When I look at today’s manufacturing environments, quality control is no longer a back-office function; it is a core strategic lever. As global competition intensifies, tariff uncertainty sets in and customer expectations rise, manufacturers are under pressure to deliver flawless products with ever-greater efficiency. Artificial Intelligence, particularly in the area of quality control and defect detection, is rapidly emerging as a transformative force, one that moves quality assurance from a reactive necessity to a proactive driver of value.

Moving from Reactive to Predictive Quality

Historically, quality control has relied on post-production inspection: identifying defects after the fact, scrapping or reworking flawed goods, and hoping to minimize losses. This approach, while necessary, is inherently inefficient. It introduces waste, delays, and lost opportunity. AI is fundamentally shifting this paradigm. By deploying computer vision and deep learning models directly on production lines, manufacturers can now detect anomalies, such as cracks, dimensional inaccuracies, or surface inconsistencies, in real time.

Consider the steel industry, where even minor surface defects can have significant downstream consequences. AI-powered inspection systems, equipped with high-resolution cameras and sophisticated algorithms, can identify deviations invisible to the human eye. This not only reduces scrap rates but also ensures that only compliant material progresses through the value chain, minimizing costly downstream rework.

The Role of Generative AI in Data Augmentation

One of the persistent challenges in building robust AI inspection systems is the availability of sufficiently diverse training data, especially for rare or subtle defects. In my own work, I have encountered this bottleneck firsthand. Traditional computer vision models require thousands of labeled images (both defective and non-defective) to achieve high accuracy. However, capturing enough real-world examples of infrequent flaws is often impractical.

This is where Generative AI (GenAI) has proven invaluable. By training GenAI models on a curated set of actual defect images, I have been able to synthesize a much broader variety of realistic defect scenarios. These synthetic images enrich the training dataset, enabling our AI models to better recognize even the most elusive anomalies. The result is a significant acceleration in model development and a marked improvement in detection rates for rare defects.

Bridging Still Images and Video: Enhancing Real-Time Monitoring

My experience has also highlighted the importance of moving beyond static image inspection to dynamic, video-based analysis. At Project X, I helped develop our VisionX AI solution for this very task. Manufacturing processes are inherently dynamic, and many defects only manifest under certain conditions or in motion. By leveraging GenAI to generate synthetic video sequences depicting a range of defect scenarios, I have been able to train models that are not only robust in still-image recognition but also highly effective in real-time video extraction.

This capability is particularly impactful in high-throughput environments, such as steel rolling mills, where continuous monitoring is essential. Video-based AI systems can track products as they move through the line, identifying defects as they emerge and enabling immediate intervention. This real-time feedback loop minimizes the propagation of defects and supports continuous process improvement.

Beyond Pass/Fail: Strategic Tiering and Value Optimization

AI’s impact on quality control extends beyond simple defect detection. Deep learning models can classify products into multiple quality tiers based on the severity and type of imperfections. For example, in steel manufacturing, material with minor cosmetic blemishes, but full structural integrity, can be automatically routed to less critical applications, rather than being scrapped. This nuanced approach maximizes resource utilization, opens new market channels, and supports sustainability initiatives by reducing waste.

Perhaps the most transformative aspect of AI-powered quality control is the actionable intelligence it generates. By aggregating and analyzing inspection data, manufacturers can identify trends, correlate defects with specific process parameters, and implement early warning systems. In my own projects, integrating quality data with operational metrics has enabled us to pinpoint root causes, optimize process settings, and even predict equipment maintenance needs based on subtle shifts in product quality.

The Strategic Imperative

Adopting AI-driven quality control is not simply a technological upgrade—it is a strategic imperative. The benefits are tangible: enhanced brand reputation, reduced costs, increased throughput, and greater agility in responding to market demands. More importantly, it fosters a culture of data-driven continuous improvement, positioning manufacturers to thrive in an increasingly competitive landscape.

The path forward requires investment in data infrastructure, talent development, and a clear vision for how AI can support organizational goals. But the payoff is clear: a transition from incremental improvement to fundamental transformation in how quality is managed. As my own work with GenAI and video-based inspection has shown, the future of manufacturing quality is not just automated, it is intelligent, adaptive, and deeply strategic.

