Tag: agent
-
Context engineering vs Prompt engineering
For years we have called “prompt engineering” the ability to get better results from language models by choosing the right words. Today, however, real applications (enterprise copilots, agents, RAG, multi-step workflows) reveal a clear limitation: it’s not enough to phrase the request well if the model doesn’t “see” the right information. This is where context…
-
Zefiro: HR that breathes
At the AI Conference held at Microsoft House Milan, I presented Zefiro, the HR agent I designed to simplify corporate life. Born within the CloudiaBrain framework, Zefiro is much more than a digital assistant—it’s a true ally for every employee. With a single, fluid conversation, Zefiro allows you to: Its impact is tangible: fewer errors,…
-
Copilot Studio – Show Image In A Chat
How can I provide visual support in a chat with a conversational agent? In other words, can I give the user the ability to see an image supporting a specific topic? Certainly! Let’s see how. Premises There are different approaches, and this makes finding the best solution a bit critical, but let’s focus on the…
-
Copilot Studio – Computer use
Microsoft has been making Computer use available within Copilot Studio for a few hours now. “The Computer use allows agents to interact with websites and desktop apps by clicking buttons, selecting menus, and typing in on-screen fields. This allows agents to manage tasks even when an API is not available to connect directly to the…
-
Copilot – Write effective instructions (Step by step)
The essential guide There are several topics I am closely following and paying attention to for the creation of an Agent. Specifically, we have two: 1- Writing effective instructions (Declarative Agent – Microsoft 365 Copilot Extensibility). 2-Optimizing prompts and topic configuration (Copilot Studio). Regardless of which tool we will use, it is good practice to…
BEYOND THE PLATFORMS 