Our platform allows you to enhance the AI's responses by providing custom training instructions that guides the agent based on specific scenarios you define.

Introduction

Locusive's training instructions enable you to create scenario-specific instructions for what our agent should do to properly respond to your users' requests. By creating pairs of scenario descriptions and corresponding instructions, you can guide Locusive to act or respond in specific ways when it encounters certain situations. This feature allows you to align the AI's behavior with your organization's policies, tone, and objectives.

How It Works

When a user makes a request, our system:

1. Analyzes the Request: The system examines the user's input to identify the most relevant scenarios from your set of training instructions
2. Retrieves Relevant Instructions: It fetches the most relevant training instructions that you have set up based on the user's request
3. Injects Instructions into the Prompt: The relevant instructions are incorporated into the prompt sent to the LLM, along with a score to identify how relevant they are
4. Generates a Response: The LLM uses the injected instructions to produce a response that aligns with your specified guidelines

This process ensures that the AI provides contextually appropriate responses based on the scenarios you've defined.

Setting Up Training Instructions

To add training instructions:

1. Navigate to Training Instructions: Log in to your Locusive account and go to Settings -> Training Instructions
   
2. Click on the button to add a new training instruction
3. Create a New Instruction Pair:
   • Scenario Description: Provide a short description of a situation where the instruction should apply.
   • Instruction: Write clear guidance on how the LLM should respond in that scenario.
4. Save the Instruction: Click "Save" to add the instruction to your set.

You can create as many instruction pairs as needed, and manage them by editing or deleting existing ones.

Example

• Scenario: 
  • Description: When a user asks about refund policies.
  • Instruction: First ask the user if they have a specific item they want to return, or if they're just looking for general questions. If you have already asked the user, and they said they wanted to return a specific item, then ask for their order number, look up the date they purchased their item from our database, and then see if they fall within the 30 day window, if they do, provide them instructions on how to send in a return from our normal returns policy. Otherwise, direct them to view our latest policy.
    

Best Practices

• Be Specific: Clearly define scenarios and provide precise instructions to avoid ambiguity.
• Use Positive Language: Focus on telling the LLM what to do rather than what not to do.
• Test Thoroughly: After adding instructions, test them by simulating user requests to see how the AI responds.
• Regular Updates: Review and update your instructions periodically to ensure they remain relevant.

Conclusion

Locusive's training instructions feature empowers you to customize the AI's behavior, ensuring it aligns with your organization's values and meets your users' needs. By defining scenario-specific guidance, you can enhance the quality and relevance of the AI's responses.