AI Assistants and datasources

Welcome to the Almitee AI Assistant Module documentation. This guide provides comprehensive details on the various types of AI assistants available within the Almitee platform, each configured with different levels of knowledge. You will also find instructions on how to configure and utilize these AI assistants effectively.

Overview

The Almitee AI Assistant is divided into four distinct types, each equipped with its own unique knowledge base capabilities:

  1. AI Chatbot

  2. AI Assistant on an Asset/Space

  3. Almitee General AI Assistant

  4. Fix with AI - helpdesk

Knowledge Base Levels

  • Level 1 (L1): Knowledge base used by the AI Chatbot.

  • Level 2 (L2): Knowledge base linked to an Asset model.

  • Level 3 (L3): Knowledge base specific to each asset/location.

  • Level 4 (L4): Knowledge base specific to the helpdesk

The knowledge hierarchy works from top to bottom (L1 to L4), where the AI Chatbot (L1) has the most general knowledge, and the knowledge depth increases with each level.

AI Assistant Types

AI Chatbot (Level 1)

  • Function: Can be integrated into your website or SAAS product.

  • Knowledge Base: Uses data from knowlegde base L1 to anwer questions

  • Explanation: When a user asks a question to the chatbot implemented on your website, the chatbot will respond with the knowledge it has configured on knowlegde base L1. The chatbot does not have any knowledge of datasources on L2, L3 or L4.

  • Detailed explanation

    • User Interaction:

      • When a user visits your website and asks a question, they interact directly with the chatbot.

    • Knowledge Base L1:

      • The chatbot is configured to have access to a specific set of information, known as "knowledge base L1."

      • This L1 knowledge base contains all the information the chatbot can use to generate responses.

      • It is essentially the chatbot's database or repository of pre-configured knowledge.

    • Response Generation:

      • Upon receiving a user's question, the chatbot searches its L1 knowledge base for relevant information to formulate an answer.

      • The response given by the chatbot is strictly based on the content and data available within this L1 knowledge base.

    • Limitations:

      • The chatbot does not have access to other knowledge bases, referred to as L2, L3, or L4.

      • These other levels (L2, L3, L4) might contain additional information or different types of data, but the chatbot is not configured to retrieve or use any information from them.

      • Therefore, the chatbot's responses are limited to what is available in the L1 knowledge base.

AI Assistant on an Asset/Space (Level 3)

  • Function: Accessible via scanning a linked QR code at specific assets or locations.

  • Knowledge Base: Comprises L1, L2 (models), and L3 knowledge bases.

  • Explanation: When a user asks a question to the assistant after scanning a QR code of an asset or space, the assistant will use the knowledge it has configured on knowlegde base L1, configured on the linked asset model (L2) and the asset/space itself (L3). The assistant does not have any knowledge of datasources on L4 or other L2, L3 knowledge bases that are not linked to the asset/space.

  • Detailed Explanation:

    • User Interaction:

      • A user scans a QR code associated with a specific asset or space. This action triggers an interaction with the assistant.

    • Knowledge Base Layers:

      • L1 Knowledge Base:

        • This is the general knowledge base that the assistant has access to by default. It contains broad information that is not specific to any particular asset or space.

      • L2 Knowledge Base:

        • This level contains knowledge configured on the asset model linked to the scanned QR code. Each asset model might have specific information relevant to that type of asset.

      • L3 Knowledge Base:

        • This level contains knowledge specific to the actual asset or space associated with the QR code. This can include detailed, context-specific information relevant to the particular asset or space.

    • Response Generation:

      • When the user asks a question after scanning the QR code, the assistant generates responses using the combined knowledge from:

        • L1: General knowledge.

        • L2: Information related to the asset model.

        • L3: Specific details about the asset or space itself.

      • The assistant integrates these three layers to provide a comprehensive and relevant answer.

    • Limitations:

      • The assistant does not have access to:

        • L4 Knowledge Base: Any additional data sources beyond L1, L2, and L3.

