Building AI Applications in the Enterprise Part 2 | WSO2Con Barcelona 2025



AI Summary

Overview

  • Topic: Building Generative AI Applications with Micro Integrator
  • Presenter: Um Nadish

Presentation Agenda

  1. Overview of Generative AI Applications
  2. Introduction to Micro Integrator
  3. Building a Generative AI Application
  4. Implementing RAG (Retrieval-Augmented Generation) Pattern
  5. Discussing Agent Functionality in Micro Integrator

Key Points

  • What are Generative AI Applications?
    • Applications that utilize large language models (LLMs) to understand and respond to human language in various formats (text, voice, image).
    • Examples include chatbots and AI agents for customer interaction.
  • Identifying Use Cases for Generative AI
    • Analyze repetitive workflows and human-intensive processes that could benefit from automation.
    • Focus on areas with high content volume, such as policy documents and FAQ guides, to extract useful insights.
  • Micro Integrator Introduction
    • A tool designed for enterprise integration that has evolved from ESB to current microservices architecture.
    • Provides a user-friendly interface for developers, integrating easily with various AI models and use cases.
  • Building a Simple AI Application
    • Demonstrated integration of an LLM into Micro Integrator for a chat application.
    • Step-by-step process involves setting up APIs for chat, and integrating LLM capabilities using a visual programming interface in VS Code.
  • Implementing RAG Pattern
    • RAG enhances the AI’s responses by enabling it to retrieve relevant information from a knowledge base.
    • Illustrated how to use a vector database for storing and retrieving information effectively during model interactions.
  • Agent Functionality
    • Agents facilitate task-specific actions performed by the AI, like evaluating customer feedback.
    • Configuration involves defining roles and objectives for the agent, leveraging existing mediators to connect different functionalities (e.g., sending emails, saving files).
    • Example: An agent that processes restaurant reviews—escalates negative feedback and saves positive reviews.

Conclusion

  • Emphasized the value of integrating AI applications within enterprise systems using low-code solutions.
  • Encouraged experimentation with AI tools available in Micro Integrator and exploration of GitHub resources for practical implementation.