Operations teams often need a small internal application: an approval process, an asset register, an inspection form, a replacement for a shared spreadsheet. Microsoft Power Platform and a custom ASP.NET Core application can both deliver it. The right choice depends less on the technology than on who uses the application, where the data lives and who will maintain it.

What each option is good at

Power Platform (Power Apps, Power Automate and Dataverse) is strongest for internal users who already work in Microsoft 365. Sign-in, hosting, mobile access and a large catalogue of connectors come with the platform. Business users can make small changes themselves, and an application can be in front of users quickly.

A custom ASP.NET Core application gives full control of the user interface, data model, performance, hosting and cost structure. It suits external users, complex business rules, high volumes and anything that will become a core product. It needs developers for every change and an operations owner for hosting and patching.

The checklist

Go through these questions with the people who will own the application. No single answer decides it; look at where most of them point.

QuestionLeans Power PlatformLeans custom .NET
Who are the users?Staff with Microsoft 365 / Entra ID accountsCustomers, partners or the public; anonymous access
How many users, and will that grow?A known, stable groupLarge or unpredictable numbers, where per-user licensing would not fit
Where does the data live?Dataverse, SharePoint or systems with existing connectorsYour own databases, high-volume or complex relational data
How complex are the rules?Forms, approvals, notifications, simple validationMulti-step transactions, calculations, concurrency, strict consistency
What does the interface need?Standard forms, galleries and viewsA specific design, heavy interactivity, accessibility requirements beyond the platform's controls
Performance and volumeModerate data sizes and request ratesLarge datasets, high request rates, tight latency targets
Who maintains it?A business team with platform skills, supported by ITA development team with a release process
How long will it live?Departmental tool, likely to change oftenLong-lived system that other systems depend on
Vendor dependencyAlready committed to Microsoft 365 and Power PlatformNeed to keep hosting or cloud options open

Two technical limits catch teams out with canvas apps in particular. Delegation determines whether a filter or sort runs at the data source; when it cannot be delegated, the app only works on a limited number of rows. And Power Platform applies request limits and allocations that matter for busy flows and integrations. Check both against your expected data volumes before you commit.

Licensing: check it early, and check it yourself

Licensing often decides the question, and it changes. We do not quote prices here because they vary by agreement and change over time. The points to check against Microsoft's current Power Platform licensing documentation and your licensing partner are:

  • Standard versus premium. Some capabilities are included with Microsoft 365 licences; Dataverse, premium connectors, custom connectors and on-premises data gateways generally require premium licensing.
  • Licensing model. Per-user, per-app and pay-as-you-go (billed through an Azure subscription) options suit different usage patterns. Model your actual user numbers against each.
  • External users. Applications for customers or partners are usually built with Power Pages, which has its own licensing.
  • Capacity. Dataverse database, file and log storage is allocated by licence; additional capacity costs extra.
  • Automation. Unattended desktop automation and high-volume flows have their own licensing considerations.

A custom application has a different cost profile: development effort up front, then hosting, monitoring and maintenance, but no per-user charge. Compare total cost over the expected life of the application, not the first year.

ALM: treat low-code like code

Many Power Platform problems have nothing to do with the platform's capability; they come from building directly in production. The same discipline as a .NET project applies (see Microsoft's ALM guidance):

  • Separate environments for development, test and production.
  • Solutions to package apps, flows, tables and components. Deploy managed solutions to test and production.
  • Environment variables and connection references instead of hard-coded URLs and connections, so the same solution deploys to every environment.
  • Automated deployment with pipelines in Power Platform, or Azure DevOps or GitHub Actions using the Power Platform CLI.
  • Source control of unpacked solution files, so changes can be reviewed and history kept.
  • Governance: data loss prevention policies to control which connectors can be combined, and Managed Environments for larger estates.

When low-code runs out: extending with code

Choosing Power Platform does not mean giving up custom code. The platform has well-defined extension points:

  • Custom connectors wrap your own HTTP API, defined by an OpenAPI description, so apps and flows can call it like any other connector. An ASP.NET Core API that generates OpenAPI is a natural fit.
  • Power Apps component framework (PCF) code components, written in TypeScript, for controls the platform does not provide, such as specialised grids, charts or input controls.
  • Dataverse plug-ins, written in C#, run business logic on the server when records change, inside the platform's sandbox and execution limits.
  • Azure Functions for heavier processing, scheduled work or integration logic, called from flows or custom connectors. New .NET functions should use the isolated worker model; support for the in-process model ended on 10 November 2026.

Hybrid patterns

Many good solutions use both. Common shapes:

  1. Power Apps front end, .NET back end. Staff use a Power App; complex rules and data live behind an ASP.NET Core API exposed through a custom connector. Business users can adjust screens while the rules stay under version control and test.
  2. Custom application on Dataverse. A customer-facing ASP.NET Core application reads and writes Dataverse through its Web API, while internal staff work with the same data in model-driven apps.
  3. Power Automate around a custom system. A custom application raises events or calls an HTTP-triggered flow for approvals, notifications and Microsoft 365 tasks, so those steps can change without a release of the main application.
  4. Start low-code, graduate deliberately. Build a first version in Power Apps to settle the process, with data in Dataverse or SQL that a later custom application can use. Agree up front which signals (user numbers, data volume, rule complexity) would trigger a rebuild.

Warning signs

You may have chosen the wrong side if you see:

  • Power Fx formulas or flows that nobody wants to change because nobody understands them.
  • Workarounds for delegation limits or request limits appearing in every new feature.
  • A custom application whose backlog is mostly form changes that business users could make themselves.
  • Licence costs growing faster than the number of people who get value from the application.

None of these is a reason to rewrite immediately, but each is worth discussing before the next phase of work.

Not sure which way to go?

We build on Power Platform and in .NET, and extend one with the other where that makes sense. A short technical call is usually enough to see which direction fits.

Power Platform services