Share

Lately I’ve been reflecting on leadership and how it can drive or hamper innovation. I was going back through my journals from 5 years ago and had time to reflect on a key learning moment. In early January 2020, my risk radar picked up on something subtle but unsettling. At that time I was the Head, Innovation and Strategic Projects for a top private school in Toronto. A few fragmented mentions of a respiratory illness emerging in Wuhan had made their way into the global news cycle, tucked between updates on trade policy and holiday season retrospectives. At the time, there wasn’t a name for what we would all come to know as COVID-19—but the indicators were enough to trigger a sense of unease. I was responsible for analyzing the risks of an upcoming international student trip, and the variables were beginning to shift. What had been a low-probability outlier was edging closer to center stage.

This is not a story about the pandemic. We’ve read volumes on that already. Rather, this is a story about a leadership lesson that has shaped how I view risk, innovation, and the responsibilities of futurist thinking. It is a story about a moment when foresight met decisiveness, and what happened when leadership created the conditions for something extraordinary to emerge.

It was Karen Jurjevich, my former boss and a leader of uncommon wisdom, who recognized that my analysis wasn’t just identifying risk. It was creating the conditions for accelerated innovation. What followed was one of the most meaningful leadership interactions I’ve had in my career.

The Risk We Don’t Yet Understand

Great leaders often recognize signals others miss, not because they have a crystal ball, but because they know how to pay attention. I brought my early concerns to Karen with the kind of cautious precision that comes with good analytical practice. I laid out what we knew, what we didn’t, and the implications for the student experience and institutional liability. The goal wasn’t to predict the future, it was to prepare for a range of possible futures.

What struck me was not just her receptivity, but the way she used that moment to pivot. Karen didn’t react with fear or bureaucracy. She listened carefully, then shifted the conversation. “So,” she asked, “what would we do if we had to cancel the trip? If school wasn’t in person? What would that look like?”

The shift from problem-identification to solution-acceleration was immediate. She wasn’t looking to slow things down or wait for more data. She was looking to speed up the right experiments and she knew I already had one underway.

Building the Future Before You Need It

Long before Zoom became a household word, I had been experimenting with virtual environments. I was particularly interested in Virbela, a platform that allowed for immersive collaboration in a digital campus. In 2019, this was still fringe territory—something between a curiosity and a forward-looking prototype. But in a world where flexibility, engagement, and global reach were becoming strategic priorities, it felt like an important area to explore.

When I mentioned the project to Karen, she didn’t just support it. She told me to go faster. The same was true for Zoom, which we had already deployed in a limited fashion for internal collaboration. Her message was clear: the signals were strong enough, the foundation was already there, and the time to scale was now.

By the time closures began rolling in March of 2020, we weren’t scrambling to stand up new systems. We were optimizing what we had already built. The acceleration she enabled in January became a strategic differentiator just a few months later.

Risk as a Catalyst for Innovation

One of the most misunderstood aspects of risk is the idea that it’s something to be managed away. In practice, risk is a lens—a way to view not just what might go wrong, but where value might be created in the face of uncertainty. Karen’s leadership modeled this beautifully. She didn’t just see risk as something to contain. She saw it as an opportunity to do something better, faster, and more aligned with where the world was headed.

That mindset—of using risk analysis as a springboard rather than a stop sign—is a hallmark of leaders who operate with a futurist orientation. It requires the humility to act before you have full certainty and the confidence to invest in ideas that aren’t yet validated. It’s not reckless. It’s rigorous. And it’s rare.

The Power of Experimental Capacity

In retrospect, what made this leadership moment so powerful was that the experimentation was already happening before the pressure hit. We weren’t reacting. We were responding, building on prototypes and pilots that had been quietly developed in the margins.

This is one of the critical lessons I share with organizations today: your capacity for innovation in a crisis is directly proportional to the quality of your experiments before the crisis. If you want to be ready when the world shifts, you need to be exploring the edges of what’s possible long before the need becomes urgent.

