---
title: "Workday Sana Self-Service Agent Admin Setup Checklist"
description: "Step-by-step Workday admin checklist for Sana Self-Service Agent: UMSA opt-in, AI Admin Security Permissions, Graph API Descriptor, skills, and tool checks."
url: https://kiweely.com/guides/workday-sana-self-service-agent-admin-setup
---

# What Workday Admins Need to Configure Before Turning On Sana Self-Service Agent

Before activating Sana Self-Service Agent, run Set Up AI Admin Security Permissions, enable Agent System of Record, and add All Users to Graph API Descriptor. Keep the Guide skill enabled and align every skill’s Available To groups with underlying tool security policies.

-   **Checked** against Workday on November 13
-   Applies to **Workday AI**
-   7 min read
-   Published November 13
-   By Kiweely

Before turning on **Sana Self-Service Agent**, administrators must confirm **Universal Main Subscription Agreement (UMSA)** eligibility, complete the one-time **Set Up AI Admin Security Permissions** task, enable the **Agent System of Record (ASOR)** functional area, and assign **All Users** to the **Graph API Descriptor** domain. Inside **Agent Management Hub**, you must keep the **Guide** skill enabled and align every skill’s **Available To** security groups with the domain and business process security policies required by its underlying API tools.

Skipping any of those steps produces confusing rollout symptoms: the legacy Workday Assistant bubble disappears, the new **Workday Chat** and **Ask Sana** controls appear in the global header, and employees immediately hit permission errors or silent non-actions.

## 1\. Contract, UMSA, and Pre-Sign Timing Checks

Check contractual and tenant prerequisites before touching security policies:

-   **Confirm UMSA status and Machine Learning data contributions:** Workday-built agents require **UMSA** for both non-Production and Production tenants, with AI data contributions kept enabled for full functionality. Check your agreement value under your organization profile on Workday Community. Signing UMSA grants complimentary [Workday Flex Credits](https://kiweely.com/guides/workday-flex-credits-explained) for **Sana Core** (requires Core HCM or Core Financials); continuing past the complimentary allotment or enabling **Sana Enterprise** (third-party connectors, custom agent builder, Meeting Agent) requires signing the Workday Flex Credits and Platform Entitlement Policy.
-   **Avoid the 14-day Policy Intelligence ingestion delay:** If you want Self-Service Agent to answer company policy questions right away, opt in to the **People Experience / Help** service on the **Innovation Services And Data Selection Opt-In** report *before* signing the UMSA or changing your data hosting region. Otherwise, policy document ingestion can face a 14-day waiting period. (Note that you no longer need the Workday Help SKU to enable the **Policy Intelligence** skill.)

## 2\. Run Set Up AI Admin Security Permissions and Enable Graph API Descriptor

Workday delivers a bundled bootstrap task—**Set Up AI Admin Security Permissions**—that initiates back-end AI processing and assigns unconstrained security groups across four administrative pillars: **Agent System of Record**, **User Provisioning Workspace for Sana**, **Platform Consumption Console**, and **Sana** workspace administration.

One-Shot Task and 30-Day Notification Expiry

The **Action Required: Review and Configure the Security for Sana and Workday Agents** notification expires after 30 days, and **Set Up AI Admin Security Permissions** can only be submitted once per tenant (if two admins open it simultaneously, Workday processes only the first submission). If the notification expired before completion, run **Notify Users to Configure Sana Agent Security**; if the task was already submitted, make later edits through standard domain security tasks.

After completing the bootstrap task, verify three foundational security items that cannot be migrated automatically across tenants:

1.  **Verify ASOR domains in every tenant:** If you configure security manually rather than through the bootstrap wizard, run **Maintain Functional Areas** to enable *Agent System of Record*, activate security, and enable six unconstrained domains: *Agent Compliance*, *Agent Management Hub*, *Manage: Agents*, *Reports: Agent Reporting*, *Setup: Agents*, plus *Reports: AI Agent Security* in the *System* functional area.
2.  **Enable Graph API Descriptor for All Users:** Enable the **Graph API Descriptor** domain security policy and assign the **All Users** security group, then run **Activate Pending Security Policy Changes**. Without this policy, Self-Service Agent tool calls that traverse Workday Graph fail at runtime.
3.  **Configure Sana Enterprise domain (if licensed):** Edit the **Sana Enterprise** domain in the *System* functional area to control which security groups can invoke cross-system connectors and Enterprise features.

