Tag: Explainer
-
AI assistants are useful. Autonomous agents are a step change. If Copilot is the colleague you ask for help, autonomous Copilot agents are the colleagues who just get on with the work. Not in a sci-fi way. In a very practical, enterprise-ready way. This article is about: First: What Is an Autonomous Copilot Agent? Autonomous…
-
There’s a common view in tech: once you build an AI model or an agent, the hard part is done. But any technology leader knows the real challenge begins after deployment, especially with Generative AI, where outputs aren’t predictable facts but fluid language and reasoning. That’s why observability (the ability to see, understand, evaluate, and…
-
AI projects rarely fail because the technology isn’t good enough. They fail because the problem wasn’t clear, the data wasn’t ready, or expectations didn’t match reality. After working with organisations at very different maturity levels, one thing is consistent:successful AI projects follow a lifecycle, whether people realise it or not. Here’s what that lifecycle really…
-
🚀 1. Microsoft Copilot Isn’t Just GPT-Powered Anymore Until recently, Microsoft 365 Copilot and Copilot Studio leaned heavily on OpenAI’s GPT models (including the latest versions like GPT-5). Now, Anthropic’s Claude models (such as Claude Sonnet 4 / 4.5 and Claude Opus 4.1) have been added as optional model choices in key Copilot experiences. This…
-
Ever been asked: “What does this user actually have access to in SharePoint?” And you know the answer may potentially be complicated 😅to report on for admins. Between site memberships, M365 groups, nested security groups, sharing links, and unique permissions, it’s rarely a simple response. Until now, answering that question properly usually meant PowerShell, compiling…
-
If you’ve ever hit the limits of simple connectors, repeated refresh logic, or If you’ve ever hit the limits of simple connectors, repeated refresh logic, or manual data wrangling across environments, chances are a Dataflow is exactly the tool you needed. Dataflows are the Power Platform’s way to centralise, clean, reshape, and standardise data before…
-
If you’ve been following Microsoft Ignite this year, you may have noticed something big quietly land among all the announcements. It’s called Agent 365, and it’s shaping up to be the missing piece we’ve all been waiting for in the world of Microsoft AI governance and enterprise-ready agents. For a while now, organisations have been…
-
Most agent scenarios I hear clients discussing during planning workshops start with chat. Generative copilots for policy Q&A. Retrieval agents that summarise information. And while custom copilots can integrate with Power Automate or use agent flows, the simplest and most direct way to extend Copilot Studio agents is through tools. These let agents move beyond…
-
My team recently came across a new feature in Microsoft’s AI prompt management interface called Test Hub (Preview), and it looks like a game-changer for anyone working with reusable prompts. Traditionally, testing AI prompts has been a bit of a trial-and-error process. You’d tweak a prompt, run it, check the response, and repeat until it…
-
Licensing isn’t the most exciting topic in technology — until it becomes the reason innovation is blocked. For many organisations, Copilot Studio licensing has been that blocker. While the per-user licensing model for Microsoft 365 Copilot was straightforward, around $40 per user per month for productivity enhancements in Word, Excel, Outlook, PowerPoint, and Teams, the…