Karen didn’t tell me to pause. She told me to push. She created a leadership context where the perimeter of the organization—the place where innovation often begins—was not only safe to explore, but strategically important to accelerate.

From Trendspotting to Organizational Readiness

Futurist thinking is not about predicting the future. It’s about preparing for multiple versions of it—and ensuring your organization has the capacity to adapt, absorb, and act. What distinguishes strong futurist leaders is their ability to turn early signals into institutional readiness.

By January 2020, it was already becoming clear that the future of education needed to be more hybrid, more resilient, and more globally agile. COVID-19 simply forced the acceleration of those truths. The organizations that thrived were not those who had the best contingency plans. They were the ones who had been building toward the future before the contingency arrived.

Karen Jurjevich modeled that posture in how she led our conversation, and in how she empowered experimentation as a strategic asset.

What This Teaches Us About Leadership

Leadership, when done well, is about more than steering. It’s about cultivating a culture where ideas are tested early, risk is examined with nuance, and readiness is built from the edge in. Too often, we associate leadership with stability and predictability. But in moments of inflection, the real measure of leadership is adaptability—and the ability to reframe uncertainty as possibility.

Karen’s leadership in that moment gave me a living example of how to create that frame for others. It showed me how to use foresight to accelerate meaningful experimentation and how to turn quiet projects into strategic pivots. It reminded me that leadership is not about being the smartest person in the room. It’s about seeing the emerging horizon and asking the right people to sprint toward it.

Innovation Requires Sponsorship

The truth is, most innovation fails—not because the idea is flawed, but because it lacks sponsorship. Karen gave me something more valuable than funding or resources. She gave me belief. And that belief was catalytic.

When leaders demonstrate belief in their people’s forward-looking experiments, they do more than de-risk innovation. They make it contagious. The result isn’t just better preparedness. It’s a culture where strategic imagination has room to breathe—and where people are willing to bet on the future because they know their leaders already are.

Looking Forward

Today, as I guide organizations through their own transformations—whether driven by AI, data strategy, or macro uncertainty—I find myself drawing on that leadership moment more often than I expected. It reminds me that risk is not the enemy of progress. It’s the compass that points us to where progress is most needed.

It also reminds me that the future is already being built, quietly, in the experiments we support or suppress. Leadership is the deciding factor in whether those experiments stay dormant—or become the differentiators that carry us through change.

The lesson from that January isn’t about how we responded to COVID-19. It’s about how one leader saw the signal before the noise—and helped accelerate the future before the world caught up.

Anthropic’s newest AI model, Claude Opus 4, recently did something remarkable: it coded on its own for nearly seven hours straight. Not just small, synthetic tasks—but the kind of real-world software development work that normally demands the focused attention of an experienced engineer. This isn’t just a technical flex. It’s a moment we’ll look back on as a shift in the relationship between human ingenuity and machine capability.

We’ve seen AI tools assist with programming before. Code completion. Syntax correction. Occasionally some debugging. But what Claude Opus 4 achieved is fundamentally different. It demonstrated what looks like autonomous software development. That’s a term we’ll start using more often, and it’s going to mean a lot. Claude didn’t just write a snippet or help you finish your function—it took a real coding task, involving multiple files, layers of logic, and codebase context, and it kept going. Quietly. Competently. For hours.

At a technical level, the numbers are impressive. Claude Opus 4 scored 72.5% on the SWE-bench benchmark—far beyond GPT-4’s 54.6%. It worked on a live open-source codebase, handed to it with minimal prompting, and iterated through multiple coherent revisions. But this newsletter isn’t about benchmarks. It’s about what happens next.

So what? Why does this matter for business, for strategy, for leadership?

Because when machines begin to hold the pencil—and do so with competence and persistence—we need to rethink what we’re drawing.

Claude Opus 4 signals that we’re entering an age where the capacity to execute is less of a constraint. If the execution of logic, design, integration, and even testing can be handled—at least in part—by AI, then the differentiator becomes something else. Your competitive edge won’t come from how fast your developers can write code. It will come from how fast your teams can imagine something worth building. That’s a profound inversion of how most digital organizations have operated for the last two decades.