## 3\. Register Self-Service Agent and Map ESS vs. MSS Skills

Open **Agent Management Hub**, switch to the **Unregistered Workday Agent** tab, click **Register** next to **Self-Service Agent**, and confirm. Then open **Self-Service Agent** from the **Agent Registry** tab and click **Configure Agent**.

Workday names the 20 delivered Self-Service Agent skills by audience (`ESS` for employees, `MSS` for managers) and risk tier (`Base` for read-only retrieval, `Advanced` for transactional writes):

| Skill Category | Delivered Skills | Operational & Security Rules |
| --- | --- | --- |
|  **Core Navigation & Search** |  **Guide** *(mandatory)*, **Global Search Answers**, **Policy Intelligence** |  **Guide** must stay active for Self-Service Agent to run at all. Turning off **Global Search Answers** reverts view-only global search queries to Assisted Search. |
|  **Employee Retrieval (`ESS - Base`)** |  **HCM - Base**, **Payroll - Base**, **Financial Management - Base**, **Talent Management - Base**, **Workforce & Frontline Management - Base** | Read-only worker profile, pay/tax forms, expense history, goals, internal jobs, schedules, and time-off balances. Typically scoped to *Employee As Self*. |
|  **Employee Actions (`ESS - Advanced`)** |  **HCM - Advanced**, **Payroll - Advanced**, **Help - Advanced Skill**, **Talent Management - Advanced**, **Workforce & Frontline Management - Advanced** | Executes real writes: personal data edits, bank accounts, payment elections, tax withholdings, expense submissions, time-off requests, shift check-in/out, and support cases. Align MFA and tenant access restrictions on any group added to **Available To**. |
|  **Manager Retrieval (`MSS - Base`)** |  **MSS - HCM - Base**, **MSS - Talent Management - Base**, **MSS - Workforce & Frontline Management - Base** | Team structure, direct-report time blocks, goal progress, and learning status. Scope to *Manager* or *Management Chain* groups. |
|  **Manager Actions (`MSS - Advanced`)** |  **MSS - HCM - Advanced**, **MSS - Payroll - Advanced**, **MSS - Talent Management - Advanced**, **MSS - Workforce & Frontline Management - Advanced** | Initiates job changes, approves/denies time entries, manages team goals/reviews, and calls **Request Payroll Inputs** (enforces Workday Payroll pay-group checks and pay component segments). |

Do not blindly place *All Users* on `MSS` or `ESS - Advanced` skills. Run **View Security for Agent Skill** for **Self-Service Agent** and click **Test My Setup** to resolve every `Config Required` or `Review Required` status before clicking **Activate**:

Fails

```
Skill: MSS - Payroll - Advanced
Available To (Agent Interaction Policy): Management Chain
Underlying Tool (request_payroll_input): Management Chain missing from Request Payroll Inputs BP / domain policy
Result: Manager sees the skill in chat, asks to submit a bonus input, and hits an API authorization failure.
```

Works

```
Skill: MSS - Payroll - Advanced
Available To (Agent Interaction Policy): Role-Based Manager Group
View Security for Agent Skill -> Functional Status: Full Match (pay component segments & BP policies aligned)
Result: Agent validates the worker's Workday Payroll pay group and initiates Request Payroll Inputs cleanly.
```

## 4\. Tune Individual Tools, Provision Sana Users, and Know the Hard Boundaries

In Workday 2026R2, you do not have to disable an entire `Advanced` skill just because one action (such as job changes or bank account edits) breaks internal policy:

1.  **Toggle individual tools and custom guidance:** On the Self-Service Agent **Skills** tab, click **Configure** in the **Tools** column. Toggle off specific tools (mandatory tools have a locked toggle) and populate **Tool Disabled Message**, **Supplemental Message**, or **Fallback Message**—for example, steering workers to a custom task or portal when `view_bonus_history` is disabled or `initiate_change_job` fails. Note that the LLM paraphrases your text rather than echoing it verbatim, and changes take 5 to 10 minutes to clear cache for new conversations.
2.  **Provision users to Sana Workspace and Home Page:** Optionally restrict eligible groups via **Set Up Security Groups for User Provisioning**, then open the **User Provisioning Workspace**, click **Configure** on the **Sana** tile, select security groups, and click **Preview and Enable Sync**. Finally, enable the **Sana** row under Web Widgets (and optionally Mobile App Widgets) in **Home Page Configuration**. Activating Self-Service Agent with at least one skill automatically replaces the Workday Assistant bubble with **Workday Chat**; if you want to hide the **Ask Sana** header button, opt out in **Maintain Feature Opt-Ins**.
3.  **Document unsupported populations before launch:** Self-Service Agent does **not** support contingent workers, additional jobs (multi-position), **Configure Optional Fields** hide/require overrides, visibility into *My Tasks*, proxy or implementer accounts, or delegated actions (acting on behalf of someone else will not create real events). Furthermore, if you test policy extraction from articles, test in an Implementation tenant rather than Sandbox, where article data extraction is unsupported.

