.png){kind=icon}
AI is changing QA work faster than HR can rewrite the job descriptions. QA teams are being asked to cover more software on shorter release cycles, while test suites keep expanding. AI is starting to absorb some of that workload, including responsibilities that once filled hours of a QA professional’s day. Watching familiar responsibilities shift to AI can make the future of QA feel uncertain. However, workforce data points to a more complicated shift. QA employment is projected to grow even as the work inside the role becomes more specialized.
QA has been through a shift like this before. Traditional test automation took repetitive checks off testers’ plates. Yet people still decided what to test, investigated failures, and judged the results. AI extends automation into some of the interpretive work previously left to people. PLUS QA has already seen that shift in AI-powered test automation, where systems can analyze previous test results and code changes to identify which parts of a product may be affected by an update.
A 2025 study from the International Labour Organization found that generative AI is more likely to transform jobs than eliminate them outright because some responsibilities can be automated while others still require people. For QA professionals, the implication is clear. As AI takes over more routine responsibilities, human judgment becomes more valuable.
Where AI in QA Stands Today
AI is being used more frequently for testing, though most organizations haven’t broadly adopted the technology. The 2025 World Quality Report from OpenText, Capgemini, and Sogeti found that 89% of responding organizations were piloting or deploying generative AI in quality engineering. Only 15% had reached organization-wide implementation.
The same research gives testers a clearer picture of the work already changing.
- 65% reported autonomous test script generation.
- 63% reported AI-powered test data generation.
- 49% reported self-healing tests that adapt when applications change.
- 67% reported AI-enhanced analytics that can support defect prediction and risk-based prioritization.
Those capabilities touch several parts of a traditional tester’s workload. AI can create first drafts of test cases, prepare data, maintain some automation, analyze defects, and repeatedly validate known behavior. One obvious fit is regression testing, where teams repeatedly check established behavior as software changes.
The report shows that productivity gains have varied considerably. Organizations using generative AI in quality engineering saw an average productivity boost of 19%, yet one-third saw only minimal gains. The technology is already saving time in some areas, but adopting AI alone does not guarantee a better testing process.
%20(1).jpg)
Where Testers Still Have the Advantage
Professional QA has never been limited to executing test cases. Testers build test plans, write custom cases, evaluate usability, investigate failures, recommend changes, and help teams decide whether a product is ready for release. Many of those responsibilities require testers to make decisions beyond the instructions in a test case.
Exploratory testing is one place where that broader role becomes especially clear. Testers decide where to look as they learn more about the product, rather than simply work through scenarios someone has already defined. They can also recognize when software technically works but still confuses users, when unusual behavior violates a business rule, or when a defect creates a bigger problem for customers than test results suggest.
The same judgment applies after a defect is found. A system can identify unexpected behavior, while a tester can weigh its impact against release priorities and explain the problem to developers, product leaders, designers, or other stakeholders. Ethical questions, cultural context, and empathy can also change what an acceptable result looks like for different people or communities.
Developers themselves remain cautious about handing those decisions to AI. In Stack Overflow’s 2026 Developer Survey, 48% of respondents said they trusted AI when they could easily verify its output. Only 6.6% trusted it for important work decisions. As AI contributes more code, tests, and analysis, verification becomes a larger part of maintaining quality.
How Agentic QA Is Changing Test Automation
AI agents can handle more of the testing workflow on their own. Instead of completing one requested task, an agent can receive an objective and work through a series of decisions.
- Interpret the testing objective.
- Decide what needs to be tested.
- Create or select the tests.
- Run them and evaluate the results.
- Refine the tests, retry a failure, or flag something for review.
When an agent makes several decisions in sequence, an error can snowball into bigger problems. Microsoft’s Agent-Pex was developed because agent behavior becomes harder to understand and validate across multi-step interactions. The system examines the record of each step an agent takes during a run, often called an agent trace, and checks whether the behavior followed expected rules.
Broader adoption creates practical challenges for QA teams as well. The World Quality Report found that organizations were already running into problems as they expanded their use of generative AI in quality engineering. Sixty-three percent cited data limitations or privacy and security risks, 64% cited integration complexity, 60% cited hallucination and reliability concerns, and 50% cited a lack of AI or machine-learning expertise. As agents take on more of the testing workflow, testers still need to review their decisions, investigate failures, and decide when the results require action.
