Welcome to Software Testing Unleashed, the podcast for testers, test automation engineers, test managers and developers who want to build great quality software. My guest today is Florian Fieber. he's Chief Process Officer at Test Solutions in Germany and a very recognized expert in the field of software testing. we are doing years recap Today We talk about twenty-twenty five And give an outlook into twenty. twenty six About topics which were discussed by communities and conferences this year.
And what testers can do to prepare for twenty-twenty six? It's the post-cast years end show and time to reflect a year, make an outlook for next year. There is one expert I know who has good overview of all testing areas. so i invited you! Thank You! And it feels like tradition here at the end of this year together with you. That's true. We did it last year at the German podcast and this here in The International one, maybe what's then up to next year with us? We'll see. Yeah we will see!
So Florian, two thousand twenty-five is nearly over when you look at the software testing area all the software quality stuff that community What are your highlights your remarkable Events and topics which shaped our community this year. Yeah, I think a good observation point is always being on conferences And listening to different perspectives into talks and presentations by people. So from that perspective this years been really interesting I Think topic-wise.
of course the one and only a meta topic was, still is something with AI. AI testing of AI, testing with AI everywhere. so this definitely the topic. it's not the only one. second topic I've seen also already last year is still accessibility testing where you see a lot interesting talks and presentations but you cannot really compare these two from a granularity perspective, right? So accessibility testing is much smaller. It's more relevant than the last years.
and of course AI testing or testing with AI, testing off-AI. we are still learning We're still challenging ourselves so I think at least most of us with the adventof of chat GPT, and we started being very... Yeah. See chances right? So okay this is great! We have these possibilities and so on. And I think this year was between dystopia and utopia. yeah On one side Okay, this will be the biggest transformation and disruption of all times. And we won't need any humans anymore.
We'll all be replaced by the AI on the other side.". This is what I specifically observed in many conferences that there's a new skepticism or realism... healthy realism at least from a software tester perspective to say, okay this is good tool. This was my key takeaway. I would say This year, right? So there was in June the Eurostar conference and they put it one sentence with their topic. It's AI on trial! And that is I think... ...the most important topic this year from my perspective.
When we stay at the AI part What are the real specific and concrete scenarios we can now use AI in for software testing? Because when I say a lot of conference talks about yes, We have to use it. Yes, we have to deal with this. Yes! We need to learn more about AI. but what is your experience really... The true benefit nowadays using AI and which direction? that the true benefit is definitely not to replace any human labor.
The true benefit, it's definitely be much more productive and fast and efficient in what you do. I think its a booster, an accelerator for being faster and more productive with things like this. This requires experienced people who know what they are doing who know which problem they want to solve and having a really, powerful tools at hand. To be fast in what you are doing so I can achieve my goals i can perform activities much shorter time than before or do other things.
that was not able driver. This applies to software testing, maybe in other disciplines or areas that might be different. I don't want to talk about them but at least from our perspective and it's also true that It requires skills and upskilling And knowing how to use these tools having some literacy. But this is probably the booster. Thousands of use cases everybody sees and we imagine.
We can generate things here, we can analyze things there And so on but...we also see that in not everywhere But some companies maybe many companies they're still missing the real strategic goal or advantage Of having AI in place. It's it's not only doing something with AI for of just doing it. So what is the real use case? And we see from surveys and from statistics that, The Real Booster Is Still Sometimes Missing.
This Does Not Apply Of Course To All Companies In All Industries But in general, in average companies are still reluctant and observing whether this really a productivity change or not. for now I don't know what would have been two years or three years, but at least for now. Yeah when we look at the test activities now where do you think in which activity has all the AI stuff most benefit to us as a tester?
In that last world quality report there was an interesting figure saying apparently activities to more generative activities. So it's less analyzing data, finding patterns for example or things like that but rather generating artifacts and so on. That is a good source. maybe we can. we can rely on that. these kinds of things are more valuable or more desirable maybe than other thing.
