Systems | Development | Analytics | API | Testing

Test automation gets harder when the application portfolio stops looking like a modern web stack. Desktop clients, packaged systems, internal applications, and long-lived business software all create different requirements for how teams build, run, and maintain tests. In an AI-disrupted SDLC, the rest of the stack changes even faster, and the risk concentrated in those systems grows with it.

Mike Dillon is a Global Platform Leader at IBM. While deeply technical and experienced in enterprise architecture, he is not a traditional hands-on software developer. Yet this spring, Dillon paired his domain expertise with IBM Bob, IBM’s AI coding assistant, to engineer his first full-scale enterprise application: WatsonAuditX. Designed to automate the documentation required for complex SAP audits, the app addresses a familiar operational roadblock.

Watch an AI agent run a regression test sometime. Really watch it. It lands on a login page. It reads the entire page structure: every element, every attribute, every nested div someone added in 2019 that nobody has dared to remove since. Then it reasons. It weighs its options. It explains to itself, at some length, that the button labeled "Sign in" is probably the one that signs you in. And then, finally, it clicks. Now multiply that by every step, in every test, in every regression run, every night.

An API is a defined way for one piece of software to request data or actions from another. The Model Context Protocol (MCP) is an open standard that gives AI agents a consistent way to discover and use tools and data sources, so they don't need a custom integration for each one. The short version: MCP doesn't replace APIs. Most MCP servers call APIs behind the scenes. What MCP changes is who the interface is designed for. APIs are designed for developers who read documentation and write code.

Two days at RAI Amsterdam (17–18 September), nearly a hundred sessions, and energy throughout the hallways. AGNT Con + MCPCon Europe is the Agentic AI Foundation's flagship event under the Linux Foundation, and this year it consolidated the European MCP content that previously lived at MCP Dev Summit Europe. If you care about MCP, this was the place to be.

In March 2024, Microsoft announced it would retire App Center in March 2025. For a lot of mobile teams, the first question was simple: what now? I was on the Release Management team at the time, and we thought CodePush was worth pursuing. We wrote up the case, got it onto the roadmap, and then waited. Nobody could say yet whether it would become a valuable product, and a team our size has to pick its bets carefully.

What are we spending? Who’s using AI? Is it improving productivity? Here’s how we use Fivetran’s own AI connectors to turn fragmented AI consumption data into answers we can act on.

AI agents that forget their context are just expensive chatbots. The difference between intelligent agents and simple assistants comes down to memory architecture, specifically how well your system tracks relationships, maintains temporal context, and reasons over evolving information. Graph databases have moved from niche infrastructure to essential AI components, with GraphRAG implementations showing 80% improvements in agent truthfulness and significant accuracy gains over vector-only approaches.