What the QA Job Numbers Show
AI is already reducing the amount of human effort required for some testing tasks. QA employment can still grow even as AI takes on more of those tasks. Employment projections and surveys of IT and QA professionals show both trends happening at once.
6% projected growth in QA employment
The U.S. Bureau of Labor Statistics projects employment for software quality assurance analysts and testers to grow 6% from 2025 through 2035, compared with 3% across all U.S. occupations. BLS says demand is expected to remain strong as software development expands for AI, connected devices, robotics, and other automation applications.
The figure describes total employment across the economy. It does not mean every software project will employ more human testers. More digital products can create more QA work overall while AI simultaneously reduces the human effort required for individual tasks or projects.
75% of IT work expected to be done by people using AI
Business and tech research firm Gartner surveyed more than 700 CIOs in 2025. CIOs expected that by 2030, three-quarters of IT work would be performed by people using AI as part of their jobs, while a quarter of IT work would be performed by AI alone. The forecast suggests people will remain involved in most IT work as AI takes on more individual tasks.
34% of testers report more time for higher-value work
TestRail’s Software Testing & Quality Report surveyed nearly 3,000 QA professionals and others involved in software testing. Thirty-four percent said AI gave them more time for complex, higher-value work, while 29% reported greater efficiency. Another 26% said they needed to learn new tools or skills, while 36% said AI had not meaningfully changed their role.
The numbers show QA employment continuing to grow as the work within the profession changes. Roles centered on repetitive work face greater pressure as AI takes on more of those tasks. Changes already underway in QA roles are revealing which skills testers will need to stay relevant.
QA Roles Changing With AI
AI is creating something new for QA professionals to evaluate. Many software tests compare actual behavior against a defined expected result. Generative AI and agents complicate that process because they can produce different responses from similar inputs, make decisions across several steps, and generate answers that appear convincing while still being wrong.
The National Institute of Standards and Technology is developing guidance for testing and evaluating AI systems, including machine-learning models, large language models, multimodal systems, and AI agents. NIST’s work reflects a growing need for people who know how to test AI systems, not merely use them.
.png)
AI Test Engineer
This work extends QA into the behavior of AI systems themselves. Testers may validate outputs, probe hallucinations, test guardrails, examine agent workflows, and evaluate whether models behave reliably across different inputs and conditions.
.png)
Quality Architect
Quality Architects operate at a broader level. The role can include testing strategy, automation architecture, risk, quality standards, and decisions about how AI fits into the organization’s overall QA approach.
.png)
AI Model Validator
Model validation focuses more directly on whether an AI or machine-learning system satisfies defined requirements. Accuracy, consistency, bias, failure modes, and behavior outside the system’s expected inputs can all become testing concerns.
Vibe Tester
Vibe Tester is still a loosely defined title rather than an established QA role. Current uses range from AI-assisted product testing and UX evaluation to broader responsibility for AI-driven testing systems.
Hallucination testing, fairness evaluation, guardrail validation, and agent testing can be responsibilities across several of these roles rather than belonging to one specific job title. In all of these roles, testers are evaluating systems that can return different answers to similar inputs, making validation less straightforward than checking whether software behaves as expected.
The Skills QA Testers Need to Stay Relevant
The strongest QA skill set now combines established testing expertise with the ability to work with AI and test products that use it. Testers still need to understand how software fails, how systems connect, and how to design useful tests. Working with AI also requires testers to evaluate generated output, test model behavior and guardrails, and understand how autonomous agents work.
No tester needs to master every skill below. The ones that matter the most depend on the role, the product, and how closely the tester works with AI systems.
Test design and exploratory testing
AI can generate test cases quickly, which makes knowing whether those tests are useful even more valuable. Testers should understand boundary value analysis, equivalence partitioning, decision tables, state-transition testing, exploratory testing, error guessing, and how to turn requirements and acceptance criteria into meaningful coverage. These techniques help testers identify missing scenarios and weaknesses that an automatically generated test suite may overlook.
API testing
Much of a product’s behavior happens behind the interface. QA professionals should understand HTTP methods and status codes, REST APIs, JSON, authentication, authorization, request and response validation, error handling, and dependencies between services. API knowledge also becomes increasingly useful as AI applications connect models with databases, external services, and other tools.
