Recent content on Luise Freese: Azure & Power Platform Architect
Part 0 showed why constant prompting, re-prompting, and steering GitHub Copilot feels fragile. Not because Copilot is unreliable, but because nothing in the interaction is stable. Every follow-up reshapes the problem again, and the system has no way to know what must remain invariant. This post answers the only question that matters after that: what exactly do you need to change, how do you do it…
Most developers think that one can’t really use GitHub Copilot wrong. There is a chat interface that lets you also choose a model, so the obvious choice is to choose a model (trust the LinkedIn bros on that) write prompts, get some code, adjust a few things, maybe ask a follow-up, rinse and repeat. Some say, it’s like working with a junior dev who has amnesia, some say it feels…
Look, I love a good keyword-based system as much as the next developer. They’re fast, predictable, and when your user says “VPN,” you know they mean network issues. But what happens when someone writes “Can’t access the customer portal from home”? Is that Network? Access? Business Applications? Remote Work Setup? Welcome to the world where simple pattern…
Client secrets are like hiding your house key under the mat; easy to grab and impossible to audit. Certificates are just slightly better, because they at least prove who made the key. Until, of course, someone forgets where it’s stored. Even Azure Key Vault, while fantastic, doesn’t eliminate the problem; it just relocates it. So while secrets always need a mechanism on where and how to store,…
Running az login feels like magic. A browser pops up, you pick an account, and from then on, everything just works . No more passwords, no more pop-ups. It’s frictionless and we all love it. But convenience always has a price, and in this case, the price is understanding what the CLI actually does with your tokens behind the scenes. What really happens when you log in Azure CLI doesn’t stash your…
I added Microsoft Dev Proxy to my Mermaid → Dataverse converter , because I wanted to test how it handled rate limits and API errors. What I got instead was literally an intervention. For everyone who doesn’t know what this solution is about: The converter takes an Entity-Relationship-Diagram written with MermaidJS as an input, parses the information and passes everything into the Dataverse…
What if your function could hit Microsoft Graph with no client secrets, no certs, and no Key Vault entries? That is exactly what a Managed identity is for: Azure issues short-lived tokens to your app on demand, and the platform rotates and protects the underlying credentials for you. The problem: credential hell // the old way — credential nightmare const clientSecret = process . env .…
Why diagrams matter (and why they usually fail us) Entity-Relationship Diagrams (ERDs) are the universal shorthand for talking about data models. They show what entities exist, which attributes they carry, and how they connect to each other. In workshops, they are invaluable: they help everyone get on the same page, and they force the team to make decisions about terminology early on. Is it a…
It’s OK to be seen trying Somewhere along the way, we started believing that you’re only allowed to speak after you’ve figured everything out. That you can only post once the product is perfect. That you can only ask a question if it’s smart enough, and that you can only contribute once you have “value”: measurable, monetizable, LinkedIn-optimizable value. And until then? Stay quiet 🤫. Watch.…
“We just need to test the AI. We’ll figure out the data later.” That sentence has quietly killed more AI pilots than any model failure ever could. Most AI pilots don’t fail because the tech didn’t work. They fail because no one did the groundwork. No clean data. No integration path. No clear problem. No business owner. Just a well-intentioned scramble to “try AI” and show something in a demo,…
There’s a narrative floating around right now: With AI, one person can do the work of a whole team. Code? Done. Design? Prompted. Copy? Generated. Voice-over, editing, visuals? All stitched together in record time. And at first glance, that sounds like progress. AI makes it easier than ever to ship something: quickly, independently, and with just enough polish to post. But it’s also creating a…
You know that neat little curve in project planning tools that suggests you’ll produce value at a steady pace from Monday to Friday? It’s a lie. Not a malicious lie. Just a completely detached-from-reality one. The myth of the 8-hour brain Most jobs still pretend that productivity is something that can be sliced neatly into eight equal hours a day. But if you’ve ever done any kind of deep,…
