---
title: "First week with Deployment Agent — Kiweely"
description: "Day-by-day setup for Deployment Agent in non-production: agreement, ASOR, the users domain, skills, a first CRAFT prompt, and how to stop Flex Credits."
url: https://kiweely.com/guides/workday-admin-first-week-deployment-agent
---

# A Workday admin’s first week with Deployment Agent

Deployment Agent is a non-production assistant for implementers and administrators. This week plan follows Workday’s setup: confirm UMSA, set up Agent System of Record per tenant, secure Deployment Agent Users, register skills, and prove one tenant-grounded prompt.

-   **Checked** against Workday on November 16
-   Applies to **Workday Extend**
-   8 min read
-   Published November 16
-   By Kiweely

Deployment Agent is Workday’s conversational assistant for implementers and customer administrators. Workday’s Workday-Built Agents reference says it uses a retrieval-augmented generation framework to give implementation guidance and to inspect in-tenant configuration. It runs in Implementation, Preview, Sandbox, and Customer Central when Customer Central is connected to non-production tenants. It is not a Production chat assistant. Workday does not turn it on across your non-production tenants for you. Each tenant needs its own Agent System of Record (ASOR) security and an activation in Agent Management Hub.

Use the first week to make one non-production tenant safe to try, prove one tenant-grounded answer, and decide whether the agent stays active. Leaving it active in Production “just so Sandbox refreshes keep the setup” also keeps a path that can consume Flex Credits until you deactivate it there too.

## Before Monday: read the agreement type

On the setup topic, Workday tells you to open your profile on Workday Community, select your organization name, and read Subscription Service Agreement.

| Agreement | What the setup topic says |
| --- | --- |
| Universal Main Service Agreement (UMSA) | The feature is automatically available. |
| Main Service Agreement (MSA) | Your organization must opt in to UMSA. |

You can only register agents included in the SKUs your organization obtained. If Deployment Agent is missing from Unregistered Workday Agent, stop and confirm the SKU before you rebuild security.

## Day 1: Agent System of Record in the tenant you will actually use

ASOR does not migrate. The setup topic says you configure it in each tenant. Prerequisites are the Security Configuration domain in the System functional area.

1.  Run **Maintain Functional Areas**. Select **Enabled** for **Agent System of Record**. Click OK, then Done.
2.  Run **Activate Pending Security Policy Changes**. Enter a comment, select Confirm, and complete the task. Enabling the functional area leaves pending security changes. Do not skip this activation.
3.  Open **Domain Security Policies for Functional Area**, choose **Agent System of Record**, and enable the domains from **Domain Security Policy > Enable**. The page names **Agent Compliance**, **Agent Management Hub**, and **Manage: Agents**, and tells you to repeat the enable for **Reports: AI Agent Security** in the System functional area.
4.  Run **Maintain Permissions for Security Group** with operation **Maintain**. Pick one **unconstrained** security group. Workday’s current note is that you can only configure unconstrained groups on ASOR security policies.
5.  On **Domain Security Policy Permissions**, grant **View** and **Modify** for:
    -   **Agent Compliance** (compliance reports and data)
    -   **Agent Management Hub** (the hub that lists and manages agents)
    -   **Manage Agents** (configure, activate, deactivate)
    -   **Reports: Agent Reporting** (agent analytics)
    -   **Reports: AI Agent Security** (for example **View AI Agent User Audit Trail**)
    -   **Setup: Agents** (setup tasks)
6.  Activate pending security policy changes again.

If the people who will chat are not in that unconstrained admin group, that is expected. Admin access and chat access are different policies. Chat access is the Deployment Agent Users domain on day 2.

## Day 2: Deployment Agent Users, then the people who may chat

Sign in to the tenant you are configuring. For Implementation or Preview, sign in to that tenant. For Sandbox, Workday recommends signing in to Production so the configuration survives the next refresh. The agent still is not usable as a Production assistant. If you need it in Sandbox before the next refresh, complete the setup in Production and repeat it in Sandbox.

Create the domain policy only if it does not exist:

1.  Run **Create Security Policy for Domain**.
2.  In **For Domain**, enter **Deployment Agent**.
3.  If the domain does not display, a policy already exists. Skip creation.
4.  If it displays, select it and create the policy.

Edit permissions on the domain named **Deployment Agent Users**. From the domain’s related actions, choose **Edit Security Policy Permissions**. On **Report/Task Permissions**, add a row and select the groups that may open the agent. Workday’s example is **Implementers**.

Assign people if they are not already in those groups:

-   Many people, one user-based group: **Assign Users to User-Based Security Group**.
-   One person, several groups: **Assign User-Based Security Groups for Person**.
-   You can roll several user-based groups into an aggregation security group.

Activate pending security policy changes and confirm.

