Author: Karl-Johan Spiik
-
Autonomous Agents in Action
A Deep Dive into the Rally Company Scenario with Business Central Autonomous agents are often discussed at a high level: orchestration, reasoning, memory, and autonomy.…
-
Do You Need a Custom Agent to Detect Anomalies?
No — you usually don’t. Power Platform and Dataverse already include built‑in mechanisms for detecting abnormal or unexpected usage. Whether you need a custom solution…
-
Play it (w)right — Automating End‑to‑End Tests in the Power Platform
At CollabDays Bremen 2026, I attended the session “Play it (w)right: automate your end-to-end tests in the Power Platform.” held by Arjan Rijsdijk (slides). This…
-
Oops!…I Vibe Coded Again
At CollabDays Bremen, I joined a session titled “Oops!…I Vibe Coded Again: AI-Assisted Development in Power Platform & M365.” prented by MVP Timo Pertilä. Going…
-
From Likes to Insights: Measuring Real User Experience in Conversational AI
At CollabDays Bremen this year, one session stood out for its clarity, practicality, and truth‑telling about a challenge many AI teams quietly struggle with. Katerina…
-
Power Platform Governance & Security Architecture
I attended CollabDays Bremen last weekend. Session led by Stalin Ponnusamy, MVP provided a comprehensive and practical deep dive into Power Platform governance, breaking down…
-
You are ready for AI! But are your users?
I attended CollabDays Bremen last weekend. I attended Leo Visser’s session without any expectations as I just arrived at the venue. Internal Reflection: Key Takeaways…
-
Fixing Prompt Injection Vulnerability
I’ve been building agents a while and after CollabDays Portugal I had the idea of hacking my own Copilot Studio agent. Other MVPs discussed how…
-
Copilot Studio Agent Security
Below are the mitigations that significantly reduce exfiltration risk. These should be in place before deploying any Copilot Studio agent to a production environment. 1.…
-
Obfuscated request patterns and rapid‑fire multi‑turn scripts
There was still couple tests to run for my agent before thinking how to fix problems. Like in any testing, it is important to run…
