AI agents don't just fail faster than humans—they fail at scale, which makes validation and testing more critical than ever. The real challenge isn't the technology itself, but solving the oracle problem and the communication problem: how do we define what we want, and how do we verify we got it? As code becomes increasingly black-box and generated on-demand, the spec and the test become the only reliable source of truth. For testers, that's not a threat—it's an invitation to lead.
In this episode, I talk with Henri Ternho, senior consultant and AI enthusiast, about why building trust in AI systems requires the same rigor we've always applied to software—just now at a whole new level. Henri explains how AI agents multiply both our successes and our mistakes, why prompting is harder than it looks, and why testers are uniquely positioned to thrive in this shift. We dig into the oracle problem, the communication trap, and why your test suite might soon matter more than your codebase.
"AI doesn't think, it doesn't analyze, it predicts." - Henri Ternho
Henri Terho is a Senior AI Consultant at Eficode with broad experience spanning regulated industries—automotive, banking, aerospace, and beyond—alongside a deep commitment to open-source collaboration. He has played a key role in fostering community-driven innovation, having served as chairman of Tampere Entreprenourship society and co-founding Tampere Tribe to support local startup culture.Henri’s passion for AI, quality assurance, and rapid software development is evident in both his industry work and ongoing PhD research on agile product innovation. He frequently shares his expertise on stage and in publications, championing lean practices and the latest AI advances to empower organizations worldwide.
The rise of AI agents is rapidly changing the landscape of software development and quality assurance. As automation becomes more embedded in our processes and decision-making, conversations around how to create trust in these systems are more urgent than ever. In a special episode of Software Testing Unleashed, recorded live at The HUSTEF Conference 2025 in Budapest, Richie welcomes Henry Tero, a seasoned consultant and AI enthusiast, to explore what trust means in the era of AI. Together, they dig deep into the practicalities, fears, and future roles for testers working with intelligent systems.
Trust isn’t a new challenge—developers and testers have always had to ensure confidence in software, whether for themselves, their teams, or millions of users. But as Henry Tero highlights, AI agents amplify this need. Automated systems can multiply both success and failure. While human colleagues may fill in gaps based on experience, AI has no intuition for context or unstated expectations. As Henry notes, it's crucial to build systems where trust in AI is as natural as trusting your teammate: “You want the system around the AI model to be that source of trust, that you have a process, checks and balances, guardrails and all this stuff built around it”.
Communicating requirements is often at the heart of these trust gaps. Whether talking to a person or an AI, miscommunication leads to wrong outcomes and disappointment. The classic “Oracle problem” emerges: How do we know what the AI provides is correct, or even useful? As Henry points out, asking for a business strategy is different if you’re in Finland or Germany—the AI must be guided by precise, detailed input, much like humans.
In practice, AI systems are often described as “fluffy” and mysterious. For many, handing over control to AI feels risky, particularly with sensitive data or critical business functions. A lack of transparency and clear validation processes makes it hard for organizations and users to trust outcomes. The unpredictability inherent in modern AI models adds an extra dimension; as Henry explains, “software should be deterministic… with AI models, of course, being non-deterministic, you have to shift your thinking”.
To build trust, familiar tools such as QA, validation, and end-to-end testing must evolve. Instead of simply checking if a single test passes, organizations need to watch trends, monitor metrics like mean time between failures, and be ready for unexpected “black swan” events. As AI systems evolve under the hood—thanks to regular model updates—the need for continuous validation and robust monitoring becomes even more pronounced.
Far from making testers obsolete, AI’s rise makes quality work even more central. Both Richie and Henry Tero emphasize the vital responsibility testers have in building trust into these systems. As Henry explains, writing clear specifications, crafting test cases, and validating actual outcomes are exactly the skills needed to manage AI-driven projects.
Fundamental practices like BDD (Behavior Driven Development) and TDD (Test Driven Development) are more relevant than ever. AI-generated code is an extreme black box—often, all that testers have is the specification and the resulting behavior. Expanding one’s skill set beyond traditional testing, exploring statistical testing methods, and understanding adjacent areas such as DevOps and development all help testers remain invaluable in an AI-first environment.
Change, especially at this scale, naturally creates anxiety. Many fear losing relevance as AI automates tasks or feel lost confronting the unknown mechanics of these systems. Yet, as Henry encourages, the very skill sets that define good testers—critical thinking, system-level perspective, and context awareness—are exactly what the AI era needs most.
Embracing new tools and techniques, focusing on communication, and learning to analyze trends instead of seeking absolute answers are ways testers can not just survive but thrive. AI won’t take the job; instead, it augments testers, opening opportunities for a more strategic role in shaping software quality.
As AI agents become ever more integral to the fabric of software, the quest for trust only intensifies. The journey is challenging but exciting—the shift in software quality from checking boxes to enabling intelligent, context-rich systems offers testers and developers alike a game-changing opportunity to redefine their craft.