We’re no longer just automating tasks. We’re now automating intention. That is: AI can take a high-level direction—"refactor this", "build that"—and carry it out across hours of sustained, context-sensitive effort. That opens the door to AI not as a tool, but as a collaborator. Not one you delegate to because you’re lazy, but one you strategically orchestrate because you’re smart.

This changes how we build teams. We’ll need fewer coders and more synthesizers. Fewer people writing code, more people deciding what needs to be built, what problems are worth solving, and how to express that clearly enough for a machine to understand. The new leverage is not lines of code—it’s clarity of intent.

This also changes what it means to scale. In traditional tech, scale meant hiring. If your backlog grew, you added headcount. Now, scale can mean spinning up intelligent agents—reliable, fatigue-free agents that work through complex problem sets without interruption. How does this change pricing models? Delivery SLAs? Time to MVP? Everything.

And here’s where it gets even more interesting. Claude Opus 4’s seven-hour coding session wasn’t about throwing spaghetti at the wall. It was methodical. It searched. It reasoned. It remembered. It revised. Anthropic calls this an "extended thinking loop." Strategically, it points to a future where we don’t just co-pilot with AI, we orchestrate end-to-end build systems that are intelligent, proactive, and deeply context-aware.

It’s easy to miss the deeper story in the headlines. The big idea isn’t that Claude Opus 4 codes for seven hours. The big idea is that we now have to reimagine the workflows, roles, and strategic decisions surrounding software development—and beyond.

The organizations that win will be those that shift their focus from operational velocity to strategic imagination. From code as an asset to capability as a service. From effort as value to insight as leverage.

This isn’t a future to fear. It’s one to shape. And shaping the future starts with asking a new kind of question: If AI can now code the solution, what’s the real problem worth solving?

Bell Canada’s recent announcement of the Bell AI Fabric signals a true shift in Canada’s technology and economic infrastructure. With plans to develop the country’s largest AI compute project, Bell is not just entering the AI infrastructure game—it is asserting a leadership position in shaping what sovereign, sustainable, and scalable AI capability should look like in a national context. The initiative begins with six hydro-powered data centres in British Columbia, forming a supercluster with a planned 500 MW capacity, and expands with clear intent across Manitoba, Québec, and beyond.

This isn’t just about data centres. It’s about future-proofing Canada’s digital economy. By building high-performance AI infrastructure rooted in clean energy and regional partnerships, Bell is betting that AI innovation will not be bottlenecked by foreign dependencies or constrained by environmental concerns. Their exclusive partnership with Groq to provide low-latency, cost-effective inference hardware puts the spotlight on performance and access, while a direct relationship with Thompson Rivers University ensures that next-gen talent and research are deeply embedded from the start.

But why should this matter to forward-thinking leaders, strategists, and innovators?

First, because infrastructure always determines what is possible. We’re on the cusp of an AI revolution that won’t be driven by platform-level generalities, but by the organizations that can access, control, and scale AI in ways that serve their unique strategic goals. Sovereign infrastructure like Bell AI Fabric becomes a critical enabler—not just a utility, but a competitive lever. Businesses that tap into this ecosystem will be able to build faster, test locally, and deploy within the regulatory and privacy frameworks of Canada. That’s not just compliance—it’s strategic alignment.

Second, sustainability is no longer an optional feature of innovation. It’s a signal of durability. As AI workloads increasingly strain power grids and raise climate concerns, Bell’s clean energy-first approach positions it as a blueprint for responsible AI scale. Their clever reuse of waste heat to power university buildings is more than a tech story—it’s a systems story. It points to a future where compute infrastructure becomes integrated, community-serving, and regenerative. For strategy leaders, this suggests a new type of innovation geography—one where business, education, and infrastructure co-evolve.

