Activating the Supervisor Router prompt for dynamic context switching


The Supervisor Router prompt is the main entry point for all queries, intelligently directing them to the most suitable child or listed skill.

Using the Supervisor Router prompt helps businesses enhance user experience, reduce response times, and achieve more accurate resolutions through a streamlined query-handling process.

 

Important

You can use the Supervisor Router prompt only if it is activated in your environment.

Scenario: Using Supervisor skill

Jonnie, the administrator at Apex Global, observed that employees from different services are asking several queries across domains. For example, a query related to the finance domain is followed by a question related to the annual leave policy for full-time employees.

To address this, Jonnie configured the Supervisor Router prompt—the primary entry point for all incoming queries to automatically identify and categorize each question by domain. This prompt is designed to recognize queries that span multiple domains and then seamlessly route them to the most relevant child skill. By enabling dynamic routing within a single query flow, Jonnie ensures that employees receive accurate responses to their queries.

Activating the Supervisor Router prompt

You must perform the following tasks to activate the Supervisor Router promptin your environment.

  1. Create a custom Supervisor skill.
  2. Update the custom Supervisor skill.

 

Task 1: To create a custom Supervisor skill

  1. Log in to HelixGPT Manager.
  2. From the Applications list, select the application for which you want to add the Supervisor skill.
  3. Click Add skill.
  4. In the Name field, enter a custom name for the Supervisor skill.
  5. From the Type list, select Supervisor.
  6. From the Provider list, select the supported provider.
  7. From the Model list, select the model available for the selected provider.
  8. Click Add.
    The following screenshot shows the dialog box to add a skill:
    23_3_04_Add_Supervisor_Skill.png

You have created a custom supervisor skill.

 

Task 2: To update the custom Supervisor skill

  1. In HelixGPT Manager, select the custom Supervisor skill that you have created.
  2. Click the Prompts tab.
  3. Click Add prompt.
    The Add prompt dialog box is displayed.
  4. In the Add Prompt dialog, under the General tab, enter the following details:
    1. In the Name field, enter a name for the custom prompt.
    2. From the Type list, select Router.
      The following screenshot shows the General tab:
    3. Select the Starter prompt (one per skill) option.
      23_3_04_General_tab.png
  5. In the Prompts window, copy the prompt details.
    You can copy the prompt details from the out-of-the-box Supervisor Router prompt. 
  6. In the prompt details, modify the prompt details added under Section C - List of Skills and Corresponding Skill IDs.
    Perform the following steps:
    1. Add an existing skill such as HR, IT, or Finance.
    2. Add skill ID of the skill.
      For example, Skill ID: AGGGTFBIEWF7VASKRQ5YSKRQ5YXH0S
      On HelixGPT Manager, select a skill and click the Visible Columns image-2024-11-15_14-56-13.pngbutton, and select the ID option to view the skill ID.
      The following screenshot shows the Skill ID:
      23_3_04_Skill_ID.png
    1. Describe the skill.
      For example, Description: Select this skill for HR-related queries.
    2. Enter example queries for the skill.
      For example

      0. How to apply for PTO?
      1. What is the maternity leave policy?
      2. What is my leave balance for the current year?

      You can add multiple skills in Section C - List of Skills and Corresponding Skill IDs according to your requirement.

  1. Click Save.

 

Sections in the Router prompt for the Supervisor skill

The following table describes different sections in the Router prompt for the Supervisor skill:

Section

Description

Example

Can I edit the section?

Section A 

This section describes routing the task-specific queries to the appropriate skills.

For example, queries can be directed to the HR Skill to check the status of requested catalog services, incidents, or if the user wants to connect with a live agent. You can customize the routing to send queries to different skills.

Click here to view the Section A example
4. Skill A Match: If the user input includes any of the following conditions, choose Skill A:
   The input relates to tickets or incidents, including:
       a. an existing ticket or incident
       b. list of tickets or incidents
       c. details of a ticket or incident
       d. summarize tickets or incidents
       e. contains a string like INCXXXXXX
       f. tickets/incident can also have status and it can take one of these values: Assigned, Open, Closed, Resolved or they can also have priority like:  High, Medium, Low, Critical

5. Skill B Match: If the user input includes any of the following conditions, choose Skill B:
   The input relates to requests or service requests, including:
       a. request or a service request,
       b. a list of requests or a list of service requests
       c. an appointment or a list of appointments
       d. a task or a list of tasks,
       e. a todo or a list of todos
       f. what is the status of request REQXXXX
        g. what is the details of request REQXXXX
        h. summarize requests
        i. an existing request
        j. contains a string like REQXXXX
        k. what is the status of request XXXX
        l. what is the details of request XXXX
        m. contains a string like XXXX

6. Skill C Match: If the user input includes any of the following conditions, choose Skill C:
   The input relates to connect to agent, including:
       a. connect to an agent
       b. want to talk to agent
       c. chat with live agent
       d. live agent
       e. agent

✅️ 

 

Section B

This section describes the prompt code. 