## Day 3: Register, enable skills, set Available To, activate

1.  Open **Agent Management Hub**.
2.  Open **Unregistered Workday Agent**.
3.  For Deployment Agent, register and confirm. Workday moves it to the registry.
4.  Open the agent profile and configure skills. To enable the Ask skill, enable both **Tenant Inspection** and **Implementation Knowledge**. The setup topic points to Enable Deployment Agent Ask Skill for the detailed toggles.
5.  For each enabled skill, fill **Available To** with the security groups that may use that skill. Those groups are the interaction policy. They must also be on the **Deployment Agent Users** domain. If they are not, Workday shows a validation error and those users will not have access.
6.  Activate the agent and confirm if prompted.

You can then open **Deployment Agent** from global search. Chat histories are isolated by tenant: a Sandbox thread does not appear in Preview. Workday keeps chat histories for 14 days, then deletes them.

Customer Central is its own setup. Configuring the agent only on Customer Central does not configure the non-production tenants connected to it. When you are in Customer Central, name the connected tenant in the prompt. The sample-prompts topic says the agent has no on-screen indicator that Ask is enabled, so include a phrase such as “in my tenant” when you need a tenant-grounded answer.

## Day 4: One proof prompt, then the task it names

Do not treat the first answer as configuration. Workday’s calculated-field example tells you to review the logic, then run **Create Calculated Field** yourself, then rate the answer with thumbs up or thumbs down.

Use a published sample, not a vague “help me configure HCM”:

> I am an implementer working on Financials. Provide step-by-step instructions on how to create a calculated field that counts Scheduled Weekly Days.

If you need the tenant, not the manual, switch to the inspection pattern:

> In my tenant, how does this calculated field work: \[Calculated Field\]

Check three things before you accept the answer:

-   The business object and function type match a field you can see in **Create Calculated Field**.
-   Security groups named in the answer exist in this tenant.
-   You can reproduce the steps without inventing a task name the answer did not give you.

## Day 5: Refresh behavior and the Production copy

| Tenant | Where you sign in to configure | What persists |
| --- | --- | --- |
| Implementation | That tenant | That tenant only. ASOR does not migrate. |
| Preview | That tenant | That tenant only. |
| Sandbox | Production, so the next refresh keeps it | Also repeat in Sandbox if you need it before the refresh. |
| Customer Central | Customer Central, plus each connected non-production tenant | A Customer Central-only setup does not cover the connected tenants. |

If you configured Production only so Sandbox refreshes inherit the agent, write that down. The deactivate topic says an agent left active in Production comes back in Sandbox on the next refresh.

## Day 6: Decide whether it stays on

Deactivate requires **Manage: Agents** and **Setup: Agents**.

1.  Sign in to the tenant.
2.  Open **Agent Management Hub**, **Agent Registry**, and the Deployment Agent name.
3.  Choose **Deactivate Agent** and confirm.

Workday hides the agent from users and leaves it in the hub so you can activate it later. Flex Credit consumption stops in that tenant. To stop consumption altogether, deactivate every tenant where you set it up, including Production if you used Production only for refresh persistence. Otherwise the next Sandbox refresh turns it back on.

## Day 7: Write the operating rule

Keep a one-page note for the team:

-   Who is on **Deployment Agent Users** and on **Available To** for Tenant Inspection and Implementation Knowledge.
-   Which tenants are active right now.
-   Whether Production holds a dormant configuration for Sandbox refresh.
-   That chat history disappears after 14 days, so a decision that matters belongs in your design note, not in the chat.
-   That a thumbs-down is how you tell Workday the steps were wrong. It does not roll back a calculated field you already saved.

## Failure modes

| What you see | What the setup topics point at |
| --- | --- |
| Agent missing from Unregistered Workday Agent | SKU, or UMSA not in place for an MSA customer. |
| Validation error on Available To | The group is not on Deployment Agent Users. |
| Answers ignore your calculated fields | Ask is not fully enabled (you need Tenant Inspection and Implementation Knowledge), or the prompt never said “in my tenant”. There is no UI badge for Ask. |
| Sandbox loses the agent after refresh | It was never configured in Production, or you deactivated Production. |
| Sandbox agent returns after you thought it was off | Production is still active, so refresh reactivates Sandbox. |
| Chat from last month is gone | Retention is 14 days, and histories do not cross tenants. |

## Frequently asked questions

### Can employees use Deployment Agent for time off and pay?

No. That job belongs to other Workday-built agents, such as Self-Service Agent. Deployment Agent’s published purpose is implementation knowledge and in-tenant inspection in non-production.

### Does setting it up in Production let people chat in Production?

The setup topic says the agent is not accessible for use in Production. The Production configuration is there so Sandbox refreshes keep the setup.

