Alhena
  • Introduction
  • Getting Started
  • Architecture
  • Reference
    • Website SDK
      • Configure Proactive Nudges
    • Product FAQs
    • Website chatsdk events
    • Website chatsdk APIs
    • Chat SDK api and events examples
      • Open other external widget once human transfer is initiated
      • Show the Alhena AI widget only when someone scroll the page by 5 px
    • Website SDK - Custom data
      • Website SDK - Customer data with Agent
    • Website SDK - Internationalization
    • API Reference
      • API calls
    • Device Compatibility
  • Tutorials
    • AI Training
      • Training Steps
      • Training Data Sources
        • Websites
        • Youtube videos
        • Google Drive
        • Twitter Pages
        • Discord Messages
        • Confluence Pages
        • Upload Documents
        • Github
        • Zendesk Tickets
        • Freshdesk Tickets
        • Freshchat Tickets
        • Custom data sources
        • Shopify API
        • Woocommerce API
        • PDF Crawling
      • Training Frequently Asked Questions
    • Tuning Alhena AI Post Training
      • Best Practices for configuring the Alhena AI’s personality and guidelines
      • Adding Human Feedback for improving specific Questions
      • Adding to your knowledge base with FAQs
      • Frequently Asked Questions - Tuning Responses
    • QAing Al Conversations
      • Smart Flagging: Streamline Your AI Quality Assurance
    • Integrations
      • Alhena Website Chat SDK
        • Customizing Your Alhena Chat Widget
      • Integrating Alhena AI With Slack
      • Integrating Alhena AI With Discord
      • Integrating Alhena With Freshdesk
      • Integrating Alhena AI With Zendesk
      • Integrating Alhena AI With Email
      • Integrating Alhena AI With Shopify
      • Integration Alhena AI With Trustpilot
      • Integrating Alhena With Gorgias
    • Notifications
    • Alhena Dashboard
      • Managing Team
Powered by GitBook
On this page
  • When to write a guideline?
  • Understanding the Alhena AI Agent's Capabilities
  • Best Practices:
  • AI Agent Identity Example:
  • AI Agent Tone Example:
  • Guideline Examples:
  1. Tutorials
  2. Tuning Alhena AI Post Training

Best Practices for configuring the Alhena AI’s personality and guidelines

Creating effective guidelines for your Alhena AI Agent, requires a clear understanding of the AI Agent's capabilities and limitations. This guide provides comprehensive best practices to write guidelines which enhance your AI Agent performance and ensure its effectiveness.

When to write a guideline?

Think of a guideline as a rule which you would want the AI Agent to follow while answering user questions. A guideline directly impacts the AI Agent behavior, so whenever there is a need to standardize the AI Agent's responses or actions across specific scenarios, write a guideline. Regular review and updating of guidelines are necessary as the AI Agent evolves and as user needs or business objectives change.

It is crucial to understand the Alhena AI Agent capabilities and then decide if you really need to add a guideline because implementing guidelines without a clear understanding of the AI Agent's inherent functionalities and limitations could lead to redundant rules or expectations that either overlap with existing capabilities or are beyond the AI Agent ability to execute.

Understanding the Alhena AI Agent's Capabilities

  • The AI Agent can have conversations, answer questions, greet, and interact in a natural manner. You can customize its name, personality, and how it speaks through settings in the Dashboard. The Agent will communicate based on these customizations.

  • The AI Agent tries to ask clarifying questions to the user in case of confusion because of the information provided to the AI Agent.

  • The AI Agent can direct users to human representatives if Helpdesk is enabled.

  • The AI Agent's capabilities can be extended by enabling agents for the AI Agent. For instance, if we add an "Order Status" agent, the AI Agent can now get and share details about orders, like what's in the order and where it is. It can also ask the user for any information it needs to use these added agents.

Best Practices:

  1. Familiarize yourself with the AI Agent's capabilities to avoid redundant guidelines.

  2. Make the guidelines clear, concise, simple, and readable. Think about how you would instruct a human!

  3. Keep guidelines brief; complex sentences are challenging to follow.

  4. If a guideline becomes lengthy, consider dividing it into separate, more specific guidelines.

  5. To adjust the AI Agent tone or identity, use the designated fields on the Dashboard for Add Tone to Your AI Agent and Add Identity for Your AI Agent instead of adding new guidelines.

  6. Familiarize yourself with the AI Agent's capabilities to avoid redundant guidelines.

  7. Avoid incorporating frequently asked questions into the guidelines, as this may overwhelm and confuse the AI Agent. Consider a situation where you are given 20 instructions to follow; it's likely you'll become confused and make errors. Instead, use re-training of the AI Agent with updated knowledge or provide human feedback to correct the answers. Here is a short video which explains the human feedback process

  8. Learn from the examples provided below in the guide.

AI Agent Identity Example:

A gaming platform company, established a clear AI Agrent identity to ensure consistent and engaging interactions with developers.

You are supporting developers building on <company_name> platform, the home of web3 gaming, and you're acting as the Prodigy AI Intern.

AI Agent Tone Example:

A gaming platform company, set the perfect tone to make the AI Agent match the tone of its audience.

Accurate, informative, helpful, business informal with a light sense of humor wherever appropriate.

Guideline Examples:

A food delivery company added a guideline to help the AI Agent address the availability of delivery options in a positive manner, improving customer satisfaction.

If delivery is unavailable for a customer, mention 'Please check back again soon as we keep expand our delivery regions.'

A company’s AI Agent operating across multiple communication channels, adapting the response style to fit the medium enhances the user experience.

If the user is asking the question from the Website channel, use emojis whenever appropriate. If the user is asking the question from the Email channel, do not generate any emojis.

A guideline added to help the AI Agent comply with the platform rules set by the company.

Do not mention the names of discord users while quoting their chat messages in your answer.

A guideline added to adapt to the changes in service offerings by the company.

Klarna is no longer a financing option. If you find mention of 'Klarna' in the information provided to you, Ignore the information.
PreviousTuning Alhena AI Post TrainingNextAdding Human Feedback for improving specific Questions

Last updated 13 days ago