My Friend Just Left QA for Product. I Called It Smart.

When I saw the LinkedIn notification — former QA engineer, now Product Manager — my first reaction wasn't surprise. It was recognition. Smart move.

Not because QA isn't a meaningful profession. It is. But because the floor is shifting beneath it in ways that aren't temporary disruptions. They're structural changes. And the people paying attention are already moving.

Where QA Stands Today

The first thing to understand about counting QA professionals is that the official numbers significantly undercount the field — and knowing why matters for understanding the scale of what's changing.

The U.S. Bureau of Labor Statistics tracks a category called "Software Quality Assurance Analysts and Testers" (SOC code 15-1253), which reported 201,700 people employed in 2024 with a median salary of $102,610. That's the number you'll see cited most often. But it captures only one slice of the quality-focused workforce.

The problem is how the BLS classifies roles. SDETs (Software Development Engineers in Test), Automation Engineers, Software Test Engineers, and Quality Engineers are frequently coded under the broader Software Developers category (SOC 15-1252) because their primary work involves writing code — even when that code's entire purpose is testing. On top of that, the industry's shift toward titles like "Quality Engineer" has blurred the line further, with many practitioners holding QE titles while doing work that is fundamentally testing-focused.

When you account for those roles — SDETs, test automation engineers, software engineers in test, and quality engineers operating under various titles — industry analysts and workforce researchers put the true US quality-focused workforce closer to 400,000–450,000 people. Think of the BLS figure as the floor, not the ceiling: it counts dedicated testers and QA analysts. Still, it misses the substantial population of engineers whose primary mission is quality, just expressed through code.

Combined, this is a workforce that represents a real, significant part of the software industry. And it's facing two compounding disruptions simultaneously.

The question worth asking isn't where QA is today. It's where it's going.

The Thought Experiment: Two Waves of Disruption

To understand what happens next, you need to understand two separate but compounding forces hitting QA simultaneously.

Wave 1: GenAI — Already Here

Generative AI has already begun replacing the most labor-intensive parts of traditional QA work. Tools powered by large language models can now:

• Generate test cases from requirements documents in seconds
• Write and maintain test scripts without a human authoring them line by line
• Analyze test results and flag anomalies with more consistency than a human reviewer working at 4 pm on a Friday
• Create synthetic test data at scale, eliminating a task that once consumed significant QA hours

This isn't hypothetical. Teams that used to staff 5-person QA squads to write and maintain regression suites are discovering that one engineer with the right AI tooling can cover the same surface area — and cover it faster.

The impact here is real but partial. GenAI accelerates QA work. It compresses team sizes. It doesn't yet eliminate the human-in-the-loop. But it's already shrinking headcount requirements by an estimated 20–30% in organizations that have adopted it seriously.

Apply that conservatively to the ~425,000 quality-focused professionals in the US today: that's 85,000 to 125,000 roles that effectively evaporate — not through layoffs necessarily, but through attrition. Companies stop backfilling. Teams don't grow. The headcount quietly contracts.

Projected QA/quality workforce in 5 years (2031): ~300,000–340,000

Wave 2: Autonomous Testing — Coming Fast

The second wave is more profound, which makes this a career-level conversation rather than just a tooling conversation.

Autonomous testing platforms don't just help QA teams work faster. They replace the core function entirely for a growing class of applications. These systems can:

• Explore a web application end-to-end without any human-authored test cases
• Detect regressions and behavioral changes across releases automatically
• Adapt to UI changes without breaking — something traditional automation famously cannot do
• Generate reports that a human would previously have needed to compile manually

This is the difference between AI-assisted testing and AI-led testing. The former still needs a QA team to drive it. The latter fundamentally questions why that team needs to exist in its current form.

As autonomous testing matures over the next decade, the roles that survive will look very different from the roles that exist today. The surviving QA professional will be less "tester" and more "quality strategist" — someone who defines coverage philosophy, interprets AI-generated results, and bridges the gap between engineering and business risk. That's a significantly smaller team than today's model requires.

Historically, analogous technological shifts — industrial automation in manufacturing, self-checkout in retail, algorithmic trading in finance — have consistently reduced the headcount of dedicated specialists by 40–60% within a decade of widespread platform adoption.

Projected QA/quality workforce in 10 years (2036): ~170,000–210,000

The Numbers, Plainly

That's a potential 50–60% reduction in quality-focused roles over a decade. Not extinction — but a dramatic reshaping of what the profession looks like and how many people it employs.

What Actually Survives

This isn't a story about QA going away entirely. Testing as a discipline matters more than ever as software ships faster and with greater complexity. The question is who performs it and in what form.

The roles that will survive — and arguably thrive — are:

• Quality Engineers, not testers. People who can architect testing strategy, evaluate AI outputs critically, and own quality as a systems-level concern rather than a task to be executed.

• Domain experts embedded in product teams. People who understand the business risk of a bug, not just its technical characteristics. This is, notably, exactly the kind of person a QA professional becomes when they transition into a product role.

• AI/Automation specialists who manage and improve the autonomous systems doing the testing — a role that barely existed five years ago.

The purely execution-focused QA role — "run the regression suite, file the bugs, write the test cases" — is the one that doesn't survive. Machines are absorbing that work, and the timeline is shorter than most people in the field want to believe.

The LinkedIn Post Makes More Sense Now

My friend didn't leave QA because she got bored. She left because she could read the trajectory. The skills she built in QA — deep product knowledge, a systematic way of thinking about failure modes, an instinct for edge cases — are exactly what make a great product manager. She leveraged the equity in her experience and moved to where the puck is going to be, not where it has been (hockey reference).

That's not a critique of the people staying in QA. There's important, meaningful work still to be done there, and the best QA professionals will find the new shape of the role and fill it well.

But if you're early in a QA career and planning the next ten years on the assumption that the field looks roughly the same, the numbers suggest it's worth rethinking that plan.

The machines aren't coming for quality. They're coming for the manual execution work. Those are two very different things, and knowing the difference is the first step to being on the right side of the shift.