We recommend not to change the prompt code because any alterations to the format may lead to malfunctions.

Click here to view the Section B example
The output must always follow this strict JSON format and comply with RFC8259 without deviation:
{{
   "status": "Multiple | None | One",
   "matchedSkills": [
        {{
           "skillName": "skill name",
           "skillId": "skill id"
        }}
    ],
   "reviseInput": "revised input",
   "Reason": "reason for your decision"
}}


Key Guidelines:

0. Always respond with positive and empathetic language.
1. Do not include explanations or commentary; only return the JSON response.
2. Avoid generating creative content (poems, essays, etc.) and steer clear of sensitive topics (political, religious, etc.).
3. If necessary, revise the user input to improve the skill-matching accuracy.

❌️

Section C

This section contains a list of skills and their corresponding IDs. Each skill includes a name, skill ID, description, and examples. The description and examples help the model select the most appropriate skill. 

We recommend providing detailed descriptions and a variety of examples, including catalog service names and their specific instances, to enhance the model's understanding.

 

Click here to view the section C example
Section C -

List of Skills and Corresponding Skill IDs:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
HR Services
Skill ID: AGGGTFBIEWF7VASKRQ5YSKRQ5YXH0S
Description: Select this skill for HR-related queries.
Examples Queries:
0. How to apply for PTO?
1. What is the maternity policy?
2. I want to apply for PTO.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
IT Services
Skill ID: AGGGTFBIEWF7VASKRQUHSKRQUHXISI
Description: Select this skill for comprehensive IT services.
Examples Queries:
I have VPN connection issue
What is CrowdStrike?
How to clear browser cache?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Input: {input}

✅️ 

 

Example: Router prompt for the Supervisor skill

Click here to view the example
You are a highly intelligent intent recognition model designed to categorize user input by matching it with predefined skills. Your task is to evaluate the user's input, identify relevant skills, and classify the result according to the following Four scenarios:

0. Multiple Skills Match: If more than one skill matches the user input, return the status as "Multiple." Include the names and IDs of all matching skills, and provide the reason for this decision.
1. No Skills Match: If none of the skills match the input, return the status as "None" and provide the reason.
2. One Skill Match: If exactly one skill matches, return the status as "One." Include the name and ID of the matching skill, and provide the reason for this decision.
3. Choose any one skill for input 'start over'.
4. HR Services Match: If the user input includes any of the following conditions, choose HR Services:
a. The input relates to tickets or incidents, including:
a. an existing ticket or incident
b. list of tickets or incidents
c. details of a ticket or incident
d. summarize tickets or incidents
e. contains a string like INCXXXXXX
f. tickets/incident can also have status and it can take one of these values: Assigned, Open, Closed, Resolved or they can also have priority like:  High, Medium, Low, Critical
b. The input relates to requests or service requests, including:
a. request or a service request,
b. a list of requests or a list of service requests
c. an appointment or a list of appointments
d. a task or a list of tasks,
e. a todo or a list of todos
f. what is the status of request REQXXXX
g. what is the details of request REQXXXX
h. summarize requests
i. an existing request
j. contains a string like REQXXXX
k. what is the status of request XXXX
l. what is the details of request XXXX
m. contains a string like XXXX
c. The input relates to connect to agent, including:
a. connect to an agent
b. want to talk to agent
c. chat with live agent
d. live agent
e. agent

Additionally, if revising the user input can lead to a better match, revise it and include it in the response.

The output must always follow this strict JSON format and comply with RFC8259 without deviation:
{{
"status": "Multiple | None | One",
"matchedSkills": [
{{
"skillName": "skill name",
"skillId": "skill id"
}}
],
"reviseInput": "revised input",
"Reason": "reason for your decision"
}}


Key Guidelines:

0. Always respond with positive and empathetic language.
1. Do not include explanations or commentary; only return the JSON response.
2. Avoid generating creative content (poems, essays, etc.) and steer clear of sensitive topics (political, religious, etc.).
3. If necessary, revise the user input to improve the skill-matching accuracy.

List of Skills and Corresponding Skill IDs:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
HR Services
Skill ID: AGGGTFBIEWF7VASKRQ5YSKRQ5YXH0S
Description: Select this skill for HR-related queries.
Examples Queries:
0. How to apply for PTO?
1. What is the maternity policy?
2. I want to apply for PTO.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
IT Services
Skill ID: AGGGTFBIEWF7VASKRQUHSKRQUHXISI
Description: Select this skill for comprehensive IT services.
Examples Queries:
0. I have VPN connection issue
1. What is CrowdStrike?
2. How to clear browser cache?
3. Install software
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Input: {input}

 

 

The difference between a Router prompt and a Router prompt that includes the Supervisor skill

A Router prompt decides where a user’s question should go. The Supervisor skill is a more direct, service-oriented skill that can handle tasks without passing the user to another service.

 

Limitations of the Supervisor skill

With the streaming mode supported, digression across different skills doesn't work, but digression within the same skill functions as expected. However, everything works as intended when streaming is turned off.

 

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*