How AI is changing work
IT
Designing, delegating the build and reviewing critically.
Developers write fewer lines themselves and spend more time reading, testing and steering. AI suggests code, writes tests and increasingly carries out whole tasks. The craft is shifting towards designing, checking and taking responsibility for what runs.
What changes
- Boilerplate code, tests and documentation arrive as suggestions from an assistant.
- Reviewing and testing code becomes a bigger part of the job.
- Agents carry out well-defined tasks, such as fixing a bug or preparing a migration.
- Teams build AI into their own products, often using text analysis and text generation.
Where people make the difference
- Architecture and the choices that keep software maintainable.
- Judging whether generated code is secure and correct.
- Understanding what the user really needs.
How AI is already used
Coding assistants
In an experiment, the group using an AI coding assistant completed a programming task 55.8% faster than the group without one.
arXivCopilot at scale in a bank
ABN AMRO provides an AI coding assistant to its developers. Anyone who wants to use it first takes a training course and signs the terms of use.
ComputableMeasuring the quality of AI code
Software Improvement Group measured the maintainability of software made with AI. With a human in the loop, quality was clearly higher.
AG ConnectReading and writing text in business software
Text analysis and text generation are the most widely used forms of AI at Dutch companies, according to the CBS, the Dutch statistics office.
CBSA Dutch language model of its own
GPT-NL is being developed for public services and later also for businesses. The government is linking projects to TNO to test the model in practice.
Digitale Overheid (Dutch government)
Where the sector is heading
Use is rising, trust is not: 84% of developers use AI or plan to, and 46% do not trust the accuracy of its output (2025).
Stack Overflow Developer SurveyThe information and communication sector uses AI the most: 54% of companies in 2025, against 17% of all companies.
CBSHuman oversight pays off: in a comparison by SIG, software built with a human in the loop scored clearly better on maintainability than software generated entirely by AI (2026).
AG Connect
What a maker can do for you here
- “Build an internal assistant that searches our documentation.”
- “Connect AI to our ticketing system, so requests come in already sorted.”
- “Set up a safe way of working with AI coding assistants in our team.”
- “Build a prototype of an app in a week, so we can test it with users.”
Researched on 5 October 2026. Know a good example from the Netherlands? Let us know.