If you only follow the loudest headlines, it is easy to believe AI is about to wipe out manual testing.
Katalon's State of Software Quality Report 2025 tells a more useful story, and it comes from inside the industry: over 1,500 QA professionals, from individual contributors to senior executives, across North America, Europe, and Asia-Pacific.
Here is the twist: two of its biggest findings share the same number, and that is not a typo. 82% of respondents say AI is critical to the future of testing. In that same survey, 82% of testers say they still use manual testing in their day-to-day work. Those two 82%s are not in conflict. They are the whole story in two data points. And despite the headlines, only 20% of respondents say they are very concerned that AI could take their job.
AI is not simply replacing testers. It is reshaping what the role is for. The people who will be displaced are not "manual testers" as a category. They are testers who only do the things AI can already do well: the repetitive, scripted, predictable work. Everyone else is not being replaced. They are becoming more valuable.
THE SHORT ANSWER
AI is replacing specific manual testing tasks, not manual testers as a role. Katalon's survey of 1,500+ QA professionals found 82% still use manual testing day to day, even as 82% say AI skills are critical to the job's future, and only 20% are very concerned about AI taking their job.
• The tasks going first are scripted regression, data validation, basic smoke tests, and documentation from known requirements.
• Most teams are still early in their automation and AI maturity, only 11% have reached the most advanced stage, so the near-term shift is upskilling, not headcount cuts.
• The professionals coming out ahead are "hybrid testers" who combine manual judgment, automation, and AI instead of treating them as competing skill sets.
The job is not disappearing. What counts as valuable inside it is changing fast.
The parts of QA most exposed to AI are the parts that were already structured, repeatable, and time-consuming:
High-maturity QA organizations push this further still: they are also 1.5 times more likely to run automated UI testing than the broader survey sample. The pattern is consistent. The more mature the team, the more systematically it moves repetitive work off people's plates.
The idea that AI makes manual testing irrelevant does not survive contact with the data. 82% of all respondents in Katalon's survey still use manual testing. Even among the highest-maturity organizations, the ones furthest along in automation and AI adoption, 72% still rely on it.
So manual testing is not going away. What is changing is where it earns its keep. It is no longer strongest as a stand-alone execution function, running the same checklist every release. It becomes more valuable applied where scripts, rules, and AI outputs are least reliable:
That is why exploratory testing still shows up in 44% of teams' workflows. As more predictable test work gets automated, exploratory coverage becomes more important, not less, because it is the check nothing else performs.
Exploratory testing requires a tester who understands how real people use software and can intuit where things are likely to break. AI can execute the steps it is given. It cannot decide what to try next on a hunch, notice that an interaction feels off even though it technically worked, or ask "why would anyone do it this way?" UAT coordination has the same ceiling: what counts as "good enough" for a specific release, market, and user base is a judgment call, not a test result.
There is also a demand-side reason this work is not shrinking:
As AI writes more of the code, the surface area that needs testing expands. That is not a smaller job. It is a different one, and a busier one.
Katalon's report does not point to the end of the tester. It points to the rise of the hybrid tester: someone who blends manual testing, automation, and AI-assisted workflows instead of treating them as competing approaches. Hybrid testers let automation and AI handle routine tasks, then apply human judgment where context, risk, and user behavior matter most.
That pattern shows up throughout the data:
That last point matters: AI is not just helping teams execute faster. It is helping them work more systematically.
It is worth naming a tradeoff, too. Hybrid testers are 1.4 times more likely to prioritize AI and machine learning than the average QA professional, but 24% less likely to emphasize attention to detail. That is not a flaw so much as the whole arrangement working as intended: handing detailed scrutiny to AI only pays off if a person is still validating what the AI produces, which is exactly the judgment layer this piece keeps coming back to.
That makes the job look less like "run tests all day" and more like "orchestrate quality intelligently." Stronger teams are not choosing between humans and AI. They are combining them, which is a much better model for the future of QA than a headline that says testers are obsolete.
| At Higher Risk | Lower Risk |
|---|---|
| Only executes pre-written test cases without understanding why | Runs exploratory and usability sessions as a core part of the job |
| Scripted regression as the primary job function | Understands what automation and AI can and cannot cover |
| No familiarity with the tools their team uses for automation | Can have a technical conversation about where a bug originated |
| Cannot read a test result and distinguish a real failure from a flaky test | Involved in requirements review before development starts |
| Works in isolation from the development team | Works inside a cross-functional squad alongside developers |
| Has not used any AI testing tool in the last 12 months | Uses AI and automation for the repetitive parts and focuses on judgment work, the hybrid-tester profile |
Another reason the "AI replaces testers" narrative falls apart: most organizations are still early on the maturity curve Katalon's report maps.
That means the majority of teams are still building the process discipline needed to use automation and AI well. The near-term challenge for most organizations is not replacing testers. It is helping the testers they have evolve into more capable, more technically fluent, more hybrid quality professionals.
Katalon's maturity data explains why teams are changing. PractiTest's 2026 State of Testing Report explains why individual testers should care, with a salary analysis across 10+ years of experience:
Technical skills get you in the door. After about a decade, the market stops paying a premium for technical execution and starts paying one for strategic thinking. Katalon's own skills data points the same direction: 68% of testers still rank automation scripting and programming as essential, but it is followed closely by API and web services testing (43%) and problem-solving and analytical ability (42%). The future QA professional is not purely manual and not purely technical. It is someone who combines technical depth with judgment, the hybrid-tester profile again, this time showing up in a paycheck.
If you work in QA, the practical lesson is not to compete with AI on repetition. That is the wrong battle.
Katalon True Platform is built for exactly the transition this data describes:
All four are orchestrated through the Katalon AI Assistant under an "AI proposes, human approves" model.
This is the practical version of what Katalon's own research and this piece both point to: AI augments hybrid testers rather than replacing them.
The mechanical parts get handed to the agents. The strategic parts stay with the team. Teams using platforms built this way report higher satisfaction and lower operational costs, not because they cut headcount, but because the people they have spend more time on work that produces quality outcomes and less time on work that just produces test counts.
So, will AI replace manual testers? Not in the simple way most headlines mean it. AI is replacing some manual testing tasks, especially the repetitive and predictable ones. Current data does not show a world where human testers disappear. It shows a world where the most valuable testers become hybrid testers who still use manual testing, just more strategically.
The repetitive layer is being automated.
The judgment layer is becoming more important.
The testers who can work across both will be the ones who stay most valuable.