Configuring Employee Navigator Supervisor agent

As an administrator, configure the Employee Navigator Supervisor agent to transform the Employee Navigator chat into a dynamic, more human-like experience. It identifies the most effective resolution path through contextual understanding, enhancing user experience. This topic explains the agentic AI configuration for the Employee Navigator.

Adopting an agentic AI approach for upgrade customers

• The previously configured Employee Navigator chat utilizes the chained-prompt method.
• If needed, you can continue using the chained prompt configuration with existing customized prompts. Employee Navigator chat supports both approaches, allowing flexibility based on your implementation needs.
• For a better user experience, you can opt for an agentic AI approach.
• To configure the Employee Navigator with an agentic AI approach, see To configure the Employee Navigator Supervisor agent.
• Note that after moving to this approach, you must redo the customizations.
• For customizations, you must clone the out-of-the-box agent. Always make changes to the cloned version. For more information, see To clone an agent and Creating and managing skills.

How the Employee Navigator Supervisor agent works for seamless support

The Employee Navigator Supervisor agent is configured in BMC HelixGPT Agent Studio. Based on the user's query, it collects relevant information and context, engages built-in tools, and delegates the flow to the appropriate sub-agent. This coordination ensures that Employee Navigator chat delivers the most suitable resolution, providing seamless support.

This agentic process follows an autonomous decision-making path based on GenAI capabilities:

• Intent interpretation: The Supervisor agent uses GenAI to understand the user's ask and determine whether it is clear or ambiguous.
• Clarification: If the user's input text in the chat is ambiguous, the agent questions the user for additional context to refine intent.
• Retrieval of information from sources: The response is generated by retrieving information from various sources, such as knowledge articles, service requests, approvals, To-dos, or service health items. For example, if the clarified ask matches a known resolution or process, the agent uses integrated knowledge tools to present a response retrieved from the relevant knowledge articles.
• Catalog suggestions: When a user enters a query, the agent configuration compares the terms against published catalog services in BMC Helix Digital Workplace Catalog. Matching is based on the service profile details like name, tags, and description. If the query contains similar terms, Employee Navigator responds with a list of relevant service links in the chat. Fields such as categories and service excerpts are not considered in this comparison. Catalog request handling happens in the following ways: 
  • If the user enters a query for a catalog service that is supported by a chat, the Catalog Request Agent walks the user through the request questionnaire conversationally. Users can submit the request via chat.
  • If the user clicks a catalog service link, they are taken to the catalog submission page in the BMC Helix Digital Workplace console.
• Human fallback support: If the user cannot self-service, the Supervisor agent provides options to connect with a human agent or raise a generic service request.

This flow ensures that users receive timely assistance regardless of the complexity or clarity of their initial query.

Scenario

Allen, an administrator at Apex Global, configures the Employee Navigator Supervisor agent in BMC HelixGPT Agent Studio. With this configuration, Employee Navigator in BMC Helix Digital Workplace can interpret user prompts, route them intelligently, and presents responses based on user intent.

Scenario of the knowledge retrieval

Britney, an HR Coordinator at Apex Global, enters ambiguous text in the chat:
“HR access”
The Employee Navigator asks:
“Are you looking to request HR portal access or find steps to grant access to a team member?”
Britney replies:
"Steps to grant access”

Employee Navigator pulls steps from a knowledge article and responds. Britney follows the presented steps and completes the task efficiently.

Scenario of the catalog request handling

Scenario when the catalog service is chat-enabled

Britney is working on the onboarding logistics for a new contractor. She opens the Employee Navigator chat. She types: “Laptop”
The term is too generic, so Employee Navigator responds with a clarifying question: “Are you looking to request a new laptop, report an issue with your current device, or find setup instructions?”
Britney replies: “Request a new laptop.”
The agent matches her input to existing services and presents a list of relevant catalog services:

• New laptop for a contractor
• New laptop for a full-time employee
• New laptop for client deputation

She clicks the most suitable one from the list: New laptop for a contractor request. This catalog service is supported in the chat. The Employee Navigator guides her through item selection, quantity, and delivery details. Britney submits the service request directly within the chat. 

Scenario when the catalog service is not chat-enabled

Mary, the admin officer, is preparing welcome kits for new hires. She enters a question in the Employee Navigator chat. "Can I request company merchandise for onboarding?"
The agent recognizes the prompt and presents a list of catalog services. Britney selects the Get company merchandise catalog service. Because it is not supported in the chat, after clicking the link, Mary is redirected to the Request submission page to submit it.

Scenario of connecting to a live agent as a fallback option

Britney, an Apex Global employee, enters a question in Employee Navigator chat: “Can I change my desk? It’s right under an AC vent, and it’s uncomfortable.” With the enhanced GenAI capabilities, the Employee Navigator Supervisor agent understands the context and finds no catalog service or article related to the query. The Employee Navigator chat offers Britney the options to connect with a live agent or to raise a generic service request. She opts for a live agent, and then it asks her to select a category for the agent. Britney selects the Facilities category, and a live chat session begins.

