Low-Code

Normally, a program is created by someone typing instructions for the computer as text. This text is called code, and writing it takes a long time and has to be learned. Low-code turns this around: you build a program mainly with the mouse, from prefabricated building blocks on a work surface on the screen. A building block might be, for example, an input field, a table, or the rule “when someone submits the form, send an email.” You only have to type a little yourself, hence the name: low code. If it works entirely without typing, it’s called no-code.

Why companies rely on it

In almost every company there are more ideas for useful small programs than people who can build them. A vacation management tool, a form for travel expenses, an overview of open repairs: such tools are technically simple, but they end up on the IT department’s waiting list for years. Low-code is meant to clear this backlog.

The second reason is money. Experienced developers are expensive and hard to find. If a clerk from accounting clicks together her own form in two days, that saves weeks of development time. There is even a term for such people: citizen developer, meaning employees who program without being programmers.

One misconception persists stubbornly: that low-code makes programmers superfluous. That’s not true. Large systems with millions of users, strict security requirements, or unusual logic are still built the classic way. Low-code covers the broad mass of small applications, not the top end.

Building blocks, rules, and the limits of clicking

A low-code platform usually consists of three parts. First, an editor for the interface, in which you arrange fields and buttons via drag-and-drop, meaning you drag them to the right spot with the mouse button held down. Second, an area for the data, where you define which information gets stored. Third, a toolkit for workflows, in which you string together if-then rules.

In the background, this still produces real code. The platform automatically translates the assembled building blocks into instructions that a server can execute. You can think of it like off-the-shelf furniture: the individual parts are prefabricated and only fit together in certain ways. This is fast, but a cabinet in a crooked sloped ceiling still needs a carpenter.

This is exactly where the small remainder of code comes into play. For special cases, you can insert your own program text in many places, for example an unusual calculation. Newer platforms supplement this with AI: you describe in plain English what the application should do, and the system generates a first draft. This is called vibe coding, though it is only a starting point and must be checked.

From Microsoft Power Apps to shadow IT

The best-known providers are Microsoft Power Apps, Salesforce, ServiceNow, Mendix, OutSystems, Airtable, and Bubble. Low-code regularly appears in stock market news because this market has been growing by double digits for years and its providers rank among the most expensive software companies of all. Website builders like Wix basically follow the same principle.

In everyday life, you encounter this idea earlier than you’d think. Anyone who writes a formula in a spreadsheet or sets up an automatic rule with a service like IFTTT is essentially already working this way. Many chatbots on company websites are also assembled by clicking rather than fully programmed.

What is discussed most critically is so-called shadow IT. When departments build applications on their own authority, the IT department eventually loses track of where which data is stored. If the creator leaves the company, no one maintains their tool anymore. That’s why larger companies set clear rules about which data may even be processed in low-code applications.