Update (May 2026): The original version of this article went out in November 2021, when “low-code” was still something we had to explain to most clients. Five years later, that explanation is unnecessary — low-code is in 85% of large enterprises and the Gartner Magic Quadrant for it now has six Leaders. What’s worth talking about instead is which of the original advantages still hold up, which got more important, and what AI has changed. That’s what this version covers.
Low-code is software development where most of what you’d traditionally write as code is generated by the platform from visual, declarative configuration. You still get the resulting application, the underlying code, and the ability to extend it with custom logic — what changes is how much of that you write by hand. This makes some projects substantially faster, some projects much cheaper, and some projects newly possible for teams that don’t have ten senior engineers sitting around.
The market has consolidated since we last updated this article. According to the 2025 Gartner Magic Quadrant for Enterprise Low-Code Application Platforms (published July 28, 2025), six platforms now sit in the Leaders quadrant: Mendix, OutSystems, Microsoft Power Apps, ServiceNow, Appian, and Salesforce. Oracle APEX continues to occupy an unusual position — it’s a Challenger by Gartner’s metrics but the de facto default in any organization already running on Oracle Database, which is a lot of organizations. SAP Build sits in the Visionary quadrant. The full list is in our breakdown of the 2025 Gartner LCAP report.
If you’re new to the space and want to go deeper on specific platforms, we’ve written tutorials for several of them:
There are many different advantages of using low-code in your company. Here’s what I consider to be the most important ones.
Software development is expensive. That hasn’t changed since 2021 — what’s changed is by how much. Average enterprise developer salaries in Western Europe and the US have continued climbing, and even with AI assistance taking some pressure off, the senior engineering hour remains one of the more expensive things a non-tech company pays for.
Low-code reduces that cost in two ways: less time spent per application, and a lower hourly rate on average (because some of the work can be done by less senior people, including business technologists). The 2026 numbers are striking. Forrester’s commissioned Total Economic Impact studies show ROI figures like 253% for OutSystems implementations (Ricoh case, full payback in 7 months) and 206% for Microsoft Power Apps. Vendor numbers, yes, but consistent across multiple independent studies and platforms.
What we see in our own projects matches the pattern: depending on application complexity, low-code typically delivers somewhere between 40% and 70% faster time-to-production compared to equivalent traditional development. The cost savings track that proportionally. The MRHT case is a good example — we built a full SaaS analytics platform in 4 months on Oracle APEX, and the client has been operating on it for over a year now. A traditional custom build would have been 18-24 months minimum.
A note of caution worth repeating: low-code is usually cheaper, but not automatically cheaper. Licensing models for some platforms (Microsoft Power Platform is the most cited example) get expensive quickly at scale. The math always needs to be checked against your specific case before signing anything.
Another very important thing is that low-code focuses on solving business problems, instead of problems with technology – or problems with the IT industry itself. For example, everyone who has tried to hire professional software developers in the last decade knows that the competition in the market has gotten really fierce. Companies do what they can to attract talent. Salaries constantly go up, accompanied by various additional – and expensive – benefits.
This means that in some cases creating and maintaining an internal IT department is kinda “a job inside the job” – a huge undertaking in and of itself. People spend so much time, effort, and money on accommodating software developers, that they lose focus of their business and goals.
Low-code addresses this problem by introducing so-called “citizen developers” (people from outside of IT space) into the process of software development, and therefore lessening the role of IT specialists. The additional benefit here is that low code makes it much easier to take into account the opinions and ideas of non-IT employees – like people these apps are made for (business users, for example). It’s also easier to train them in the software’s use. Gartner forecasts that by the end of 2026, 80% of low-code tool users will be outside formal IT departments, up from 60% in 2021. We’ve seen this play out very directly — in the past three years, the most active “citizen developers” in our client organizations have been finance analysts, operations leads, and HR specialists who got tired of waiting six months for IT to build them a simple form. Mostly they build small departmental apps, which is exactly what they should be building. The role of IT shifts from “build every app” to “set the guardrails and review the few that touch production data” — which, in practice, is a happier job than chasing departmental requests all day.
The talent gap argument needed updating. In 2021 the story was simple: not enough developers, salaries climbing, and forecasts about hundreds of millions of new applications no workforce could possibly deliver. Five years later, the gap is still there, but the picture is more nuanced.
On one hand, AI coding assistants (Claude Code, Cursor, GitHub Copilot, the Oracle SQL Developer for VS Code AI features) have meaningfully raised the throughput of individual senior developers — by some internal estimates, 30-50% on routine work. Some of the pressure on the talent market has come off.
On the other hand, the United States alone is projected to face a 1.2 million developer shortfall by the end of 2026. 82% of organizations report struggling to hire engineers (Gartner, 2025). The applications that need to get built keep multiplying — every regulatory deadline, every legacy modernization, every new AI-powered process. Even with AI assistants in the picture, demand outstrips supply for the foreseeable future.
Low-code addresses this from the supply side. A developer who knows Oracle DB, SQL and PL/SQL can be productive in Oracle APEX in two or three weeks. A business analyst can be productive in Mendix or Power Apps in two or three months. You don’t need to compete for the rarest, most expensive engineering talent for projects that don’t actually require it. That logic was true in 2021. It’s more true now.