So, meaning when you think about generating artifacts with typical and need for software testing being its ideas tests cases with test steps automated scripts test data I think this is probably one advantages we can have if you do it right, solving this problem. So all these kinds of things are really helpful. And you mentioned on the other side accessibility, it's much less complex than AI. But where are we now with all the accessibility stuff?
We have now since half a year the regulatory is in charge of us. so but what were the companies nowadays there? First we can say that if they would not be their regulatory part We are only there because the EU Accessibility Act, I would say. If that was not to be there this topic is a nice-to-have topic for everybody of course but many decision makers who spend money and things like that right?
So it's always very important we see as big drivers at least in our discipline being accessibility or AI with the AI act or security and things like that. What we have observed, I've observed is that companies often underestimate The effort you need for good accessibility testing in comprehensive accessibility testing That it's definitely not just using one tool Or one browser add-on accessibility checks on your website for you.
It's much more things to do because only some parts can be tested automatically, but there is still a big need for humans in place who do the manual testing and it's complementary approach. so what I often have seen this year that clients underestimated The efforts we would typically need as a service provider to do good accessibility testing on some website. So, the accessibility test might be bigger than the original design and implementation of some websites maybe.
And in your experience when you did this project I can think of a lot of projects which are already in development since years, and the product is released there in production. And now someone's doing this accessibility test... ...and then another effort to refactor all these stuff…. …to rework those things so that they really work in the sense of accessibility. Is it such an enormous effort? Do you see that a lot re-factoring new architecture has to be done or more or less…? bug fixing style?
Yeah, it depends. Of course I mean if you have never invested in a sustainable architecture of your let's say for example website and there are no means in place that help to provide good accessibility architecture or technologies, or front-end frameworks are something like that.
Coming from the testing perspective and quality of course it's not only technology being used but also having an understanding in mind set to understand that accessibility is not tested at the end and included or introduced, or injected into the product but of course should start as early as possible with a good design built in quality. With build-in accessibility and then of course accessibility testing. it's only some validation validating everything.
so I think this is similar to other non-functional testing, or test types we see being it's maybe security testing. Or usability testing loading performance testing and things like that. so We have two may be perspectives we have here. one might be...we've never made up our minds about accessibility in our product? We start with some accessibility testing during UAT Things to do. we are you might be effective in what?
We do because we find the accessibility issues before we go on production, but of course it's not efficient. In what they're doing. and Do you think there is still? Companies who aren't really not aware of the accessibility stuff where you see okay You come there until there's no awareness for this stuff at all. Yes definitely Being it on purpose or not I don't know. One reason could be that the penalties are not as huge or big, for example with GDPR. Or with an upcoming AI act.
so there aren't such big penalties dependent on your annual revenue. you have. Let's see what happens. I think still many companies or decision makers who don't take that topic maybe serious enough, so definitely i would say it is a non-functional criteria and has the same challenges as all other non functional criteria we know with the products and might be considered as a nice to have topic, which is definitely not.
close the year of twenty-twenty five and make an outlook to twenty, twenty six. What do you think? When when you ask your glass or looking in your teapot something like that what are topics especially for next year? we have to deal with some additional topic which is on the horizon. not all aware Yeah, so of course nowadays we typically don't look anymore in our MUX. We typically probably use GPTSA as an oracle.
but Of course I would say looking ahead into next year Still AI and testing will be the topic. this will also be the top again two years i guess. So from that side it might become a bit boring because we have only monotopic discussions, but of course it's big enough to have enough sub-topics or contents in. Of course the Agentec AI approach technology is one thing that will see emerging meaning there transformation of the role of software testers.
It will continue to be a maybe only manual activity, doing manual tasks but rather more automation using AI technologies and using AI agents that help me fulfill some task. so... And my role may shift like orchestration having one or more agents helping me here. I'm the pilot, i use a co-pilot and not the copilot in using a pilot. this is very important thing right? And another thing that would say for example when we look at what EU AI Act demands from companies or their employees.
so if you look at Article IV it's called AILiteracy AI literacy, meaning enough skills building up experience having the right tools and technologies in place knowing how to handle them. And two work with them from all their staff of other employees. so I think this will be definitely a big task for companies next year not only for the experts but everyone being able being competent in using AI technology for their work. It's a commodity, right? So it will not change.
