Featured
Picked from the blog
-
Same score, opposite meaning: what my Nordic Summit demo actually proved
I gave “From Smart Answers to Trustworthy AI: Designing Governed Copilot Studio Agents” at Nordic Summit 2026 in Billund on 21 September, in the Ideas room. I forgot to press…
-
Copilot Cowork, Claude Cowork and Claude Code: five comparison tables
Most comparisons of AI tools ask which one is smartest or latest. In my daily M365 and Power Platform work that is almost never what decides it. What decides it…
-
Scheduling a Copilot Studio agent test: the evaluation call returns a run, not a result
Agent instructions change constantly. Someone widens a trigger phrase, someone adds a knowledge source, someone rewrites a topic to remove…
-
Defensible beats impressive: designing a governed Copilot Studio agent
A confident wrong answer sounds helpful. It is the worst possible outcome, because nobody thinks to check it. “I don’t…
All posts
-
Trust people, but do not just trust: insider threat at HELish Summit
The 15.00 slot in Runway 1 at HELish Summit 2026 was the one session of the day that was not about building anything. Tatu Seppälä,…
-
Intelligence on tap, and the human who still has to read the code
Vesa Juvonen opened HELish Summit 2026 in Vantaa this morning with a keynote called “Intelligence on Tap: When AI Becomes a Utility”. Vesa is a…
-
A good agent does not know a lot. It knows what matters here.
Two talks on the same evening, from opposite directions, arrived at the same place. I spoke at the PowerUp! Meetup (Finland) Live Meetup on 2…
-
Two new agents spent 400,000 Copilot credits in a day, and nobody used them
On the GitHub Copilot harness, a Copilot Studio agent spends credits while you build, test and evaluate it, not only when it runs. At one…
-
Copilot Studio agents have no voice, so I built the microphone
I have been building Copilot Studio agents for several years now, and in the car I use the voice feature of M365 Copilot a lot.…
-
Your Copilot Studio credit alert is going to the wrong person
Agents on the GitHub Copilot harness consume Copilot credits while you build them, not only when they run. Copilot Studio has per-agent credit limits for…
-
Test-Driven Agent Development
I have been building quite a different kind of Copilot Studio agents, but this latest project was the most challenging one. The knowledge source list…
-
How I Work With A Coding Agent: What’s In My Toolkit After Three Months
Three months ago I wrote about setting up a Linux box on Azure to develop Copilot Studio with Claude Code. That post ended where the…
-
Unified Integrations Done Right: Governance Isn’t Optional, and Neither Is the Middleware
EPPC26 Session Recap — Filip Kocevski European Power Platform Conference, Copenhagen | Thursday, July 2, 2026 | Track: Power Platform | Level 300 Filip Kocevski…
-
Vibe Apps and Code Apps: Trust Is Earned, Not Assumed
EPPC26 Session Recap — Charles Sexton European Power Platform Conference, Copenhagen | Thursday, July 2, 2026 | Track: Power Apps | Level 300 Charles Sexton’s…
- Agent 365 (3)
- AI Builder (6)
- AI Hub (12)
- AI Translator (3)
- Application Insights (2)
- Azure (19)
- Azure DevOps (7)
- ChatGPT (13)
- Claude Code (18)
- Cognitive Services (2)
- Copilot (40)
- Copilot Studio (68)
- Custom Connector (2)
- Database (3)
- Dataverse (37)
- Dynamics 365 (15)
- Error handling (2)
- GitHub (4)
- GitHub Copilot (7)
- Helpers (1)
- Logging (3)
- Logic Apps (4)
- Microsoft 365 (9)
- Offline (1)
- OpenAI (5)
- OpenCode (2)
- Planner (1)
- Power Apps (32)
- Power Automate (69)
- Power Pages (1)
- Power Platform (64)
- Security (16)
- SharePoint (10)
- Teams (15)
- VS Code (5)
- Yleinen (19)