Low-Code Platforms in Enterprise — Salvation or Trap?

“A citizen developer clicks together an app in an afternoon.” That’s what low-code platform marketing says. After a year of experience with Power Apps, OutSystems, and Mendix in an enterprise environment, we have a more realistic view. Low-code works — but not where marketing promises.

What We Tried¶

Power Apps: Microsoft ecosystem, integration with M365 and Dataverse. Our clients (mostly on the Microsoft stack) want it automatically.

OutSystems: A full-fledged low-code platform. Generates real code (.NET/Java), enterprise features (CI/CD, monitoring, scaling).

Mendix: Similar to OutSystems, strong in business logic modeling. SAP integration.

Where Low-Code Works ✅¶

• Internal forms: requests, approval workflows, records — an ideal use case
• Rapid prototyping: show the client a concept in 2 days instead of 2 months
• Process automation: Power Automate for M365 service integration
• Simple CRUD apps: asset tracking, contacts, simple CRM

Where Low-Code Doesn’t Work ❌¶

• Complex business logic: once you need more than CRUD, you hit the limits
• High performance: thousands of concurrent users, real-time processing
• Custom UI/UX: pixel-perfect design or non-standard interactions
• Offline-first: mobile app with offline synchronization

Real Costs¶

“Low-code is cheaper” — not always. Power Apps Per App Plan: ~$8/user/month. OutSystems license: tens of thousands annually. Plus: specialized consultants (scarce and expensive), migration when the platform ages, vendor lock-in.

For an internal app with 50 users and simple CRUD: Power Apps is cheaper than custom development. For a core business system with 5,000 users: custom development is cheaper and more flexible.

Shadow IT Risk¶

The biggest risk of low-code: a “citizen developer” builds a critical application without IT governance, without backups, without a security review. Then they leave the company. Solution: a governance framework — who can build, where it gets deployed, who approves, how it’s backed up.

Our Recommendation¶

Low-code belongs in the toolbox, not as a strategy. Excellent for simple internal tools and prototypes. For core systems and complex applications, stick with custom development. And always have an exit strategy — what if the vendor triples the license cost?

Low-Code Is a Compromise¶

Speed of development vs. flexibility. Simplicity vs. vendor lock-in. Citizen development vs. governance. No answer is universal. Understand the trade-offs and choose by use case, not by marketing.