        • Other L2 or L3 Knowledge Bases: Any other asset models or specific assets/spaces that are not directly linked to the scanned QR code. This ensures that the assistant's responses are tailored to the specific asset or space in question and do not inadvertently include irrelevant information from unrelated sources.

Fix with AI - Helpdesk (Level 4)

  • Function: Ask AI to find a fix for the problem described in the ticket.

  • Knowledge Base: Comprises L1, L2, L3 and L4 knowledge bases, thus providing the most detailed and contextual information.

  • Explanation: When a user asks the helpdesk to fix a problem, it will use the knowledge it has configured on knowlegde base L1 and L4, the asset/space the ticket is linked to (L3) and the linked asset model (L2) if a ticket is linked to an asset. It does not have any knowledge of datasources on other L2, L3 knowledge bases that are not linked to the ticket.

  • Detailed explanation:

    • User Interaction:

      • A user contacts the helpdesk to report a problem and requests assistance in fixing it.

    • Knowledge Base Layers:

      • L1 Knowledge Base:

        • Contains general information and data that the helpdesk uses to address a wide range of issues. This includes general troubleshooting guides, common problems, and their standard solutions.

      • L4 Knowledge Base:

        • Contains additional specialized or advanced information that can be used to resolve more complex issues. This might include detailed technical documentation, advanced troubleshooting procedures, or escalated support resources.

      • L3 Knowledge Base:

        • Contains data specific to the asset or space directly linked to the ticket. This includes the historical data of the asset, previous issues, maintenance records, and other specific details about the asset or space in question.

      • L2 Knowledge Base:

        • Contains data specific to the asset model linked to the problem. This includes model-specific troubleshooting steps, detailed technical specifications, and common issues related to that particular model.

      • The L2 and L3 knowledge base is utilized only if the ticket is specifically linked to an asset or space.

    • Response Generation:

      • When a user submits a ticket to fix a problem, the helpdesk assistant uses a combination of information from:

        • L1: General troubleshooting and problem-solving information.

        • L4: Advanced or specialized information for more complex issues.

        • L3: Specific details and historical data about the asset or space linked to the ticket.

        • L2: Model-specific information if the ticket is linked to an asset model.

      • The integration of these knowledge bases ensures that the response is comprehensive and tailored to the specific issue at hand.

    • Limitations:

      • The helpdesk assistant does not have access to:

        • Other L2 or L3 Knowledge Bases: Any data from other asset models or specific assets/spaces that are not directly linked to the current ticket. This ensures the helpdesk assistant’s responses are relevant to the specific asset or space associated with the user’s problem.

Almitee General AI Assistant

  • Function: General AI assistant of Almitee, accessible only when authenticated on Almitee.

  • Knowledge Base: Includes all levels of knowledge bases (L1, L2, L3 and L4).

Configuring Knowledge Bases

Knowledge bases can be enhanced with various data sources:

  • Websites: Add webpages for automatic information extraction.

  • Documents: Upload PDFs for AI learning.

  • Custom Questions: Create and add custom questions to assist users.

  • YouTube Videos: Include videos for content learning.

Data Source Training

Adding a new data source triggers an AI learning stage. Each data source can be independently retrained by clicking the retrain icon.

Document Handling

  • Adding documents to Almitee (PDFs) initiates an automatic learning stage for the General AI Assistant.

  • Documents added as a data source in any knowledge base allow other AI assistants to learn from them, making the information publicly available across corresponding AI levels.

Implementation of AI Chatbot

The AI Chatbot can be implemented using two lines of code available under the "Implementation" tab on the Chatbot page. Access to our API for this implementation requires a paid organizational plan.

Security and Access

For security reasons, we ensure that only organizations with a paid plan can integrate the AI Chatbot. Contact our support team for assistance if issues arise.

Conclusion

The Almitee AI Assistants are designed to provide tailored and scalable support across various platforms and specific locations. By leveraging detailed documentation and diverse data sources, these assistants enhance user interaction and operational efficiency.

For further assistance or inquiries, please reach out to our support team at support@almitee.io.

Last updated