Concept: Deployment Agent in Implementation Environments
Overview and Purpose
The Workday Deployment Agent is an AI-powered conversational assistant that you can access by entering
Deployment Agent
on the global navigation bar. It leverages Google Gemini Large Language Models (LLMs) and a retrieval-augmented generation (RAG) framework to provide instant guidance. The primary purpose of the agent is to equip Workday implementers and customer administrators with configuration intelligence while working in Implementation environments. Implementation environments include:
- Implementation tenants.
- Preview tenants.
- Sandbox tenants.
The Deployment Agent isn't available as a conversation assistant in Production tenants.
The Deployment Agent acts as a specialized technical architect that doesn't just search for keywords, but understands the logic, dependencies, and structural requirements of the Workday platform. It functions as a knowledge assistant that draws from and synthesizes a repository of curated information, including:
- Workday Administrator and User Guides, to provide step-by-step how-to instructions and functional definitions.
- Community articles and Q&A, to leverage accepted, real-world peer solutions and troubleshooting tips.
- Implementer-only documentation, to offer technical guidance previously restricted to service partners.
- Web service descriptions (WSDLs), to interpret SOAP or integration errors and identify missing data fields or security permissions.
Business Benefits
- Provides immediate answers to how-to questions, reducing the time spent searching external sites.
- Helps identify the root causes of web service errors, like SOAP errors, or business process logic failures.
- Offers an environment for users to quickly get up to speed on unfamiliar functional areas.
- Maintains system stability and security by routing implementation tenant AI traffic through the Agent Gateway, ensuring all interactions are authenticated and audited.
Sandbox Tenant Refreshes
When you refresh a Sandbox tenant, Workday permanently deletes all Deployment Agent chat history data. You cannot recover this data after a refresh.
To avoid manually reconfiguring the agent in your Sandbox tenant after each refresh, we recommend setting up the agent in your Production tenant. You can't use the Deployment Agent in the Production tenant. However, establishing the configuration there ensures it persists and propagates to your Sandbox tenant during standard refresh cycles. If you do not configure the agent in Production, you must manually reconfigure it in your Sandbox tenant after every refresh to continue using it.
Limitations
The agent:
- Might occasionally produce inaccurate or incomplete information. Users must validate all suggestions against official documentation.
- Can’t access your specific data or live tenant configurations. It also can’t detect your current security policies or existing instances.
- Has a limit of 60 questions per conversation You may continue to ask questions in the conversation but the Deployment agent can access and use only the most recent 60 questions (and associated answers). This is to maintain system stability and performance during peak usage. Once you reach this limit, you can continue to ask questions by starting a new chat conversation.
- Is designed to assist, not replace, human expertise.