Workday’s Sana is a conversational interface and agent layer for starting work in Workday. Employees can ask questions, find information, and trigger HR or finance workflows in natural language instead of navigating menus. Its agents can assist with actions, but they operate inside Workday security roles, policies, delegations, approval chains, and administrator settings—not as unrestricted autopilot.
For administrators, Sana is a rollout and governance project as much as a user-interface change. Security configuration, agent registration, provisioning, connector settings, approvals, monitoring, and tenant-specific commercial terms determine what employees can actually use.
As an Amazon Associate I earn from qualifying purchases.
What Sana is and why Workday introduced it
Workday completed its acquisition of Sana on November 4, 2025. Workday said the deal combined Sana’s enterprise knowledge capabilities with Workday data and context. Gerrit Kazmaier, Workday’s president of product and technology, described the goal as creating “a single, intelligent interface that connects the systems, data, and actions employees rely on — becoming the new front door for work.” That is Workday’s product vision, not an independently measured productivity result.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
On March 17, 2026, Workday announced Sana for Workday, the Sana Self-Service Agent, and Sana Enterprise. Workday characterizes Sana for Workday as a unified conversational AI experience that can replace traditional menus and navigation for many interactions. In practice, Sana is the place where a worker asks for an answer or action, while agents provide the task-specific reasoning and workflow execution behind that request.
#1 Best Overall
Workday said the Self-Service Agent launched with more than 300 skills covering areas including pay, time, and absence. “300+ skills” is Workday’s vendor-reported launch figure; it is not an independent inventory, accuracy test, or promise that every tenant has every skill enabled.
What employees can do in the Sana experience
Ask questions in ordinary language
An employee can ask for help locating information instead of opening several Workday pages. Workday’s product description says Sana can answer questions, summarize information, and search Workday and other approved knowledge sources. A response may explain a policy, identify a record, or point the employee toward the next step.
Start or complete routine work
Workday says the Self-Service Agent can assist with or handle everyday HR and finance actions. Its examples include pay, time, and absence questions and tasks. The exact actions depend on the skills enabled in an organization, the worker’s role, available data, and the tenant’s configuration.
Move between answers and actions
The important change is that Sana is not limited to a question-and-answer window. A conversation can lead to a workflow, approval request, or action in a connected system. Sana’s product materials describe an experience in which employees can see what needs attention, obtain an explanation, and move work forward from the same interface.
Expect approval gates for sensitive changes
For a sensitive action, Sana may show a proposed change and ask for approval before anything is committed. Workday says security roles, delegations, policies, and approval chains continue to apply. An employee should therefore treat a conversational response as a guided route into Workday’s controls, not as permission to bypass them.
Why “autopilot” does not mean unrestricted autonomy
An AI agent can perform more steps than a conventional chatbot: it can interpret an objective, retrieve information, call an allowed skill, and submit a workflow. Workday’s own descriptions still place those operations inside configured permissions and policy checks. The agent cannot legitimately grant itself access to a payroll record, skip a required approver, or change a domain policy.
Workday documents two execution patterns in its Agent System of Record:
- Ambient execution: the agent acts with permissions assigned to the agent itself.
- Delegate execution: the agent acts on behalf of a worker, using that worker’s permissions.
The distinction matters when an administrator investigates an action. The same natural-language request can have a different result depending on whether the agent or the requesting worker owns the relevant access. Administrators must review those permission models rather than assuming every agent behaves like the person who typed the request.
Workday also describes audit trails and reports for agent activity. Those records support oversight, but they do not replace periodic review of security roles, policies, connector access, and approval behavior.
Sana Core and Sana Enterprise: what changes
Workday separates the experience into Core and Enterprise scopes. The practical difference is the systems an organization wants Sana to reach, not a simple “basic versus better” ranking.
Rank #3
| Comparison point | Sana Core | Sana Enterprise |
|---|---|---|
| System scope | Conversational and agentic use across Workday HCM and Financials. | Core Workday use plus third-party connectors and cross-application workflows. |
| Actions | Requests and tasks within configured Workday capabilities. | Requests and actions that can be orchestrated across Workday and connected systems. |
| Administration | Security permissions, agent access, provisioning, and workspace settings are configured in the Workday tenant. | All Core administration plus Enterprise domain setup, connector configuration, and cross-system access controls. |
| Governance | Agents are registered, authorized, activated, monitored, and reviewed through the same lifecycle controls. | Those controls extend to connected systems; Workday says cross-system actions are logged and available for audit. |
| Commercial prerequisites | Applicable agreement, product, and Flex Credits conditions vary by customer. | Additional Enterprise purchase and Flex Credits policy requirements may apply, depending on the customer agreement. |
Workday says role and permission mirroring is used for connected work. That means an Enterprise rollout should map not only which connector is enabled, but also which users and agents may perform each class of action in that external system.
What administrators must configure before employees use Sana
Workday’s setup documentation presents Sana as a tenant-level implementation. A typical sequence is:
- Configure AI administrator security permissions. Give the administrators responsible for Sana the domains and privileges needed to configure its workspace, agents, and policies. Separate administrative access from ordinary employee access.
