
IBM i DevOps TechTalk #29 - AI - PART 2
In this episode, Ray Bernardi welcomes Alan Ashley for the second part of their conversation about AI on IBM i. Together, they explore how MCP servers work, why they can be viewed as a toolbox connecting AI assistants to trusted application data, and how ARCAD MCP Server uses structured metadata to provide accurate context without exposing source code. They also discuss the importance of governance and security, including controlled access, API keys, audit capabilities, and restrictions on sensitive actions. Finally, they explain how ARCAD’s deterministic tools can reduce token consumption, improve the reliability of AI-generated answers, and help organizations control the growing costs associated with AI models.












