At MoodleMoot Spain 2026 in Valladolid, we had many interesting conversations. One of them was John’s talk, in which he addressed a topic that is becoming increasingly important: how to apply artificial intelligence to QA.
Because we all know that QA is important. But we don’t always stop to think about whether we’re doing it the best way possible.
Why QA is important
QA isn’t just another phase of the project. It’s what makes the difference between a platform that works… and one that causes problems.
The session was summarized very well in three key points:
- User experience guarantee
- Safety net
- Prevention
Ultimately, it’s about ensuring that everything we build works the way we expect it to and, above all, the way the user expects it to.
How do we do QA these days?
This is where we often encounter a fairly common situation. The standard QA process follows a fairly clear logic:
- Active navigation: A user opens the browser and navigates through the platform step by step.
- Rigid instructions: When we automate, we write code that looks for exact matches.
- Validation: A human inspector checks that everything is working as it should.
- Loop: The process is repeated for each browser, device, and version, as many times as necessary.
And this is where the problems begin. The problem isn’t that it doesn’t work. It’s that it becomes unsustainable as the project grows.
The shift in focus: incorporating artificial intelligence
The question raised during the session was: What happens if we incorporate artificial intelligence into the QA process?
It’s not about replacing everything that already exists, but about changing the way we work. It’s about moving from running tests manually to defining the workflow and letting AI run it as many times as we want.
AI enables:
- Tests that adapt automatically because we don’t need to specify their exact identifier (self-healing)
- Natural language interaction makes it easier to write commands
- Autonomy in implementation
- Integration with CI/CD
- Improved accessibility
There are also limits to keep in mind
That said, it’s not quite that simple. There are several factors to keep in mind:
- Model usage costs (tokens) and latency
- Early adopter technology
- Non-determinism
- Page loading
This is where it’s easy to make a mistake: thinking that AI replaces all QA. But it doesn’t. What it does is complement QA, but you have to know how and when to use it.
Tools that are already leading the way
During the session, several tools that are already being used for this purpose were also mentioned:
- Stagehand
- ZeroStep
- TestRigor
- Datadog
- Applitools
Some are more advanced than others, but they all point to the same goal: smarter, less manual QA.
Are you interested in exploring this approach?
At 3ipunt, we’re already applying this kind of approach to our projects, always with one clear goal in mind: that technology should be meaningful and add value.
Incorporating AI into QA isn’t about changing everything overnight, nor is it about following a trend. Rather, it’s about starting to question how we work and seeing where it makes sense to evolve.