## How to Verify Your Setup Before Opening Access to All Employees

Pilot with a small user-based security group—never with **Start Proxy**, because Self-Service Agent blocks proxy accounts:

1.  Sign in as a real test employee and test manager in an Implementation tenant, open **Workday Chat** and the **Sana Workspace** (and Slack, Teams, Copilot, or Gemini Enterprise if using **Workday Everywhere**), and test one `Base` read, one `Advanced` write, and one disabled-tool prompt.
2.  Open **Agent Management Hub > Self-Service Agent > Monitoring** to confirm session IDs are recording, and check **View AI Agent User Audit Trail** to verify transactions log the agent ASU as *By User* and the worker as *On Behalf Of User*.

Because Self-Service Agent exposes 20 skills backed by dozens of REST and Graph tools, [Kiweely Vine](https://kiweely.com/vine) can inspect your live [security group permissions](https://kiweely.com/guides/workday-security-audit-before-enabling-ai-agents) against every tool’s domain policy—catching mismatches and testing real non-proxy worker personas in a visible, stoppable browser before rollout.

## Frequently asked questions

### Why can employees open Sana Chat in Workday, but every Self-Service Agent action fails?

Workday evaluates two independent security gates at runtime: the skill's Available To interaction policy in Agent Management Hub and the invoking user's standard domain and business process security policies for the underlying API tools. Also verify that the Graph API Descriptor domain is enabled with All Users assigned.

### Can I disable one risky action inside a Self-Service Agent skill without turning off the whole skill?

Yes. Open Self-Service Agent in Agent Management Hub, go to the Skills tab, and click Configure in the Tools column to toggle individual tools off and set a Tool Disabled Message, Supplemental Message, or Fallback Message. Allow 5 to 10 minutes for tool and skill cache updates to apply to new conversations.

### Do I still need the Workday Help SKU to let Self-Service Agent answer policy questions?

No. Workday removed the Workday Help SKU requirement for the Policy Intelligence skill, and you can also connect external policy sources such as Microsoft SharePoint or OneDrive on the Documents tab. However, opt in to People Experience Help on Innovation Services And Data Selection Opt-In before signing the UMSA to avoid a 14-day policy ingestion wait.

### Can contingent workers, proxies, or delegates use Sana Self-Service Agent?

Self-Service Agent does not currently support contingent workers, additional jobs (multi-position), implementer or proxy accounts, Configure Optional Fields overrides, or delegated actions. Even if a delegate chat appears to succeed, Workday does not create the underlying event.

## Sources checked

-   [Set Up Sana for Workday](https://doc.workday.com/admin-guide/en-us/workday-ai/sana-for-workday/steps--set-up-sana-for-workday.html)
-   [Set Up AI Administrator Security Permissions](https://doc.workday.com/admin-guide/en-us/workday-ai/sana-for-workday/set-up-ai-administrator-security-permissions.html)
-   [About Self-Service Agent](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/self-service-agent/concept--self-service-agent.html)
-   [Set Up Self-Service Agent](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/self-service-agent/set-up-self-service-agent.html)
-   [Self-Service Agent Skills and Configure Self-Service Agent Tools](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/self-service-agent/disable-individual-self-service-agent-tools.html)

## Related guides

[All guides](https://kiweely.com/guides)

