About Self-Service Agent
The Self-Service Agent is a conversational AI agent that retrieves and summarizes broad self-service data and takes action on common interactions, such as:
- Updating personal information and requesting time off.
- Managing organizational goals and viewing team learning statuses.
- Viewing organization policy information from knowledge articles stored in Workday.
When you activate the Self-Service Agent and enable at least 1 skill, Workday displays the Workday Chat bubble at the top of your tenant and hides the Workday Assistant chat bubble to reduce duplication. When you activate another agent, such as the Payroll Agent, but not Self-Service Agent, you might see 2 chat bubble icons (Workday Assistant and Workday Chat) in your Workday tenant.
Skills and Tools
Workday uses the term
skill
to refer to specific actions or tasks that an agent can take. Each skill is further made up of tools
, which agents use when executing skills. Tools are an abstraction layer (facade) over Workday APIs that the agent calls to perform tasks.The Self-Service Agent skills span varying levels of complexity and intended audiences. Workday makes this distinction in the skill name itself by including these prefixes and suffixes:
- ESS (employee self-service). Skills specifically designed for all employee types.
- MSS (manager self-service). Skills specifically designed for managers.
- Advanced. Complex action skills.
- Base. Simple action skills.
Example: When you want to make the
MSS - Payroll - Advanced
skill available to the Management Chain
security group, you must ensure the the Management Chain
security group has access to the domains that the request_payroll_input tool is secured to.Use the:
- Agent Management Hubto view the descriptions of each Self-Service Agent skill and which tools are associated with each skill.
- View Security for Agent Skillreport to view the required and recommended security policies that must be configured for each tool.
Limitations
Self-Service Agent currently doesn't support:
- File uploads.
- Contingent workers.
- The ability to hide or require fields using theConfigure Optional Fieldstask.
- Custom labels (tokenization).
- Additional jobs (multi-position).
- Usage with implementer and proxy accounts.
- User preferences indicated in profiles, such as locale.
- Delegated actions. While actions initiated by a delegate through Self-Service Agent might initially appear successful in the chat window, they won't create actual events or requests in the system. When you perform an action in Self-Service Agent and specify that you’re acting on behalf of someone else, the agent explicitly states that the action can’t be performed.
- Data extractions from articles in Sandbox. We recommend using IMPL tenants instead for test cases related to this.
- Reading document attachments on help articles.
Self-Service agent uses a non-deterministic LLM model which might produce slight grammatical and syntactical variations in responses. Additionally, Self-Service agent's reasoning and responses are updated continuously over time, which might results in slightly differing interactions as the agent improves week over week.
If a skill or tool involves tenant access restrictions, such as multi-factor authentication (MFA), ensure that any security groups that you assign to that skill or tool on the
Available To
column of the Configure Agent
task are also set up for those same tenant access restrictions so that they are enforced appropriately. We recommend that you opt in to the service on the
Innovation Services And Data Selection Opt-In
report before you sign the UMSA or change your data hosting region. This prevents a 14 day waiting period limitation for Self-Service Agent to ingest your policy documents.Language Support
The Self-Service Agent is optimized and validated for:
- English
- French
- French-Canadian
- German
- Japanese
Currently, the Policy Intelligence skill only supports English and French-Canadian. We plan to support additional languages for the Policy Intelligence skill in the future.