QA trends for 2026: Insights from Tricentis Transform

AI is fundamentally reshaping software quality, and the organizations leading this shift aren’t waiting to adapt. In October 2025, we brought together over 1,000 quality engineering leaders, practitioners, and innovators for Transform, our annual conference exploring what’s next in software delivery. Across all expert-led sessions, keynotes from industry pioneers, and real-world success stories, we saw the trends defining 2026: AI-driven quality engineering is happening now, and it’s redefining what’s possible at scale.

Whether you joined us in person or are tuning in now, Transform 2025 was the definitive event for discovering the insights and trends that are shaping the next era of software quality.

AI is significantly expanding the surface area for business risk

In the opening keynote, Tricentis CEO Kevin Thompson emphasized that over 40% of code written last year was generated by AI. However, we have far less data on how much of this code survived review and made it into production. The confidence gap is real: in a Stack Overflow survey, 88% of respondents said they weren’t confident deploying AI-generated code, while a GitLab survey found that 29% had to roll back releases due to AI errors. This accelerating pace of change demands a fundamental shift in how teams work, marking software quality engineering’s entry into a new AI-driven era.

To address this gap, Tricentis has developed AI agents that create tests autonomously, self-drive execution, problem-solve, and deliver results.

David Colwell, VP of AI, and Eran Sher, Chief Product Officer at Tricentis, shared what Tricentis is building to meet these emerging needs:

• Agentic test generation creates complete test cases from natural language prompts, user stories, or requirements — dramatically reducing manual authoring effort.
• Agentic quality intelligence continuously analyzes code changes and coverage to identify testing gaps, then automatically generates tests to close them.
• Agentic load testing builds sophisticated load test scenarios with realistic user behavior patterns.
• A conversational testing interface ties it all together, letting teams interact with all agentic capabilities through intuitive, chat-based tools that make AI-powered testing accessible to everyone on the team.

These innovations signal Tricentis’s commitment to not just keeping pace with AI’s disruption of software development, but actively solving the quality challenges it creates.

QA is becoming an accountability layer for AI

When AI fails, it fails at enterprise scale. A central takeaway from Transform this year was how quality assurance is becoming the critical accountability layer for AI-driven software delivery.

Where QA teams once wrote scripts and ran tests, they now define quality objectives, oversee AI-generated results, and ensure automated decisions align with business priorities. This shift reflects lessons from the field: AI projects fail not because of the technology itself, but because of poor data, adoption challenges, and scalability issues, areas where domain expertise proves essential. Modern QA teams are building closed-loop AI systems where agents can author, execute, and analyze tests autonomously, but with human oversight providing the governance needed to keep these systems reliable, scalable, and aligned with real-world needs. In this new paradigm, QA isn’t just testing software, but validating the AI that’s increasingly building and testing that software.

Human-in-the-loop is non-negotiable

One theme echoed across many sessions at Transform: AI is probabilistic, not deterministic. This is a fundamental difference that changes everything about how we deploy it. Scaling AI without proper governance doesn’t just create risk; it leads to systemic failures that can undermine entire initiatives. The numbers tell the story: 95% of AI pilots fail due to lack of appropriate guardrails, with 60% of those failures involving compliance issues that could have been prevented.

The path forward requires a deliberate strategy built on clear KPIs, continuous monitoring that treats AI systems as living processes rather than one-time implementations, and parallel testing to validate AI decisions against established benchmarks. In an era where AI can transform quality engineering at unprecedented speed, the organizations that succeed will be those that build human oversight into the foundation as a core principle.

Early AI adopters are pulling ahead

In a session about setting the workforce training agenda, Sophia Velastegui, former Chief AI Officer at Microsoft, put it bluntly: “This is the smallest gap right now between people who know AI and don’t. But that gap is going to get bigger.” In the same session, Vittorio Cretella, former CIO of Procter & Gamble, was even more direct: “You’re going to lose your job to AI if you don’t learn to use AI.”

The path forward isn’t about fearing displacement; it’s about seizing the opportunity to lead. Organizations and individuals who demonstrate tangible business value with AI, who lead by example rather than waiting for permission, and who embrace change as a competitive advantage are the ones who will define the next decade of software quality. The question isn’t whether to adopt AI, but whether you’ll be among those who shape how it’s used.

The role of QA is fundamentally changing

For decades, QA professionals have been the executors: writing scripts, running tests, and manually validating results. That era is ending (catch our recent webinar on how QA is changing here).

Today’s QA leaders are becoming orchestrators, defining quality objectives and overseeing AI-driven outcomes rather than executing every test themselves. We can already measure the impact of this: One Tricentis customer achieved an 85% reduction in manual effort and a 60% increase in productivity using AI agents. As organizations face an avalanche of AI-generated code, the ability to orchestrate quality at scale rather than test line by line has become the difference between keeping pace and falling behind.

AI Agents are co-workers, not just tools

Not all AI is created equal, and understanding the difference matters. While AI assistants suggest, recommend, and advise, they don’t work for us. But agents do. An AI agent operates as your co-worker, combining memory, reasoning, orchestration, and instruction to act autonomously on your behalf.

However, at least 60% of AI-generated code contains issues that require intervention, demanding entirely new approaches to testing and quality assurance. The organizations that will succeed with agents aren’t just those that deploy them, but those that build the guardrails to ensure they deliver reliable, compliant results.

Maximize risk coverage over test coverage

The traditional goal of maximizing test coverage is giving way to something smarter: maximizing risk coverage. Instead of testing everything equally, intelligent orchestration focuses effort where it matters most. This could potentially reduce overall test time by 40% while improving quality outcomes.

This shift is essential because GenAI testing is inherently probabilistic. You’ll never reach 100% coverage, and chasing it wastes resources that should be directed toward what protects your users and your business. In this new paradigm, edge case testing and exploratory testing become critical disciplines. Identifying the scenarios where AI might fail in unexpected ways and ensuring that risk, not just code, drives your testing strategy.

Looking to the future

Beyond the frameworks and statistics, some of the most valuable insights at Transform came from leaders who’ve already navigated these transitions. In a session on building enterprise-scale AI processes, Luke Mahon, Product Marketing Director at Tricentis, emphasized the importance of starting small and scaling thoughtfully. In a session on how testers can become agents of change, Nibs Mishra, AVP, Customer Solutions Technology Leader at Nationwide cut to the heart of a common mistake: “Don’t throw tools at the problem…bring your workforce with you.”

Tom Sweeney, Vice President of Enterprise Technology at Ford, delivered a rallying cry that captured the spirit of the entire conference: “The faster you adopt, the better off you’ll be. Don’t be afraid—embrace it.” The path forward requires both courage and strategy, and the organizations and ideas represented at Transform proved that the combination is already happening.

Stay tuned for Transform 2026 announcements, and watch 2025 session recordings here.