Skip to main content

Conversation Automation

note

For access to the new AI Agent UI, reach out to the Yellow.ai support team.

Yellow.ai's Automation introduces Conversation, enabling you to build Gen AI-powered bots in minutes, even with zero AI knowledge—just by typing what you need. No training is required for you or the bot to resolve customer queries.

How is this achieved?
Gen AI-powered Orch-LLM + Conversations.

Conversations are driven by Orch-LLM (Large Language Model), fine-tuned conversational AI model. It empowers chatbots to manage multiple goals within a single conversation. Orch-LLM understands user queries, retains context throughout the interaction, and triggers the appropriate flows or responses. It keeps track of the entire user context and can dynamically switch between different goals, guiding users to complete their tasks efficiently.

What is an AI Agent?
Our bots have evolved into highly intelligent systems, so advanced that we now call them Artificially Intelligent Agents (AI Agent).

In this section, you'll learn how to set up an AI Agent:

  1. Agent: Configuring the character and persona of your AI Agent.
  2. Conversations: Customizing conversations by defining goals, rules, and instructions to handle queries.
  3. Configuration: Fine-tuning the responses generated by the selected LLM.

image

Key features of AI Agent​

Dynamic chat and static flows are a thing of the past. The AI Agent is designed to eliminate the need for manually creating flows. Now, you can simply write prompts and publish them live with ease.

Key features of AI Agent are:

  • No training needed: The AI agent understands user queries without requiring specific utterances or intents to be trained.
  • High intent identification accuracy: The agent accurately identifies user intent in most cases.
  • Context retention: It maintains conversation context, ensuring smooth, continuous interactions.
  • Context switching: Seamlessly handles multiple topics within a conversation without losing track.
  • Flow orchestration: Efficiently manages multiple conversation flows and triggers the right one based on the user's context.
  • Small talk & off-topic handling: Engages users with small talk and handles off-topic queries effectively.
  • Dynamic flow switching: The AI agent can switch contexts and adjust responses dynamically at any point in the conversation.

Move NLP bots to Agent AI​

To transition existing NLP bots to Agent AI, follow these steps:

  1. Add Agent AI page details: Ensure the Agent AI page is configured with the Agent persona details.
  2. Move flows to conversations: Transition existing NLP flows to Conversations. Test the flows in the new system to validate proper behavior.
  3. Add descriptions to static flow triggers: For any static flows, add clear descriptions in the triggers. Review and update trigger conditions to ensure they align with the dynamic nature of Agent AI.
  4. Configure start and welcome messages on the Agent AI: Update the start and welcome messages on the Agent AI interface.