payroll, flagging duplicates and policy violations before payment.

Sitting across all of it is the Self-Service Agent, the one Workday leads with. It's the conversational entry point for employees and managers: ask a question in natural language, get an answer grounded in your policies and Workday data, and let the agent carry the task from question to completion.

Notice the pattern. None of these agents make the contract decision, hire the candidate, or approve the close. They compress the time between "the information exists somewhere" and "a person can act on it."

What's Available Now

Sana became available worldwide in March 2026, and Workday keeps adding new agents to the lineup. Expect that roster to keep growing.

Names, groupings, and availability shift often, so confirm the status of any specific agent with your Workday account team or AMS partner before you plan a rollout around it. And whatever the roster looks like when you’re reading this, the practical move is the same: don't build your AI roadmap around a single agent's launch date.

The Hard Part Isn't the Agent

Widely cited MIT research from 2025 found that only about 5% of enterprise AI initiatives were delivering measurable ROI. That number wasn't about the quality of the AI. It was about organizational readiness: whether the data was clean enough, whether the process was documented well enough to automate, and whether anyone had defined what success looked like before flipping the switch.

Workday isn't shy about this either. Its own product and technology president, Gerrit Kazmaier, has argued that AI value stalls when tools sit outside the workflows that actually run the business, disconnected from the core systems that hold the data and context they need to be accurate. His metaphor is worth borrowing: raw AI without process and context around it is like a train without rails.

For HR and Finance, where a payroll run or a book close has to be right every time, an agent that's "almost right" is wrong. The rails matter as much as the engine. That's the case for embedded AI over bolted-on tools: intelligence that sits inside the workflow, with the data and context around it, rather than reaching in from the outside.

That's the real story underneath every Workday AI agent. The capability is arriving fast. Whether it delivers value depends almost entirely on how ready your Workday environment, your data, and your processes are to receive it. An agent that automates a broken process just breaks it faster.

This is where configuration and governance work matters more than most teams expect going in: mapping which processes are stable enough to automate, cleaning up the data an agent will draw from, and setting up the right guardrails before an agent starts acting on employee or supplier records rather than just surfacing information.

Four Questions Worth Asking First

Before activating any agent, run your own process through a short readiness check rather than assuming the agent will adapt to whatever exists today:

1. Is the process stable? An agent set to automate a process that changes every quarter will need constant reconfiguration, which erases most of the time savings.
2. Is the underlying data clean? The Accounting Agent and the Legal Agent are only as accurate as the records and documents they read from. Garbage in still means garbage out, just faster.
3. Who owns the exceptions? Every automated process has edge cases. Decide upfront who reviews what the agent flags, rather than discovering the gap after something slips through.
4. What does success actually look like? A faster close or a shorter time-to-fill only counts as ROI if you defined the baseline and the target before the agent went live, not after.

None of these questions are complicated. They're also the exact questions the MIT research suggests most organizations skip, which is a large part of why so few AI initiatives show measurable return.

Where to Start

If you're an HR or Finance leader looking at this list and wondering where to begin, start with the process that already frustrates your team the most. If contract execution is a bottleneck, look at the Legal Agent. If scheduling eats hours every week, look at the Workforce Management Agent. Match the agent to the problem you actually need solved.

The Planet Group is Workday's first global staffing and AMS partner across North America and Europe, with more than 2,000 Workday placements, roughly 300 customers, a 61% repeat-client rate, and a 400-plus person Workday Certified Network behind it.

If you're weighing which Sana agents fit your environment, or you need the configuration and governance work done right before you turn one on, that's the conversation worth having next. Contact us today and let's get started.