“Our AI chatbot is live- in a test environment, since … uhm… 18 months.” Sounds familiar? If your organization has a hallway legend about a “pilot” project that never saw daylight, you’re not alone. Across industries, AI pilots have become the business equivalent of Schrödinger’s cat: both alive and dead, depending on who’s presenting the PowerPoint. They give the illusion of…
When distributing tasks or tickets across a team, fairness matters, and so does flexibility. With Power Automate, we can implement a round robin pattern that always picks the next person in line, and wraps around once we hit the last. I got asked by a customer on how to achieve this - so it might be helpful to you as well. What we’re solving Let’s say your team handles incoming forms—support…
If you made it this far in tis series, you’ve seen the messy middle: the risks, the lifecycle debt, the cost of inaction. Now let’s zoom out. Copilot Studio isn’t just another way to automate tasks, but a design surface for how your organization thinks, acts, and responds. But only if you treat it like a capability, not a tool. The real value isn’t in building faster but in thinking better Most…
Agents that outlive their creators – governance, risk, and the long tail of AI In most organizations, apps get decommissioned. Workflows get retired. People move on, and someone eventually archives the mess they left behind. But agents? Agents keep acting, even when their creator is gone. That’s not a hypothetical risk. It’s already happening. Because most Copilot Studio deployments are building…
tl;dr AI can read every policy, spec, and wiki page in your company; it even knows the formal org chart. But it doesn’t feel the informal paths: the backchannels, the trust networks, the unspoken approvals. Institutional knowledge lives in the spaces between people. When you drop an AI agent into an organization and expect it to “just know”, it plays the wrong game. Map the…
Talk to be announced soon - with Julian Kusenberg (SoftwareOne) Register Get your ticket here
Full Day Tutorial: Getting Started with Power Platform & AI on SharePoint Together with Robin Rosengruen (Microsoft) SharePoint is the backbone of collaboration in Microsoft 365, and with the Power Platform, you can take it to the next level: Build custom apps and automate processes to make work easier. In this tutorial Luise and Robin will walk you through the essentials of Power Apps Canvas Apps…
Codemotion 2025 The ultimate event for developers, tech leaders and industry changers Your Testing Strategy is broken - lets fix it! Testing is essential for delivering reliable software, but too often, teams fall into the trap of testing everything and testing it poorly. Endless unit tests for trivial code create a false sense of security while ignoring the areas that matter most; like how your…
Some organizations break things by moving too fast. Most break things by doing nothing at all. Inaction is comfortable. It looks responsible. Safe. “We’re just waiting until the platform matures.” But the bill is still running, just not where you’re looking. Copilot Studio isn’t expensive because of licenses. It’s expensive because of what happens when you treat it like a…
Incident tracking in SharePoint lists often involves too many clicks. Users need to open each item, change a value, wait for the value to be saved, and hope they didn’t miss anything. With JSON column formatting and Power Automate, this process can be streamlined using inline action buttons. Use case A SharePoint list tracks incidents with fields for title, priority (low, medium, high), and…
The comforting danger of “we assumed…” You’ve heard the saying: assumption is the mother of all fuck-ups . It’s crude, yes, but also painfully accurate. Trust me, been there. “We assumed each customer would have one account” “We assumed the Excel template wouldn’t change” “We assumed we would only need one language” These statements don’t come from…
You’ve built a Copilot Studio agent. It works in the demo. It talks to systems. You click “Publish” and it goes live. Except… nothing about this is live in the way that matters. No pipelines. No environments. No version control. No audit trail. No rollback. If that’s your lifecycle, you haven’t shipped an agent. You’ve just made a sandbox louder. What “Publish” actually means Clicking…
There’s a moment in nearly every development task where you hit a small decision: Do I clean this up now? Or do I just get it working and promise myself I’ll come back to it? 🤞🏻 The second option always feels easier. More efficient. Responsible, even: “I’m not skipping it forever, just for now.” But I’ll let you in on a secret that you won’t like: “I’ll refactor this…
