GenAI in Testing - First Entry in a Series

AI and especially GenAI has been the hype since the end of 2022 when OpenAI released ChatGPT publicly.

At first, people were having fun, learning new stuff, or using it for generating silly content, and explore uncharted territories, kinda testing ChatGPT from functionality, ethical or security point-of-view.

But then, they started realizing that it could be useful at work. You could generate e-mail responses to your colleagues and clients. Especially those who hated the corporate wording style, or lacked proper English (or whatever language was used at their job), started benefiting from using GenAI. You can generate test cases, snippets of code, help filling in your yearly tax report, write a contract without the need to pay for a lawyer. You see where this was and is going.

Obviously, this generated skepticism and even fear in most people, myself included.

I personally thought that it's still just mostly hype, why would I need this to write any letters or documents for me? Or how can this create better test automation code than I could? I wasn't afraid of AI (at first), and I spent most of 2023 not diving into it, just exploring it. But then at the end of the year my management at EPAM asked me to start investing more time into learning about LLMs, and how GenAI can be best used in Testing. And that's how I became one of the first QA personnel at the company who was capable of developing reusable prompts and AI Agents to provide efficiency gains in multiple aspects of the SDLC.

Today, I am one of the main consultants of GenAI in Testing in the Americas for EPAM. I am mentoring other consultants, engineers, analysts and managers on the usage of GenAI tools and platforms, AI Agents, and basics of prompt engineering.

My intent with this blog series is to show some insights into the work I do, and how you, my reader, could leverage it at work. I won't sell you the bs that GenAI will replace humans, because I don't agree with that statement.

Moreover, in my opinion, most of the GenAI tools and platforms out there are just embedding ChatGPT or some other LLM, only providing little additional features and benefits. I don't think it's fair from any tool vendor to sell the idea that their fancy GenAI tool is so outstanding until they really provide some groundbreaking real project results.

So I don't currently plan to pitch any such tools, but of course if necessary for demonstration purposes, I will obviously use some GenAI platforms, but will try to teach concepts that are tool agnostic, or can be accessed or achieved in many alternatives.

I also won't sell you the idea that by leveraging GenAI you automatically become let's say 50% or 80% more productive. It doesn't happen that way, and definitely not overnight, and not at that extent. At the beginning, you will have productivity losses as you need to learn and your people need to invest into learning as well, setting some of the regular work items aside for a while.

But more to that in one of the next blog entries. :-)

What I will share in this series:

• Personal experience using GenAI in software development

• Easy-to-follow instructions and guidelines that you can apply at your job

• No-bs, honest style of where I see GenAI trending in the near future and long-term

The only AI-generated content of this blog post is the cover image, all text was written by a human, by me.