Charm Universal ML Engine

Elloe AI's Charm Universal ML Engine™is the brain behind your top-performing AI Agent.

Our Charm Universal ML Engine ensures your AI Agent grows, learns and works 24/7 to continuously improve your business.

By combining safety checks, guidance, advanced knowledge retrieval, a suite of cutting-edge LLMs, and action capabilities between your business systems and integrations, we’ve given you the control you need to coach a problem-solving powerhouse. Watch your AI Agent understand and isolate customer issues, gather knowledge from diverse sources and create an action plan to resolve the inquiry—all while remaining relevant, accurate and safe every single time.

Get a Demo

With the Elloe AI Charm Universal ML Engine™, you can trust your AI Agent will tap into the best artificial intelligence models to maximize resolution, and align with your business to deliver personalized customer experiences.

OpenAI Logo OpenAI Logo OpenAI Logo OpenAI Logo OpenAI Logo OpenAI Logo

We’re taking you
under the hoodSee the thought processing power
of the Elloe AI Charm Universal ML Engine™

Understand how your AI Agent thinks by seeing how our Elloe AI Charm Universal ML Engine™ processes a customer inquiry with advanced problem-solving capabilities.

  1. Engine Track

    What do I already know
    about this customer profile?

  2. Engine Track

    What are the problems
    this customer is facing?

  3. Engine Track

    What rules have been set to ensure my
    knowledge is relevant and accurate?

  4. Engine Track

    Does this customer issue
    require me to take action?

  5. Engine Track

    Is this solution accurate, relevant and safe?

  1. 1. Understand the Inquiry

    The Elloe AI Charm Universal ML Engine™ processes each customer query individually to understand the conversation context, apply your personalized guidance and review with safety checks to ensure it aligns with your business.

    Customer Inquiry:
    This is the starting point of the process, where a customer asks the AI Agent for assistance.
    Conversation Context:
    The system assesses the context of the conversation. This involves understanding the current state of the dialogue, the customer’s query, and any previous interactions that may influence the response, as well as the customer’s profile.
    Guidance & Coaching:
    Relevant and applicable guidance provided by the AI manager is applied and the AI Agent reflects on how it has solved similar problems for customers before.
    Safety Check:
    A safety check is applied to both the customer’s question, and the AI Agent’s response. This is a critical step where the system checks for safety, and ensures its response is accurate, relevant and aligns with your business.
  2. 2. Isolate the problem

    The Elloe AI Charm Universal ML Engine™ reasons through the customer issue — no matter how complex the inquiry — isolating each problem it needs to solve.

    LLM Supervisor:
    At the heart of the system is the LLM Supervisor, which orchestrates the AI Agent’s operations. It decides when to search for knowledge, write a reply, take action, or make an API request based on the input it receives.
  3. 3. Retrieve the required knowledge

    The Elloe AI Charm Universal ML Engine™ connects with the knowledge sources it deems most relevant to the issue, while adhering to your customized rules set by the AI manager to ensure the relevance and accuracy of the information in accordance with your business.

    Knowledge Source:
    The AI Agent can search an external knowledge source to gather information necessary to respond to the customer inquiry accurately. This process adheres to rules set by the AI manager to ensure the relevance and accuracy of the information retrieved.
    Query Writing & Retrieval:
    The system may rewrite questions to improve the quality of knowledge retrieved - for example, rewriting “how much does it cost?” to “price of iPhone 15”, using the conversation context.
  4. 4. Create an action plan

    The Elloe AI Charm Universal ML Engine™ determines if the problem requires a specific action (like updating an account or processing a transaction) and connects with your business systems—just like a human agent.

    Use Action:
    If the customer's request requires a specific action (like updating an account or processing a transaction), the system can perform this action in another business system. This process adheres to rules set by the AI manager to ensure the relevance and accuracy of the information retrieved.
    API Request:
    Sometimes, to fulfill a request or to get additional data, the AI Agent needs to make an API request to a business system.
    Business Systems:
    These are the backend systems that manage customer data, process transactions, and perform other business-related functions. The AI Agent interfaces with these systems to execute actions required to resolve the customer's inquiry, just like a human agent.
  5. 5. Automatically resolve the inquiry

    The Elloe AI Charm Universal ML Engine™ reviews to ensure the reply includes all necessary information, aligns with the controls you’ve set and completes a safety check before sending a response back to the customer.

    Write reply:
    Once the necessary information is obtained and a safety check is completed, the system writes a reply to the customer's inquiry.
    Send message:
    Finally, the AI Agent sends the crafted message or confirmation of an action taken to the customer, completing the interaction loop.