Supported skill, agent, and models

Component	Name	Reference
Out-of-the-box agentic skill	Employee Navigator for Helix	Out-of-the-box skills in BMC Helix Digital Workplace
Out-of-the-box agent	Employee Navigator Supervisor	Creating and managing agents by using the Agent Studio wizard
Models	GPT-4.1 GPT-4.1 mini Gemini 2.5 Flash	Models in BMC HelixGPT

Sub-agents used in the Employee Navigator Supervisor agent

Sub-agents are specialized components that handle specific tasks within the overall conversation flow. The following sub-agents are set up in the out-of-the-box configuration for the Employee Navigator Supervisor agent:

• Catalog Request Agent
• Live Chat Agent

Sub-agents configured in the Employee Navigator Supervisor agent

Catalog Request Agent for catalog request handling

The Catalog Request Agent supports conversational catalog submission within Employee Navigator. Employee Navigator Supervisor agent leverages it to interpret user intent, suggest relevant catalog services, and submit service requests via chat. When a user queries about submitting a request, Employee Navigator chat presents a list of relevant catalog services.

• If users select a chat-supported catalog service, it guides them through a conversational questionnaire. This questionnaire is pulled directly from the catalog service, as designed by catalog administrators in the Question Designer.
• If the selected service is not supported in chat, the user is taken to the catalog submission page in the BMC Helix Digital Workplace console. Catalog services that include questionnaires with the following question types are not supported for the request submission via chat:
  • Table (multi-select)
  • Table (single-select) 
  • Repeatable group of questions

The Employee Navigator chat presents each question in a conversational format, allowing users to respond naturally, ensuring clarity and consistency throughout the submission process.
To learn more about the Catalog Request Agent, see Using BMC HelixGPT to make services requestable via chat.
 

Click to view example of catalog submission via chat.

Live Chat Agent for connecting with a human agent

The Employee Navigator Supervisor agent triggers the Live Chat Agent workflow when users request a connection to a human support or if no suitable resolution is found for their questions.

Employee Navigator chat behaves in the following ways:

• If a user asks to connect with a live agent in the Employee Navigator chat, the chat responds with contextual follow-up questions and asks for a category to route to the relevant human agent.
  As an administrator, you can set a default category shown to users during live agent routing. You need to copy and then edit the agent. Later, modify the default prompt instructions to include the default category.
• If no satisfactory resolution is found for their queries, Employee Navigator chat offers users the option to connect with a human agent or submit a generic service request as a fallback.

View example of connecting with a live agent

  

After selecting a category, the following conversation happens:

View example of a fallback option

  

 

Considerations to support live chat capability

Note the following considerations for the live chat support capability provided by the Live Chat Agent:

• Localization support is not available.
• Attachments cannot be shared between agents and users.
• This capability is not supported in channels, such as Microsoft Teams.

Tools used in the Employee Navigator Supervisor agent

Tools are functional programs that assist the AI agent in performing targeted actions. The following tools are set up in the out-of-the-box configuration for the Employee Navigator Supervisor agent for the specific tasks:

• Knowledge Retriever Tool—Search, retrieve, and suggest relevant knowledge articles, improving issue resolution by providing accurate information quickly.
• DWP Profile Search—Search and retrieve people profile information from BMC Helix Digital Workplace, personalizing responses and streamlining user interactions.
• DWP Event Tool—Capture and process events in BMC Helix Digital Workplace, triggering appropriate actions based on user interactions with their service requests.
• DWP Service Health Tool—Provide insights into IT service availability and health status to guide user responses.
• Fallback Tool—Offers buttons as a fallback option in the chat.

For more information about managing tools, see Adding tool and toolkit for an AI Agent.

To configure the Employee Navigator Supervisor agent

1. Log in to BMC Helix Innovation Studio and navigate to the Workspace tab. 
2. From the Application Launcher, select HelixGPT Agents Studio.
3. Click Visit deployed application to open the console.
4. From the Application list, select BMC Helix Digital Workplace to add the Supervisor agent.
5. On the Skills tab and perform either of the following steps to select the skill:
   • Out-of-the-box skill—Select the Employee Navigator for Helix skill from the list.
     Best practice: We recommend that you must not modify the out-of-the-box skill.
   • Custom skill—You can select a custom skill that you have created and proceed with the next steps.
     For more information about creating a custom skill, see Creating and managing skills.
6. From the Agent tab, select Employee Navigator Supervisor agent from the list to link it to the selected skill.
   

Result

Click to view the knowledge retrieval example

Click to view the example of the catalog request handling

Related topics

Using Employee Navigator to find AI-generated answers

Configuring BMC HelixGPT to offer options for unanswered questions

Use case coverage matrix