Finally, let’s put the futurist lens on this. Bell’s move reframes Canada’s role in the global AI economy. This isn’t just catch-up; it’s a potential leapfrog. If Canada can build high-performance, low-cost, sovereign AI infrastructure now, it might just be in a position to shape the rules, ethics, and business models of the next AI wave. We’ve talked a lot about AI agents, LLMs, and vertical stacks—but none of that matters without robust, secure, performant infrastructure. This announcement shifts the conversation from AI hype to AI readiness. And readiness, when linked to national infrastructure and visionary design, is what drives long-term competitive advantage.

This isn’t just Bell’s announcement. It’s a moment for every leader in Canada to ask: are we building for today’s tools, or are we preparing our systems to shape the future?

In my most recent Shape of Tomorrow podcast, I explored a shift that’s not yet splashed across headlines but is already shaping our daily experience—the slow, quiet death of the app and the steady dissolution of platforms as we’ve known them. It’s not speculative. It’s not futuristic. It’s happening right now. But in this article, I want to push the conversation further. Because if you’re a leader building technology, shaping user experiences, or guiding strategy, there are urgent implications to unpack.

Apps used to be endpoints. Little icons we tapped to unlock function. That model is fading. Today, we are moving into a world where software is no longer a destination. It’s a presence. The difference isn’t just technical. It’s philosophical. Instead of designing things that users open, we must now ask how we create systems that live alongside them. Seamless, anticipatory, context-aware. We’re evolving from icons to outcomes. And if your product, your service, or your team is still optimizing for interaction rather than intention, you’re already out of step.

Look closely and you’ll see this shift everywhere. You use Slack to collaborate, but you might never open Google Docs separately to comment on a document. You ask Siri to book a table and it confirms the reservation by quietly working with OpenTable and your calendar without fanfare. Your AI-enhanced maps suggest your Saturday errands before you type a word. These are not just smarter features. They’re early evidence that the interface is dissolving. The app no longer needs to be the container. What matters is the orchestration, not the click.

Too many organizations, however, are stuck in an old model. They track metrics like time spent in app, daily active users, or screen flow. They polish UX and refine design systems while missing the more urgent transformation. Value is migrating from interface to intelligence. Time spent is no longer a sign of success. It’s a sign of friction. The real win today is time saved. Seamless resolution. Invisible enablement.

There are concrete steps to take. Start by shifting how success is measured. Move away from engagement as your guiding metric. Build KPIs around usefulness and seamlessness. Rethink product development as a service-layer strategy rather than a visual one. This means investing in systems, APIs, and data orchestration, not just front-end polish. Your app doesn’t need to be opened. It needs to be accessed by the systems your users already live inside.

Most importantly, begin designing for ambient intelligence. Ask whether your user experiences live in the interface or in context, preference, memory, and flow. Think less about where users are and more about who they are, what they need, and when. Intelligence is not about input. It’s about awareness. This is the leap that will define the next generation of product and strategy leadership.

I believe we are entering a phase that will make super apps irrelevant. They were one attempt to consolidate experiences into a single container. But containers are giving way to connective tissue. The winners in this next phase will not be those who control the destination. They will be those who own the journey. Not those who monopolize screen space. They will be those who inhabit decision space. Agents, not interfaces. Super context, not super app.

The questions we must ask ourselves now are more strategic than technical. Are we designing systems that our users visit, or ones that move with them? Are we spending our time refining buttons, or building orchestration? And is our competitive edge embedded in the user interface or in the unseen intelligence that powers outcomes?

If your AI strategy culminates in a chatbot or an app, you’re missing the point. Your AI must roam. It must anticipate. It must act on behalf of users without being summoned. This is not a nice-to-have. It is the new battleground, especially for those building in enterprise, in consumer tech, or in public systems.

As I said in the podcast, the transformation underway is not just technical. It is poetic. We are leaving behind a world of commands and icons, and entering one where the best software simply… shows up. Quietly. Thoughtfully. Helpfully. It doesn’t ask for attention. It earns trust by disappearing into the background and delivering value when and where it’s needed.

If that sounds like the future, it’s because it is. But it’s also the present. It just hasn’t been evenly implemented. And that’s where you come in. The end of the app is not the end of design. It’s the beginning of designing for something far more powerful: intention.