Shift left might need failed – however AI seems to be set to ship on its promise
“AI will replace QA.” It was not the first time I had heard this claim. But when someone said it to me directly, I asked them to demonstrate how and they simply couldn’t.
That exchange occurred shortly after my co-founder Guy and I launched our second company, BlinqIO. This time, we focused our efforts on building a fully autonomous AI Test Engineer.
We developed an advanced platform that was not only capable of understanding applications under test, generating and maintaining robust test suites, but also recovering from failures independently.
I’m pleased to say that the technology worked. However, from speaking with numerous global enterprises, there was constant concern, not about functionality but trust and control when it comes to AI tools.
CEO and Co-Founder, BlinqIO.
The Limitations of Shift Left
Across industries, we are seeing organizations come under intense pressure to release software faster than ever.
Methodologies like Agile, CI/CD, DevOps, and Shift Left were all introduced to accelerate delivery without compromising quality. However, as Shift Left was implemented, its original intent was often either misunderstood or misapplied.
Originally intended to embed testing earlier in the development lifecycle, Shift Left too often resulted in the marginalization or elimination of dedicated QA roles altogether.
Developers were soon being asked not only to build features but to verify their correctness without independent validation. On paper, this may appear to be an efficient way of working; however, in reality, the consequences are evident.
From experience, developers lack incentives to test their own code, and, as a result, coverage frequently becomes deprioritized.
From my perspective, Shift Left did not fail because it was inherently flawed; rather it failed because it wasn’t yet complete and ready to be used the way it should.
I believe successful implementation of Shift Left requires rethinking collaboration models, redefining shared accountability, and embedding quality throughout the software life cycle.
In companies where it works, teams are not simply writing tests earlier but reframing how risk is assessed, how requirements are defined, and how feedback is used to drive continuous improvement.
Simply removing QA and assuming innovation will compensate is a false economy. Trust me when I say it doesn’t work.
FOAI: Fear of AI
Today, Artificial Intelligence is poised to give Shift Left a second chance. But widespread adoption remains hampered by a new and growing barrier, this is something I like to call ‘FOAI’, which is Fear of AI.
This fear is not rooted in science fiction. It is being replaced among even the most innovative employees. The fear of being accountable for decisions made by systems they don’t understand.
Most importantly, it’s about the fear of relinquishing control to technology introduced without adequate explanation or transparency.
In theory, I think most tech founders would agree that AI should be embraced. In practice, it is often introduced as a black box-opaque and seen as unexplainable yet mandatory.
Teams are expected to trust something they cannot interrogate. This, in turn, undermines confidence and fuels resistance. I have witnessed how quickly resistance can dissolve when people are invited into the AI adoption process.
When teams are able to fully understand how AI actually functions and the ways in which it prioritizes tests and why it flags certain failures, their entire perspective tends to shift.
Teams that began with skepticism are now using our platform to autonomously manage thousands of tests with confidence. This transformation was not just about the technology but the trust that developed once transparency and control were brought into the mix.
Leadership in AI – A Personal Perspective
I believe trust is key when it comes to technology adoption. In addition, I believe it’s helpful to also identify who within a team shapes these technologies and helps implement them.
Working in AI and deep tech as a female founder means navigating often subtle, persistent barriers. There is usually an unspoken expectation to prove one’s technical authority over and over. These reflect deeper assumptions about just who is seen as qualified to help companies build their future with AI in it.
What has helped me, personally and professionally, is visibility. When women are seen founding and leading AI companies and not just using AI, but building it, this challenges some deeply rooted biases.
This is why I remain active not only as a speaker at events, but in mentorship groups, panels, and one-on-one conversations. It’s to help with AI transition and acceptance.
To me, inclusion must go beyond representation. It requires access to influence. It means being present in the rooms where decisions about technology, ethics, and impact are being made. I think that the future of AI should be co-created by everyone using it.
Decoding the Language of AI
In the current landscape, artificial intelligence is surrounded by an often overwhelming level of jargon. From LLMs, agents, neural networks and synthetic data to autonomous systems. While AI-related terminology can be daunting, it needs to be understood.
In high-stakes domains such as healthcare, finance, and enterprise software testing, AI must be accountable. Teams need to know not just what happened, but why. Another is agentic behavior systems that operate autonomously on behalf of humans.
This kind of functionality is already present in modern AI platforms. But to use it safely and effectively, teams must be able to monitor and adjust how AI systems function in real-time. Without this, building that much-needed trust is nearly impossible.
The Future of AI – Quiet, Powerful, and Integrated
I don’t believe AI will change the world through one dramatic breakthrough. I think that its most powerful effects will unfold quietly and that this will happen within infrastructure, beneath user interfaces, and behind the scenes.
Future-ready AI will not necessarily announce itself with glossy demos. Its contributions will be measured not in headlines, but in release stability, in faster recovery cycles, and in the confidence with which teams ship software.
This shift will also reshape the value we place on human capabilities. As AI increasingly automates repetitive, mechanical tasks, the skills that rise to prominence will be curiosity, strategic thinking, and the ability to frame complex problems.
In my view, these are the traits that will define effective leadership in an AI-enabled world and not just one of technical proficiency.
The companies that will thrive in the future will be those that integrate AI in a thoughtful manner. Those that treat trust, quality, and explainability as essential design principles and not afterthoughts will be setting themselves up for success. I also think those that view AI not as a replacement for human insight, but as an enabler of it will perform well.
Trust me when I say that AI will not replace workers. However, I do believe that ignoring its potential or implementing it without transparency may hinder your organization’s future.
As for Shift Left? It may have fallen short the first time. But I think that with the right application of AI, we have an opportunity to try again, this time with the tools, mindset, and visibility to get it right.
We’ve featured the best Large Language Models (LLMs) for coding.
This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro