In 2025, QA teams are experiencing real change. Product delivery cycles are becoming shorter, Agile is maturing, and there’s increasing pressure to launch new software quicker and without flaws.
There’s also evident talk surrounding AI as the biggest factor for change, and its impact is also increasing. But the truth is AI is just one factor in this story. Other trends - like continuous automation, DevOps integration, and the role of QA - are redefining the way we test and ensure quality.
In this article we’ll evaluate: what’s happening, what’s still new and buzzing, and how tools like Xray aid teams to stay on top of this evolution.
AI in QA: the buzz vs. the reality
AI is officially everywhere, and QA is no exception. There’s talk surrounding test cases being automatically generated, smart code analysis, behavior detection, among others. Sounds promising. But what about in practice? Most teams are far from applying AI consistently in their QA tasks.
There’s many reasons for this: the LLMs need a lot of data - sometimes they reach conclusions that aren’t very coherent or show the rationale behind it, so there’s a lack of trust.
But even more fundamentally, AI in QA suffers from a data integrity problem. To work reliably, AI tools need access to structured, consistent, and context-rich data - and most QA environments simply aren’t set up for that. Test results are often unstructured, coverage maps are incomplete, and traceability between user stories, test cases, and production defects is fragmented.
As a result, AI struggles to make meaningful connections or learn useful patterns. A test generator trained on inconsistent test data will generate more noise than value. A bug classifier fed with vague or outdated labels will misidentify root causes. If data integrity isn’t addressed, teams risk turning AI from a smart assistant into an unreliable distraction.
It’s important to keep a balanced approach. When using test management tools like Xray, which are essential for teams to build a strong foundation of planning, traceability, and collaboration - you can have an external AI assistant aiding you on smaller more automated tasks. While AI keeps gaining maturity, the main focus needs to be on keeping QA processes well aligned with the rest of development.
According to a recent study on Top Predictions of AI for 2025 from Gartner, organizations leveraging AI in operational roles like QA must prioritize data integrity and human oversight to avoid unreliable AI outputs. Gartner predicts that by 2026, 20% of organizations will use AI to streamline management layers, emphasizing the importance of human-machine collaboration to maintain workforce morale and operational continuity.
QA is moving: from phase to embedded mindset
The way QA is changing is not just a matter of where we can fit testing cycles. It’s also a matter of mentality.
QA starts earlier with more impact
The concept of shift-left testing isn’t new. Testers are entering the scene in early stages of analysis and planning, they’re more active in refinements, challenging poorly defined requirements while helping clarify criteria before the code is written. This allows quality to be well-thought since the beginning and not only verified toward the end. Besides that, this approach promotes closer collaboration between QA, developers, and product teams, which reduces mistakes and accelerates the delivery cycle.
QA doesn’t end at deployment
On the other hand, there is shift-right: QA continues after releasing. Today, testing isn’t only ensuring something “works” before reaching the user. It includes log analysis, behaviour monitoring, A/B testing, and automatic alerts based on production metrics.
This approach allows teams to learn with the real use of the application and also to quickly adjust. QA supports dynamic adaptation by monitoring user experience data in real time to maintain quality standards.
QA lives inside the cycle, not outside of it
With CI/CD pipelines running, QA can’t just be a bottleneck. Tests need to be continuous, and integrated. This means automation is required, but it’s also important to know when to execute, which tasks to prioritize, and how to analyze results. Quality Assurance is no longer a “service” that is presented to the team, but is now part of the team. You need to be able to read code, understand pipelines, and essentially speak the product’s language and be able to translate all that into a sign of trust for your business.
QA evolution from Tester to quality partner
For a while, testers were only viewed like “bug hunters”. They would receive software, test it, find issues, write a report, and move on. They were someone more technical and didn’t have a close relationship with product decisions or business logic. This doesn’t work anymore. QA is no longer solely responsible for verification, but is now a strategic partner - someone that truly helps their team to make better decisions early on, with an improved focus on overall quality and not only on defects.
Today it's expected that whoever works in QA receives the product, understands the potential users, and sees the impact of each feature. QA is constantly challenging vague requirements, helps to prioritize risks, gives guidance on what should or shouldn’t be automated, and participates in technical discussions about coverage, test architecture, or continuous integration. It’s not only about “testing well” anymore - it’s about ensuring that your team is building well from the start.
Quality powered by people and AI as an ally
The Quality Assurance role in 2025 is clearly evolving and it’s not only due to AI. Of course Artificial Intelligence is gaining space, but it’s not the only transformational factor. The real influence comes from how teams approach quality in development: from the beginning, with a transversal mindset, more automated processes and increased collaboration between technical and business sides.
AI can help generate ideas, maybe analyze some patterns, and automate certain more repetitive tasks, but it doesn’t substitute the critical thinking, product knowledge, and informed decision-making we need.
The main goal is to find a healthy balance where synergy between humans and AI is prioritized, and also stay informed, so that your team can adapt. There’s still more to come for AI in QA and we’re here to keep you posted.
