Testing approaches are drifting towards the category of generative test cases by AI, automated repetitive work. The bigger concern now being, if tools could take away some aspect of our work, where should a tester channel his/her work and learning time?
I would love to hear your thoughts on what you must have the skills to be a good tester in the future.
My two cents:
On the bright side of life: AI explosion is coming to an end. More and more companies are trying to find out the ROI of it, and it does not look good at all. AI will be another tool for limited purposes.
The skills required to be a good tester have not changed. The only new skill you need to acquire is the ability to persuade your manager that it would not be useful to incorporate AI into your testing. Right now, it’s an unnecessary distraction. If it ever becomes useful, you will be able to learn what you need very rapidly. I haven’t encountered any problem for which AI would be the solution. Has anyone?
We have been experimenting with using AI to create simple testing tools for our internal use, and it may be worth putting some effort into that. It’s certainly fast, but I hate the whole iterative process of working out what it’s just built, why it still doesn’t meet my needs and how I need to change the prompts to get it to build exactly what I want.
I hate that I don’t learn anything from the process. I hate that it needs so much more testing than if I wrote the code myself. And I’m scared to look at the hideous code it’s generated, but fortunately it will never go into production.
There are likely multiple areas to learn but the one area I think will become fairly standard very quickly is having testers with access to the code and doing local builds.
A lot of testers have been doing this for years and AI adds quite a bit of testing opportunities that leverage from direct code access, whether its API testing, automation or deeper investigative white box analysis I would expect it to shift further in this direction and become more of a standard for testers across the board.
I’ve found a lot of value in doing local builds and using the developer tools available for testing in general.
When it comes to automation there are quite a few experiments going on that point to AI accelerators, some suggesting things like four times multipliers. The interesting thing for me on this is there have been times UI automation was rejected as it tended to have a poor ROI, a potential productivity multiplier may require those decisions to be revisited and it might just be worth that half day a week focus on now.
AI allows us to produce a lot of cheap code / cheap tests. Of course, we need to learn how to use AI effectively.
But we, as test engineers, will be valuable in the AI era for other characteristics: the ability to validate and judge the AI tool output. Ability see the big picture and how the pieces fit together. The knowledge of what to test and what NOT to test. Critical thinking, deep domain knowledge.
As AI handles more test automation and repetitive tasks the role of testers is shifting. Which skills will keep us relevant and future ready? I’d love to know what you think are the essential ones every tester should build.
I presented something related to this at a testing conference recently:
My view is that the core competence of a tester is solving complex testing problems, requiring “deep work”.
If we can use AI for all the “shallow work” or administrative tasks we do - great. That leaves us more time to focus on our core competence.
Just like with test automation, AI is something we should learn about, to understand how it could help us become more efficient. How AI can help someone - if it is generating basic test cases, writing up test reports, summarising quality data, or being used in automated tests - is very dependent on context I think. But definitely something we should all be looking into.
Obviously this is only my view and not a universal truth.