A while back, I wrote about A hill to die on: there are no tiny changes , just changes we underestimate. If you missed it, give it a read. But there’s a related myth we need to dismantle: “It’s just a temporary fix.” Ah yes, the sacred chant of backlog triage. We all know how it goes: “This is just to get us unblocked.” “Let’s be pragmatic; we’ll clean it up later.” “We’ll write a proper solution…
Ever wanted to share your LinkedIn profile straight from your Apple Watch, without fumbling for your phone? I’ve built something that makes it simple: QR Watchface , a sleek web app that turns any URL into a perfectly sized, scannable QR code for your watch face. Why QR Watchface? Imagine this: You’re at a conference, or maybe a coffee meet-up, and you want to share your LinkedIn profile,…
It started with a fish. Or more accurately, a fish pic guy . You know the one: holding up a trout like it’s a personality. And when I saw Sara Lagerquist compare him to local choices in Dataverse , I couldn’t unsee it. ( She really did. ) Suddenly, the entire dating app experience started looking like a metaphor for Power Platform architecture. The red flags, the weird flexes, the mysterious…
We need to talk about something that keeps showing up in LinkedIn posts, company updates, and personal branding pep talks: fancy fake bold text. You’ve seen it. You’ve maybe even used it. 𝗟𝗶𝗸𝗲 𝘁𝗵𝗶𝘀, or 𝘁𝗵𝗶𝘀, or even ⓉⒽⒾⓈ. 💡 For everyone using a screenreader: There is a line that uses unicode characters to display something that resembles bold letters. Screenreaders can’t…
Autonomy scares people. Not because they’re against AI. They’re against losing control, and rightly so. Most organizations aren’t afraid of what their Copilot Studio agent might do. They’re afraid of what it might do without asking . And yet, that’s the whole point of agents. If you need a human in the loop for every decision, you didn’t build an agent. You built a clunky wizard. Autonomy, done…
We’ve entered the age of automation. But most organizations are still treating it like the age of macros. A little generative garnish, a few nicely structured prompts, and suddenly every flow is called an agent. It sounds impressive until you try to scale it. Then the cracks start to show. Copilot Studio is not just another way to automate tasks. It’s an entry point into something much more…
Software development is a decathlon (and low-code only gives you running shoes for two events) Building apps with Power Apps feels easy. Drag some controls onto a canvas, connect to a SharePoint List, throw in a few formulas — ta-da! A working prototype. Except… that’s not the race. That’s the warm-up. Shipping real apps (you know, the ones that solve real problems, scale beyond one user, and…
Fewer tests. Faster feedback. Testing is broken; not because we don’t do it, but because we waste time doing it wrong. We obsess over unit tests that don’t matter. We write coverage reports that lie. We mock everything until our app is disconnected from reality. Then we go live and wonder why things break. Again. The test pyramid wasn’t meant to be worshipped The old test pyramid said: write lots…
We love to pretend that business apps are boring. Predictable. Safe. Low risk. After all, we’re not launching rockets or handling billions of users… we’re just managing leave requests, budgets, logistics, workflows, customer data, and access rights. You know, just the things entire companies run on. But here’s the uncomfortable truth: Business apps are some of the riskiest systems we build And…
There’s something oddly reassuring about blaming the vendor . The timeline slipped? Must be their fault. The app doesn’t work on tablets? Scope misunderstanding. We spent $32 million and still don’t have a working product? Clearly a delivery issue. But behind every disaster labeled vendor failure , there’s usually a deeper truth: someone on the client side stopped paying attention. Or was…
Most teams treat testing as a cost. Something you do because you have to — like documentation or status reports. It’s seen as a necessary drag on velocity. A checkbox to tick before the sprint ends. A stage in the pipeline you hope turns green quickly so you can merge and move on. But what if we’ve been seeing it wrong all along? Good testing isn’t a drag. It’s leverage It doesn’t slow you down…
Your password isn’t protecting you. At least not in the way you think. Not anymore. We’ve built a whole mythology around passwords — how long they should be, how often they should change, what characters to include. We treat them like they’re the foundation of our digital security. They’re not. Source: https://xkcd.com/1200/ Usually, I don’t run security workshops. Most of the time, I’m…
The pyramid that never was You’ve probably seen Maslow’s hierarchy of needs: a five-level pyramid that starts with food and shelter and ends with self-actualization. It’s tidy. It’s visual. It’s wrong. Maslow never drew a pyramid . He described needs as overlapping tendencies, not a strict ladder. You don’t need to finish belonging before pursuing esteem. You can feel creative even if you’re…