- Set up the Agent System of Record. Register the agents your organization will use, define their configuration, select the execution mode, and establish who can activate, deactivate, and manage them.
- Edit relevant domain security policies. Review the Workday domains behind each skill. A conversational interface does not change the underlying authorization model.
- Provision employees. Assign the appropriate user populations and security groups. Provisioning determines who can see Sana and which actions they can request.
- Enable the Sana card or access point in Workday. Make the interface available in the employee experience only after security and provisioning are ready.
- Complete Sana Enterprise domain setup when applicable. Enterprise customers must configure the additional domains, connectors, and cross-system settings required for their chosen orchestrations.
- Set workspace restrictions. Workday documentation says administrators may disable web search, support chat, or individual connectors. They can also restrict agentic actions to approved email domains.
- Test approval and audit paths. Use representative worker roles to verify that sensitive actions stop for approval, denied actions remain denied, and activity appears in the expected monitoring and audit reports.
Skipping the permission and provisioning work can produce a polished interface that answers too little—or exposes an action to the wrong population. A controlled pilot with several worker roles is safer than enabling every skill and connector at once.
How the Agent System of Record supports governance
The Agent System of Record is Workday’s central framework for the agent lifecycle. Its documented functions include defining and registering agents, configuring them, activating and deactivating them, managing access, monitoring activity, and producing audit reports.
Registration and ownership
Every production agent should have a named owner, a documented purpose, an execution mode, and a defined set of skills or connectors. Registration gives administrators a place to identify what the agent is supposed to do before it is activated.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Rank #4
Activation and deactivation
Administrators can control when an agent is active. Deactivation is an important incident-response step if a connector, policy, or agent behavior needs investigation. It is different from deleting the configuration, because the organization may need the historical record for review.
Monitoring and audit
Workday recommends regular review of agent activity. Monitoring should look for unexpected volumes, repeated denied requests, unusual data access, failed approvals, and actions made through a connector. Audit reports help establish what happened, but administrators still have to interpret the events and correct the underlying configuration.
Permission choice
Choosing ambient or delegate execution affects both risk and accountability. An agent with its own narrowly scoped permissions may be easier to constrain for a fixed process. Delegated execution can reflect a worker’s access, but it requires careful testing of what different roles can request through the same conversation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Availability, Flex Credits, and tenant eligibility
Workday’s March 17, 2026 announcement said Sana for Workday and the Self-Service Agent were available to customers worldwide through Workday Flex Credits without an extra license or separate paywall. That statement describes Workday’s announced commercial model; it does not establish that every customer can turn the feature on immediately.
Free tools Windows power users keep installed
One-click scans. No signup required.
Workday’s administrator security guidance lists agreement and product prerequisites, including UMSA and applicable Flex Credits terms. It also notes that conditions can depend on the organization’s service agreement. Sana Enterprise has additional product and Flex Credits policy requirements. Before planning a rollout, an administrator should confirm:
Best Value
- which Sana edition is covered by the organization’s agreement;
- whether the tenant has the required UMSA and product entitlements;
- how Flex Credits are consumed for the selected skills, agents, or connectors;
- which third-party systems are supported for the tenant’s Enterprise configuration; and
- whether regional, security, or data-governance conditions affect enablement.
Those checks are contractual and tenant-specific. The launch announcement should not be read as a universal guarantee of eligibility or unlimited usage.
What Sana is not yet proven to do
The official Workday materials establish intended scope, controls, and setup responsibilities. They do not provide an independent study of Sana’s accuracy, productivity gains, response times, or error rates. Workday’s examples and customer outcome statements should therefore be treated as vendor claims rather than verified performance benchmarks.
Employees should check important results against the underlying Workday record, especially for pay, leave balances, compliance-sensitive requests, and finance transactions. Administrators should validate each enabled skill with realistic data and approval paths instead of assuming that a successful demonstration represents every worker or exception case.
Practical questions for an implementation decision
When Core may be enough
Core is the narrower choice when the objective is conversational access to Workday HCM and Financials without cross-application orchestration. The implementation still requires security, provisioning, agent lifecycle management, and audit review, but it avoids adding third-party connector scope.
When Enterprise is justified
Enterprise becomes relevant when a process genuinely crosses Workday and another approved system—for example, a workflow that must retrieve context from one application and submit a controlled action in another. The benefit depends on connector availability, permission mapping, logging, and the organization’s ability to operate those integrations.
Questions to answer before launch
- Which employee populations need conversational access, and which should be excluded?
- Which actions are informational, and which require approval or delegated authority?
- Will each agent use its own permissions or a worker’s permissions?
- Who owns the agent after activation, and who can deactivate it?
- Which connectors, web search settings, support channels, and email domains are allowed?
- What audit report will be reviewed, by whom, and how often?
- What Flex Credits and agreement terms apply to this tenant?
Bottom line for employees and administrators
For employees, Sana changes the starting point for Workday work: ask, understand, and act in one conversational flow rather than hunt through menus. For administrators, the interface is only the visible layer. The real implementation is the combination of security roles, policy checks, provisioning, approvals, connectors, agent permissions, lifecycle controls, and audit review. Workday’s “autopilot” language describes a more capable agent experience, not permission to operate outside those controls.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
