Create Your Own Service Design AI Assistant - Building your own service design large language model



AI Summary

Meeting Summary (Outline Form)

  • Introduction
    • Host: Greg Ly
    • Co-hosts: Brandon (Precocity), Brenda
    • Absent: Robin
    • Sponsors: Slalom and Precocity
    • Series: Part two of AI-powered service design
    • Previous session: Available on YouTube (by Leton Lucky)
  • Recap
    • Service Network (S network) Dallas chapter event
    • Over 50 past sessions available on YouTube
    • Topics range from journey mapping to advanced service design
  • Presentation by Preston
    • Presenter: Preston McAfee, Director of Emerging Technologies at Tonic 3
    • Focus: Building an AI agent for service design
    • Overview:
      • Essential AI terminology
      • Tools in the ecosystem
      • Building GPT agents
      • Live demo
  • AI Terminology
    • AI (Artificial Intelligence)
    • Agent: AI system directed at a single purpose
    • Co-pilot: AI working alongside a user
    • Generative AI: Generates text, images, music
    • Large Language Models (LLMs)
    • Retrieval Augmented Generation (RAG)
    • Vector: Numerical representation in 3D space
  • Tools for Local AI
    • LM Studio: Local LLMs
    • GPT for All: Local chat with LLMs
    • Autogen Studio: Create agents with a front end
    • Crew AI: Build agents without a front end
  • Lexi: AI Agent for Service Design
    • Created to assist users in integrating AI into service design
    • Adapts responses to user’s expertise level
    • Grounded on service design knowledge and reports
    • Can be upgraded with voice synthesis (Lexi 2.0)
  • Building a Service Design Team with AI Agents
    • Agents with different roles and tasks
    • Researching service design topics and historical data
    • Generating insights and reports
  • Learning Curve and Tool Evaluation
    • Approximate time to learn and use tools effectively: 1 month
    • Evaluation of tools based on ease of use, performance, community support, etc.
    • Current leading tools: Claude, GPT-4, Sonet 3.5
  • Q&A Session
    • Addressed concerns about data privacy with local AI models
    • Discussed the potential impact of AI on the critical thinking aspect of service design
    • Explored business cases for AI application in service design
    • Mentioned the use of virtual machines for running models with massive compute power
  • Closing Remarks
    • Acknowledgment of the rapid evolution of AI technology
    • Encouragement for attendees to share their experiences with AI tools on LinkedIn
    • Anticipation of future developments and discussions in the field of service design and AI
  • Resources
    • YouTube channel for further technical details
    • LinkedIn for community engagement and sharing experiences