Copilot Studio gives you a fast, secure way to build conversational agents and deploy them into the Microsoft 365 environment without writing code. But what happens when the default models aren’t enough? That’s where Azure AI Foundry comes in. Copilot Studio now supports direct integration with models deployed via AI Foundry. That means you can choose exactly the model you need, like a…
Test coverage is not quality That’s a hard truth most teams don’t want to hear. But we need to. Because too many development teams are stuck in the wrong mindset: chasing 100% test coverage as if it guarantees safe deployments, stable apps, and happy users. Spoiler alert: It doesn’t. In fact, the more you worship coverage metrics, the more likely you are to miss the actual point of testing: to…
The feed vs. the floor Open LinkedIn and you’ll see the same recycled energy over and over: AI will reinvent your entire industry! The companies that don’t adapt will die! Here’s how I use AI to 10x my thinking in under 3 minutes! Then, in the comments: 🔥 “Incredible insight!” 🤝 “Exactly what I’ve been saying!” 🧠 “Have you tried asking ChatGPT politely ?” And so many rocket 🚀🚀🚀 emojis.…
Somewhere between your fourth copy-pasted email of the day and yet another spreadsheet update, you wonder: should I just automate this? It’s a fair question; and honestly, a tempting one. But context is everything. Before diving in, it’s worth pausing and asking yourself a few grounded questions. Think of this as a guided conversation with your future self (and hopefully someone from governance).…
Where does the path toward a proper data strategy begin? With technology, processes, or mindset? It’s a good question, which I received in a discussion on LinkedIn and it gets right to the heart of why so many organizations feel stuck. Excel got us this far. but it won’t get us further Excel filled the gaps. When systems were too rigid, when processes weren’t documented, when no one had the time…
I often hear some version of this: We should probably modernize it… but to be honest, the Excel spreadsheet works just fine . It’s usually said with a shrug, as if running a key business process on a two-decade-old file is just how things are done. We joke about it. We meme about it. But the truth is this: some of your most essential operations are still running on duct-taped Excel files built…
Here’s a fun experiment: when was the last time you memorized a phone number? Or did math in your head? Or remembered your colleague’s birthday without Outlook whispering it to you like a digital butler? Exactly. Welcome to the age of cognitive offloading, where our brains have quietly delegated entire functions to apps, calendars, and lately: AI. It’s efficient, it’s seductive, and depending on…
Don’t get knocked out by licensing costs Power Platform licensing is often seen as the scary part. Mention a 5€ per user license, and someone inevitably reaches for a calculator: “Wait, with 3,000 users, that’s €15,000 a month — €180,000 a year… for a single app?!” That reaction is common. And completely understandable, if you stop at the price tag. But if that’s the entire discussion,…
From chaos to clarity – how to build AI projects that actually work If you’ve worked on an AI project recently, this probably isn’t a joke, it’s your backlog. For all the buzz around AI, many initiatives are still chaotic, short-lived, and poorly aligned with real business needs. The problem isn’t the technology but how we approach it. And the landscape isn’t getting simpler. Especially in the…
There’s no such thing as a small change “ Can we just move that button? ” “ It’s just a quick fix. " “ Tiny change, shouldn’t take long. " If I had a euro for every time I’ve heard that, I’d have long since retired from software, though probably not from cleaning up after so-called small changes. Let me say this loud and clear for the people in the back: There is no such thing as a small change in…
Lately, I keep seeing weekend seminars popping up: Promising to turn you into an “ AI leader ”, prepare you for the role of Chief AI Officer , or help your company tick the compliance box for the EU AI Act. The messaging is clear: show up for two days, leave with a certificate after a multiple-choice test, and suddenly your organization has AI competency. Does that sound convenient? Yes! Is it…
Most organizations are asking the wrong question about AI. They’re asking: How do we win with AI? What they should be asking is: How do we keep playing? The difference is subtle, but transformative. It’s the difference between approaching AI as a finite game or an infinite game. Simon Sinek popularized this distinction (originally developed by James P. Carse), and it offers a fresh, much-needed…