Let’s talk about something I’ve been thinking a lot about lately, and if you’ve ever built anything in Power Apps (or a similar low-code tool) or dipped your toes into the world of AI tools like ChatGPT, you might be wondering the same thing:
Is AI about to replace low-code platforms like Microsoft Power Apps?
Because let’s face it, things are moving fast. A few years ago, learning Power Apps was the smart way to build internal tools without relying on developers. Now, you can prompt an AI model with plain English and get a working app prototype in seconds.
Sounds like game over for low-code, right?
Maybe, not quite.
Let me explain why.
Why Power Apps Still Matters
Power Apps was designed for enterprise use cases. It’s not just about making things quickly, it’s about building secure, compliant, internal tools that run on top of your Microsoft stack that includes: SharePoint, Teams, Dataverse, Excel, Power Automate to name a few, all with admin-level control, DLP policies, and identity governance baked in.
This is a platform made for organizations that care about:
- Access controls
- Data protection
- Audit trails
- App lifecycle management
It’s drag, drop, connect, and critically secure. That’s a huge deal when you’re building tools that handle real business data.
AI Is Evolving Fast (and That’s Exciting)
Now let’s talk about the other side of the coin.
AI tools like ChatGPT, Claude, and Gemini are capable of generating:
- Code
- Database schemas
- UI layouts
- App logic
- and much more
All from a single prompt. You can say:
“Build me an app to track customer issues, assign owners, and send an email when a case is resolved”
… and you’ll get something eerily close to a working app.
It’s fast, it’s fun—and yes, it’s impressive.
But here’s the critical thing to understand…
Speed Doesn’t Equal Security
Just because an AI can build an app quickly doesn’t mean it’s safe to use inside your business.
When you use a public AI tool to generate an app, you’re potentially exposing sensitive data, inviting vulnerabilities, and skipping all the governance steps that matter in enterprise IT.
- No identity access controls
- No endpoint security
- No audit trail
- No guarantees on data residency or compliance
- Potential code hallucinations or misconfigurations that could get you hacked
So while AI-generated apps might look shiny on the surface, they’re not ready to replace enterprise-grade platforms like Power Apps. Not without layers of review, security checks, and a dedicated dev/sec team behind the scenes.
Where It Gets Interesting: AI Inside Power Apps
What’s actually happening now is more nuanced.
Microsoft has started embedding AI into the Power Platform through Copilot. You can now describe the app you want in natural language, and Power Apps will scaffold it for you—within your tenant, using your data sources, and under your governance policies.
It’s not about AI replacing low-code.
It’s about AI enhancing low-code, making it faster, more intuitive, and easier for anyone to build.
That’s where the future’s heading:
🧠 You prompt.
🔐 Power Apps builds—safely.
⚙️ You refine and publish—with control.
Skills Are Shifting, Not Disappearing
Let’s be clear, AI isn’t removing the need for skill, it’s just shifting them.
You don’t need to know every formula in Power Apps anymore, but you do need to:
- Learn how to write effective prompts
- Understand your data structures
- Spot security gaps in AI-generated outputs
- Govern what gets deployed (and by whom)
The best builders moving forward will be fluent in both AI and low-code platforms. It’s not either/or. It’s both.
My Take
Can AI replace Microsoft Power Apps?
Not yet, but maybe sometime in the future.
At least not yet for businesses that care about security, compliance, and building tools that can scale across their organization.
💬 What’s your take? Have you tried building with AI yet? Are you still using Power Apps daily? I’d love to hear how your workflows are evolving.
👉 If you’re curious about this AI journey, stick around as I’m documenting the whole thing.
This is Barry, aka Hugo on AI.
More soon.