High costs are understandable if they are accompanied by the quality that’s equally high, but sadly, that’s not always the case. Once upon a time, IT has been identified with progress and acceleration of changes. Investment in this area was seen as the best way to speed your business up. Currently, that’s not entirely true. Complex software projects take a lot of time to finish, and even when they’re technically “done”, it often turns out more work is still required. Time to market counted in months or even years, is far from optimal in most circumstances.
Low-code application development can offer several unique advantages here. Firstly, it allows you to create software a lot faster. I’ve already mentioned that in the section about IT cost, but it’s a factor that affects a company in several ways – and therefore it’s worth repeating. Often, the first effects can be seen after a few days or even hours of work. That basically means you can answer challenges as they arise, instead of tackling them in the more or less distant future.
Secondly, low-code standardizes the process of making software. Companies that designed low-code development platforms (such as Oracle or OutSystems) chose well-tested components and defined how the applications are developed and deployed. Developers can change these things if they really need to, but the key idea here is that they have a baseline, a set of “good practices” that’s usually worth following (there are also tools that can help you make sure that they are kept, such as APEX Advisor or APEX Standard Tracker). With APEX 26.1 from May 2026, Oracle went a step further — applications can now be exported in APEXlang, a human-readable specification format that’s reviewable in Git, diff-able in pull requests, and inspectable by AI coding agents. That’s a meaningful upgrade to governance and a clear example of where the major platforms are heading: code that’s still generated, but no longer opaque. Also, the apps created on these low-code platforms are available through them, which means that some things – integrations, user management, authentication, and so on – will be the same in the case of every piece of software. They’re defined at the level of the platform itself, instead of each individual application.
This standardization results in maintenance being easier and less problematic. I’m sure you can already see why: many things are the same in every application, and they use components that have already been extensively tested by someone else (i.e., the company responsible for the low-code development platform) on various systems, browsers, and so on. That means much less work is required.
All of the above also makes it a lot harder to blur the responsibility for mistakes – if you follow the standards, you see who doesn’t, and the platforms log most things developers do very scrupulously. That makes certain quality bars easier to enforce. To put it simply, you’ll not only spend less money, but it’ll be easier to make sure it was actually spent well.
Most low-code platforms are cloud-native or cloud-ready, which in 2026 isn’t an “advantage” so much as a baseline expectation. What’s worth noting instead is how comfortably they sit with sovereign cloud requirements. Oracle APEX runs on the Oracle EU Sovereign Cloud at no premium over the standard public cloud price. Mendix and OutSystems both offer EU-hosted deployments. For organisations facing DORA, NIS2, the EU AI Act, or local equivalents like Germany’s BSI C5, having a low-code platform that already speaks “EU-resident data, EU-resident operations” out of the box is more valuable than it was even two years ago.
This advantage didn’t exist in 2021. In 2026, it might be the most important reason on this list.
Generative AI is genuinely good at producing application code. It’s much less good at producing application code that works inside an enterprise — applications that respect role-based access control, log audit trails, integrate cleanly with whatever ten other systems your company runs, and survive a security review. Low-code platforms are designed for exactly that environment. When you plug AI into them, you get the speed of generation with the governance of the platform.
Every serious LCAP has moved in this direction. Oracle APEX 26.1 (May 2026) shipped APEXlang and AI Agents with Tools — AI doesn’t just chat inside your app, it can call governed functions to actually do things. Mendix has built natural-language application generation into its IDE. Microsoft Copilot Studio for Power Platform lets you spin up departmental AI agents in hours. OutSystems and Salesforce have parallel offerings.
The pattern across all of them is the same: AI generates intent, the platform translates intent into governed, inspectable artifacts. You get faster builds without the “I asked Claude to write our approval workflow and now I can’t audit who modified it” problem. For regulated industries — banking, healthcare, government — this is the difference between AI being interesting and AI being usable.
For organisations that haven’t picked a low-code platform yet, this might be the dimension that matters most when choosing one. AI capabilities in 2026 vary substantially between platforms. The differences in 2027 will be larger.
In November 2021, I closed this article by calling low-code “the future.” Five years on, that framing doesn’t quite work anymore. Low-code in 2026 isn’t where things are heading — it’s where they are. Gartner projects that 75% of new applications built in enterprises this year will be developed on low-code platforms, up from less than 25% in 2020. The market itself crossed $30 billion in 2026 and is on track to multiply several times over by the end of the decade.
What’s still genuinely “future” is what happens when low-code and AI converge — when natural-language prompts, governed metadata languages like APEXlang, and AI agents with tool access become the default authoring experience inside enterprise platforms. That convergence is happening now, and the platforms that get it right will define what enterprise software development looks like in 2027 and 2028.
If you’re considering low-code for the first time, or trying to figure out which platform makes sense for your stack, this is a useful moment to start the conversation — early enough to make a real choice, late enough that the technology is genuinely mature. Drop us a line at hello@pretius.com or use the form below. We’ve been doing this since 2008. The initial consultation is free, and we’ll be straight with you about whether low-code is the right call for your specific situation.