Test automation
AI can write automation code, but testers still need to understand what the code is doing. Important skills include building and modifying automated tests, debugging failures, managing test data, handling asynchronous behavior, maintaining tests as products change, and identifying flaky tests. Familiarity with at least one major automation framework also helps testers evaluate code generated by AI instead of accepting it at face value.
Git, CI/CD, and troubleshooting
Technical QA roles increasingly require testers to understand where tests fit into software development. Useful knowledge includes Git, pull requests, build and deployment pipelines, CI/CD, application logs, test logs, and differences between testing environments. These skills help testers trace failures through a system and work more closely with developers when something goes wrong.
Using AI for testing
Prompt engineering is becoming a practical QA skill. Testers should know how to give AI enough information to understand a testing problem, define constraints, specify expected outcomes, generate test ideas or data, and refine instructions when the output misses the mark. Just as important, they need to recognize when an AI-generated test, explanation, or recommendation is incomplete or wrong.
Testing AI-generated output
AI products introduce failure modes traditional applications may not have. Testers working with generative AI should understand how to evaluate hallucinations, factual accuracy, relevance, consistency, bias, safety, and guardrail failures. They also need to recognize responses that sound convincing while failing the requirements they were supposed to meet.
Testing AI agents
Agents require testers to follow behavior across several decisions rather than evaluate only the final result. Useful skills include reviewing agent traces, checking tool calls, testing multi-step workflows, examining failure recovery, finding errors that carry into later steps, and using adversarial tests to expose unexpected behavior. Testers also need to verify that an agent stayed within its instructions while completing a task.
Model and data validation
QA professionals do not need to become data scientists to understand how models and data affect product quality. Useful knowledge includes accuracy, consistency, robustness, bias, test-data quality, behavior outside expected inputs, and the basics of how training and evaluation data are used. Testers working with retrieval-augmented generation should also understand whether the system retrieved appropriate information and used it correctly.
Risk-based testing and test strategy
As AI makes it easier to generate and execute more tests, deciding where to focus becomes increasingly valuable. Testers should understand how to rank failures by likelihood and impact, identify high-risk user journeys, determine what deserves automated coverage, recognize where human testing adds more value, and judge whether testing has provided enough confidence for a release. Product and industry knowledge become especially important here because the consequences of a failure depend on what the software does and who relies on it.
What QA Professionals Should Do Right Now
The skills list may look long, but most QA professionals do not need to tackle everything at once. The better approach is to identify where your current work is changing, strengthen the skills closest to your role, and build from there.
- Audit your current work
Look at which responsibilities are repetitive, which require investigation or product knowledge, and where AI is already changing the amount of time you spend. - Choose a few skills to deepen
Use the previous section as a guide. Focus first on the areas closest to your current role or the job you want next. - Use AI in your existing QA work
Apply it to test design, test data, failure analysis, or another task you already understand well enough to evaluate. - Challenge the output
Check what AI missed, where it made bad assumptions, and whether the tests or analysis it produced are useful. - Build proof of what you can do
Keep examples of projects where you used AI, found problems it missed, improved its output, or tested an AI feature directly.
Using AI in day-to-day testing helps testers learn where it performs well, where it fails, and when human review is still needed. Direct experience teaches those lessons faster than reading about the technology alone.
What This Shift Means for QA Teams and Businesses
AI can help QA teams run more tests with less manual work, especially when the testing is repetitive and clearly defined. Experienced testers remain important when testing requires exploration, product knowledge, risk assessment, or an understanding of how people will use the software. For teams, the result is broader test coverage without losing human decision-making.
As companies give AI more responsibility in testing, they also have to address privacy, integration, reliability, and governance. Teams need clear rules for when AI can act on its own, how its results are checked, and who is responsible when those results influence a release decision. AI can take on more testing work, while teams remain accountable for product quality.
PLUS QA is already combining those approaches. MAX AI handles repetitive, high-volume testing and expands coverage, while PLUS QA testers concentrate on complex scenarios, exploratory testing, and UX analysis. AI can give QA teams more capacity, while experienced testers determine whether the results are useful and whether a product is ready for its users.