The revenue line is the headline, the product list is the work
Workday's fiscal 2027 second-quarter release, published on August 27, 2026, leads with the numbers you'd expect. The release reports total revenue of $2.649 billion, up 12.8 percent on the year, and subscription revenue of $2.471 billion, up 13.9 percent. Aneel Bhusri's quote says AI drove more than 25 percent of new annual contract value and that more than 5,500 customers now use at least one of Workday's organic agents. Our sibling piece on what the adoption statistic can and cannot tell you takes the 5,500 apart properly.
We read the release for something else. Further down, past the guidance, there is a list of product items that will land on a platform owner's desk within a quarter or two, whether or not anyone in the tenant asked for them. That list is the part of the Workday newsroom post we'd take into the next platform review, because each item changes who can do what inside a Workday tenant.
Two items are generally available, so someone will ask for them
The release says the Financial Audit Agent is now generally available, and so is Workday Learning, powered by Sana. Generally available is a different category from announced. It means the switch exists in a production tenant, a partner can put it in a statement of work, and a finance director who read the same press release can ask on Monday why it isn't on yet.
For the Financial Audit Agent, the question we'd settle before the switch is which controls it reads and where its findings are written. An audit agent that raises a flag into a queue nobody owns is worse than no agent, because the flag creates a record of something seen and not acted on. The release doesn't describe that workflow, so pin it down with your account team before a controller sees a demo. Our piece on why deterministic rules still belong in an agent strategy covers the same ground from the rules side.
Workday Learning with Sana is a different kind of change. It arrives on the HCM side of the house, where the owner is usually an HR systems lead rather than a platform architect. The release doesn't describe what it does with employee records, so the governance question is the one every HCM team already has for talent data. Which fields feed it, and who signed off on that.
Agent Passport is the item to read twice
The release says Agent Passport tests and verifies agents before they reach production and then monitors them continuously. That is a governance surface, and it speaks to the question every tenant has asked about organic agents. What evidence exists that this agent behaves the way the vendor says it does.
What the release doesn't say is who holds that evidence, whether a customer can see the test results, or whether Passport covers agents a customer builds as well as the ones Workday ships. Those three answers decide whether Passport replaces any part of your own review or simply sits beside it. Until they arrive, we'd keep running the questions in our agent readiness check.
Developer Agent changes who can build in your tenant
The release also announces a Developer Agent, which Workday says lets developers build AI applications using natural language. Read that sentence with an admin's eyes rather than a developer's. Every platform that has shipped natural language building has had to answer where those apps go, who reviews them, and whether they follow the same release path as everything else.
The post gives no availability date, so there is time to decide before the first one appears. Anything a Developer Agent produces enters the same sandbox to production path as a hand-built integration, with the same reviewer. If your tenant doesn't have that path written down yet, this announcement is the reason to write it, and our Release Radar note on separating announcements from changes that affect owners is a workable template.
The data boundary moved in two directions
Two partnership items in the release change where Workday data can live and where Workday agents can run. Workday Data Cloud is integrating with AWS for what the release calls bi-directional, zero-copy access. And Workday agents are integrating into Google's Gemini Enterprise.
Zero-copy reassures a data team and should get a security team's attention, in the specific sense that nothing is duplicated but everything is reachable. The question for your next review is which Workday objects become queryable from an AWS account, under whose identity, and whether the row-level security you configured in Workday carries across. The release doesn't answer that, and we'd expect the answer in integration documentation rather than a press release.
The Gemini Enterprise integration runs the other way. A Workday agent inside a Google surface means an employee can ask a Workday question from a tool Workday doesn't administer. Whoever owns identity on the Google side joins your Workday access model, and they may never have been told. Have that conversation with the Google Workspace owner before the integration is enabled. After the first ticket about a pay slip appearing in the wrong chat is too late.
Book the review before the switch appears
Adaptive Decision Intelligence rounds out the list, described in the release as a new capability for finance and operations teams. The post says little more than that, and it belongs in the same Financial Management discussion as the audit agent, because both put a model's output in front of a controller.
The numbers will be stale by the November call. The product list will still be shaping your tenant well into 2027. Book the platform review for the week the Financial Audit Agent toggle shows up in your tenant, and send the three Agent Passport questions to your account team before that meeting.
