7 areas, 2 updates so far. Checked every morning; 2 slower-moving areas are checked on Mondays.
Notified launched a beta MCP server that exposes its public-relations and investor-relations data and services to AI assistants. It is another example of enterprise data vendors shipping MCP endpoints instead of custom plugins.
Why it matters: Shows MCP spreading into niche enterprise data verticals.
At its first Config event in India, Figma expanded its MCP server with 20 new partner integrations including Linear, Notion, Mistral and JetBrains Junie. The server is faster and lets agents create code layers, edit specs and generate images on the canvas. CEO Dylan Field also cautioned that MCP is imperfect and favors frontier labs.
Why it matters: Design-to-code agents can now read and write Figma files across many popular tools.
Getting reliable behaviour from a model through instructions, examples, structure and well-managed context.
Grounding model answers in your own data with search, chunking, reranking and evaluation.
LLMs that call tools in a loop to complete multi-step tasks, and the patterns that keep them reliable.
Standard ways to connect models to tools, data and other agents.
Knowing whether your AI system works, and catching regressions before users do.
When and how to change model weights instead of prompts.
Putting the pieces together into architectures you can defend in a design review or interview.