### Do I need a separate Integration System User for this agent?

The setup topic does not tell you to create one. Access is the Deployment Agent Users domain plus the **Available To** groups on each skill, and those groups have to match.

### Where do I look if a security answer seems wrong?

Use the Ask skill pattern on the security example: name your role, say “in my tenant,” name the field and the role that cannot see it, and ask which domain policy and which groups have View. Then compare that to the domain policy yourself before you activate security changes.

## Sources checked

Pages in the frontmatter, fetched 2026-11-16. The sample prompt quoted above is the published “New configuration” example.

## Sources checked

-   [Set Up Deployment Agent in Implementation Environments and Customer Central](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/deployment-agent-in-implementation-environments-an/set-up-deployment-agent-on-implementation-environm.html)
-   [Set Up Agent System of Record](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/set-up-agent-system-of-record.html)
-   [Deployment Agent Sample Prompts for Implementation Environments and Customer Central](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/deployment-agent-in-implementation-environments-an/reference--deployment-agent-sample-prompts-for-imp.html)
-   [Deactivate the Deployment Agent in the Agent Management Hub](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/deployment-agent-in-implementation-environments-an/deactivate-the-deployment-agent-in-the-agent-manag.html)
-   [Workday-Built Agents](https://doc.workday.com/admin-guide/en-us/workday-ai/agents/workday-built-agents/reference--workday-built-agents.html)

## Related guides

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

-   [### Set up Workday Agent System of Record for HRIS admins ASOR is the tenant control plane for agents. HRIS admins enable the functional area, activate security twice, and permission an unconstrained group. The configuration does not migrate, so Production, Sandbox, Preview, and Implementation each need the same setup.](https://kiweely.com/guides/workday-asor-setup-for-hris-admins)
-   [### Stop or audit a Workday AI agent: an admin runbook When an agent must stop, deactivate it in every tenant where it was set up. Hiding it is not enough if Production will refresh Sandbox. Use View AI Agent User Audit Trail and View Security for Agent Skill to see what it did and what it can still do.](https://kiweely.com/guides/stop-audit-workday-ai-agent-admin-runbook)
-   [### CRAFT prompts for the Workday Deployment Agent Workday’s Deployment Agent sample-prompts topic, updated 2026-11-06, says accurate answers come from five CRAFT elements. This guide restates that method, shows which elements you can omit, and rewrites the published samples so an administrator can see what each word is doing.](https://kiweely.com/guides/workday-deployment-agent-admin-prompt-patterns)
-   [### Workday-built AI agents catalog: what each agent is for Use Workday’s Workday-Built Agents reference as the configuration catalog. This guide restates each agent’s published purpose, notes SKU and UMSA gates, and separates marketing names from what Agent Management Hub can register.](https://kiweely.com/guides/workday-built-ai-agents-catalog)
-   [### 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)

### More guides on Custom agents and Workday AI

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

-   [Workday Core Reporting Agent: when conversational reporting is enough](https://kiweely.com/guides/workday-core-reporting-agent)
-   [Workday Adoption Agent: Release Impact Analysis for Preview admins](https://kiweely.com/guides/workday-adoption-agent-release-impact)
-   [The 2026 Workday AI map: Sana, ASOR, Workday-built, custom, partner, and external agents](https://kiweely.com/guides/workday-ai-ecosystem-map-2026)
-   [Why Generic IDE Agents Still Need Workday Developer Agent for Extend](https://kiweely.com/guides/why-generic-ide-agents-need-workday-developer-agent)
-   [Complimentary Workday Flex Credits: What They Are and What Happens When They Run Out](https://kiweely.com/guides/complimentary-workday-flex-credits-what-happens-when-exhausted)
-   [AI Agents That Edit Workday Studio XSLT Safely: Diff, AUnit, Then Deploy](https://kiweely.com/guides/ai-agents-edit-workday-studio-xslt-safely)
-   [Best AI Agent Setup for Workday Security Policy and Pay-Access Audits](https://kiweely.com/guides/best-ai-agent-setup-for-workday-security-audits)
-   [Forecasting Workday Flex Credits for HCM Agents vs Financials Workloads](https://kiweely.com/guides/forecasting-workday-flex-credits-hcm-vs-financials)

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

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

-   [AI tools for Workday consultants in 2026: ChatGPT, Claude, Workday agents and Kiweely Vine](https://kiweely.com/blog/ai-tools-for-workday-consultants-compared)
-   [Tenant Inspection vs. Tenant Execution: Where Workday’s Deployment Agent Stops and Kiweely Starts](https://kiweely.com/blog/workday-deployment-agent-vs-execution-agent)
-   [Why Flex Credit Forecasting Gets Harder When Sana, Data Cloud, and Custom Agents Share One Pool](https://kiweely.com/blog/estimating-workday-flex-credits-for-sana-and-data-cloud)

## 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.