Am I using AI and my work or is that there?". And we definitely have to know how to handle the risks... ...know the challenges, know the borders of limitations,... ...and things like this.". So, there's a whole part about using AI in my work context as a tester or company. But... In relation to testing of an AI system do you see what are the critical paths for next year? Or is it a niche problem and topic to test on AI? It depends! The consultant answer.
So, the question is with what we compare it. If you have two pillars or parts testing off AI and testing with AI... We currently have a bigger impact on more things happening in the whole field of the AIS as testware supporting us as testers doing our work. This impact we currently see where it's mainly used. Of course, testing off AI is important but compared to the first part its still a niche.
But of course for those using AI in their products or AI components In that product It not a niche right? So Its importance too To handle it properly. The test object system and of course this is, I would not say completely different but it's different from traditional approaches. We still have our methods and tools and technologies in place. we know as testers right?
how to test some systems, but of course here we have another Test object that behaves differently and We have To make sure how to come up with quality assessments How to validate these? These the system's yeah And the more we use this technologies. Of course The more it will have risk there.
This is why they have AI act Giving us boundaries maybe or recommendations how to work with systems that are chat assistants, that I use in health insurance for example or they're used in self-driving cars and things like these. These are systems not only important but also bear high risks maybe in the usage of them. so testing Quality assurance is key in these parts of the industry.
Yeah, yeah It's now when... When the testers began to build up their learning plan for next year and getting their budget for learning from the companies. What do you recommend? To a tester today to gain up their skills next year? are there any specific seminars, conferences or meetups or certifications that you recommend doing next year to grow in the area of software testing? I would first say there's not one channel type of upscaling.
All what you just said is important and Like some like more conferences. Some like more YouTube videos, some like More go to trainings and things like that. so at the end I would say it's a complementary approach. of course i Would encourage everybody for example going to conferences. this is always great exchange with people from The industry With thought leaders. you You meet there? You get inspirations And things like That. one part.
other part Is what always recommend if you ask me, for example training and certification in that field. For example we have the ICTQB certified tester portfolio. now I mean not now. it's been there three years one syllabus for AI testing. so part of a system which is therefore three months, half a year or three months about is testing with GenAI.
And this something I would definitely recommend to everybody because it's definitely helps in that field of AI literacy supporting and maintaining your qualification and skills using AI technologies for very concrete disciplines. so these two could be too interesting parts, pillars in your qualification plan next year. And do you have any plan for yourself? You want to share with us for your next year in test software testing? We'll see it!
To be honest I haven't created yet my plans so i will definitely upscale myself also in this wild field of AI literacy. I would personally like to go deeper into data analytics, machine learning and more hands-on using the technology training models. so that's something i'd love. What will be there? Yeah. Sorry, what about you? yeah for me too I also want to do more with the AI stuff especially using it in the projects and enabling my client teams to use it more in their daily life.
And that was one of my question before, what is the concrete way we can or a specific way? We can do it now and there. I want to give you an appropriate answer next year for clients into your team. There are also some gaps where i have gained skills in this area. The reason why think its really important might be so If you take the example with accessibility testing we had before, so what I meant when i said okay it's not that complex.
You only realize that when for example do several trainings or attend different presentations of talks then you see ok now listen to let say five or ten speakers talking about some topic and identify a pattern in Say and apparently okay, this is the common ground. The common sense in one area that's important to learn. so it not only... That what I also meant with a complementary thing. i have different sources.
Let say YouTube videos are learned from on-site trainings or webinars meetups whatever might be some repetition of things. And this is also very important to understand and identify the importance of a topic, really being able to assess it. Yeah that's so true! Yes Florian thank you very much for your insights from twenty-twenty five and twenty-two six. I hope we will see soon again in May next year. Yeah, I'm looking forward to that and thank you so much for being here on the show.
Happy New Year because Christmas is already over when the podcast episode is released! Okay thanks also from my side. it's been a pleasure as always but still Merry Christmas or happy holidays over to you and good new year. Thank You bye-bye.