-   [### What Is Workday Agent Passport and How Do Admins Audit or Stop an AI Agent? Workday Agent Passport shows verified security stamps, assessors, and grant dates on each agent profile in Agent Management Hub. To audit or stop an AI agent, inspect View AI Agent User Audit Trail and click Deactivate Agent or disable its Agent System User OAuth 2.0 client.](https://kiweely.com/guides/workday-agent-passport-governance-explained)
-   [### How to Audit Over-Permissive Workday Security Groups Before Turning On AI Agents Delegate AI agents inherit the union of an invoking user’s security groups, while ambient agents act through dedicated Ambient Agent Security Groups. Before activating any agent, audit unconstrained User-Based groups, stale role assignments, and View Security for Agent Skill pathways.](https://kiweely.com/guides/workday-security-audit-before-enabling-ai-agents)
-   [### Workday Domain Security vs. Business Process Security Policies: Which Controls What A domain security policy controls View/Modify (and Get/Put) access to reports, report fields, setup tasks and web services. A business process security policy controls who can start, approve, cancel, rescind or view a transaction. Both need activation.](https://kiweely.com/guides/workday-domain-vs-business-process-security-policies)
-   [### Workday Agent System of Record (ASOR): An Admin’s Security and Setup Checklist Workday’s Agent System of Record (ASOR) governs both Workday-built and custom agents through the Agent Management Hub. Here is how ASUs differ from ISUs, how delegate and ambient execution evaluate security, and the exact steps an administrator takes to register, secure, and audit an agent.](https://kiweely.com/guides/workday-agent-system-of-record-security)
-   [### Workday Flex Credits Explained: What Counts as an Action Flex Credits replace per-employee pricing for Workday AI with credits drawn down per metered action. What Workday publishes about the Rate Card, the Platform Consumption Console, non-production use and how to reduce consumption.](https://kiweely.com/guides/workday-flex-credits-explained)

### More guides on Custom agents and Workday AI

[All 91 guides](https://kiweely.com/guides#custom-agents)

-   [Calling External Glean and Sana Agents From Workday Extend Without Hitting 408 Request Timeouts](https://kiweely.com/guides/workday-extend-external-agents-glean-sana-408-timeout)
-   [AI Gateway Document Intelligence (/ml/docint/v1/ids): Converting Workday Drive Base64 to Binary, Resume Layout Pitfalls, and Fixing Opt-In Errors](https://kiweely.com/guides/workday-document-intelligence-base64-binary-opt-in-errors)
-   [Why a Workday Custom Agent Has No Credentials Tab in Agent Management Hub, and How to Expose It Through A2A](https://kiweely.com/guides/workday-custom-agent-missing-credentials-tab-a2a)
-   [Passing Large CSV File Attachments From a Workday Custom Agent Flow to Orchestrate Without Hitting Context Limits](https://kiweely.com/guides/workday-custom-agent-csv-file-attachment-orchestrate)
-   [Preventing Duplicate Initiate\_Change\_Job Business Process Events When a Workday Custom Agent Calls a Write Tool](https://kiweely.com/guides/workday-custom-agent-duplicate-initiate-change-job)
-   [Why Custom Agent Actions Fail With S22 Permission Denied or Task Not Authorized in Agent Chat](https://kiweely.com/guides/workday-custom-agent-actions-permission-denied-tool-wids)
-   [Fixing invalid\_request OAuth Token Exchange Errors When Connecting Microsoft Copilot Studio to Workday A2A](https://kiweely.com/guides/workday-a2a-copilot-studio-oauth-invalid-request)
-   [Why Custom Agent Skills Deployed From the Developer Site Don’t Appear in Agent Management Hub, and ASOR Tenant Prerequisites](https://kiweely.com/guides/workday-agent-management-hub-skills-not-showing)

### From the blog on Custom agents and Workday AI

[All posts](https://kiweely.com/blog)

-   [How to Give Employees Live Workday AI Answers Without Flex Credits or Workday Extend](https://kiweely.com/blog/connecting-ai-to-workday-without-flex-credits-or-extend)
-   [Bringing Developer-Built Web Apps and HRIS-Built Screens Together in One Employee App](https://kiweely.com/blog/combining-agent-built-and-developer-built-workday-web-apps)
-   [Using an AI Admin Agent to Configure Workday API Clients and Test Role Boundaries](https://kiweely.com/blog/ai-admin-agent-for-workday-connector-setup-and-role-testing)

## Let Workday check the work.

Kiweely writes WQL and builds orchestrations in your own tenant, and Workday, not Kiweely, says whether they passed.

[Start for free](https://kiweely.com/download) [See how it works](https://kiweely.com/vine/how-it-works)

## Get the next guide in your inbox.

One email a month with new guides and one Extend tip. Nothing else.
