As artificial intelligence reshapes industrial automation and drives manufacturers to find new ways to improve productivity and remain competitive, the Association for Advancing Automation (A3) at Automate 2026 is bringing together industry leaders to share their perspectives on the current state of automation, the biggest opportunities and challenges facing businesses today, and where the industry is headed.
Being held in Chicago at McCormick Place, the opening keynote on June 22 — The State of the Automation Industry: Leadership Roundtable — includes a panel of industry leaders exploring how manufacturers are integrating AI with automation systems to improve productivity, build more adaptive operations, and enable the workforce to scale new technologies responsibly. We spoke with one of the panelists, Matt Moschner, President and CEO of Cognex, about the most important trends shaping automation today.
Moschner oversees global engineering, products, sales, and operations at the company. Since joining Cognex in 2017, Moschner has held key positions of leadership across product and engineering teams, helping grow the company’s barcode reading portfolio and its strategic technology and product innovation process. Moschner holds a Bachelor of Science in Electrical Engineering & Economics from Duke University and a Master of Business Administration from Northwestern University’s Kellogg School of Management.
Tech Briefs: The leadership roundtable keynote at Automate 2026 this week focuses on the current state of automation. What are the biggest opportunities and challenges facing businesses today?
Matt Moschner: The automation landscape is shifting from rigid, rules-based programming to adaptive, AI-driven platforms. The biggest opportunity is flexibility. Production environments are becoming more complex — more SKUs, shorter run cycles, and more dynamic inputs — and AI allows systems to adapt to that reality rather than forcing operations into predefined workflows. But flexibility alone isn't enough if organizations aren't ready to act on it. The primary challenge is trust. Companies are being asked to rely on AI for decisions that directly affect quality, uptime, and compliance — and many aren't yet structured to support that. We see it play out in misalignment between IT and operations, where rich data goes unused because there's no shared framework for acting on it. That tension — between what the technology makes possible and what organizations are ready to absorb — is what we'll be working through together over the next five years.
Tech Briefs: We’re hearing a lot about AI reshaping automation. Where are you seeing real, production-level impact today versus hype?
Moschner: The difference between hype and real impact is whether AI is performing reliably at production scale — and in industrial automation, it is. We're seeing measurable results in inspection and identification tasks that have historically been out of reach — too variable, too subtle, or too fast-moving for traditional approaches to handle consistently. Take defect detection: A system that used to require hundreds of labeled images to reach acceptable accuracy can now learn from dozens — and hold that accuracy at full line speed, in variable lighting, with product variations that would have tripped up earlier approaches. That's not a pilot result; that's production. Full autonomy is still ahead, but manufacturers aren't waiting for it. They're deploying AI today in targeted, well-defined roles — augmenting human decision-making and expanding scope as confidence builds.
Tech Briefs: Cognex has been developing AI-powered machine vision tools. What’s the biggest breakthrough in the last few years that’s actually changed what vision systems can do on the factory floor?
Moschner: The most significant breakthrough has been the shift from programming vision systems to training them by example. Instead of trying to anticipate every acceptable condition in advance — every scratch pattern, every weld variation, every label placement — manufacturers can now train systems on real production images and let the model learn what good looks like. That change alone unlocked a category of inspection tasks that were simply too variable to define by hand. At Cognex, that meant rethinking not just the algorithms but the entire deployment model — an edge-to-cloud architecture that makes AI practical at production scale, not just impressive in the lab. The result is a fundamental shift from systems executing predefined checklists to systems that can perceive, interpret, and judge like a skilled human inspector but with a level of consistency and scale no human can match.
Tech Briefs: What types of inspection or perception tasks were historically too complex for machine vision but are now becoming feasible with AI?
Moschner: AI is enabling automation of inspection tasks that were previously impractical — and the categories are broader than most people expect. The first is subtle, hard-to-define defects. Surface scratches, cosmetic inconsistencies, structural irregularities — these don't follow explicit rules, which is why they historically required human judgment. AI can now detect them with greater consistency than a human inspector, across an entire shift, without fatigue. The second is extreme diversity. Vision systems in logistics encounter an enormous range of shapes, sizes, and packaging conditions — but diversity shows up in unexpected places too. We have a customer in Asia using our AI vision tools to inspect shrimp deheading on a processing line. Every shrimp is different. That's exactly the kind of task that was simply out of reach until now. The third is generalization — deploying across new products and lines without rebuilding from scratch. That's what moves AI vision from isolated use cases to enterprise-wide automation.
Tech Briefs: Looking ahead five years, how do you see machine vision evolving alongside robotics and physical AI? And does it become the key enabler for more autonomous systems?
Moschner: Over the next five years, machine vision will evolve from a point solution into a continuous intelligence layer across the factory. As robotics and physical AI systems become more capable — machines that don't just execute instructions but sense, adapt, and act in the physical world — vision will become the critical link between perception and action. That's why the "eyes of automation" framing is increasingly incomplete. Modern vision systems aren't just capturing images — they're making decisions at the edge in real time, and through cloud-connected platforms, sharing those learnings across lines and sites. Think of it less as a camera and more as the nervous system of the factory, linking perception to action, and detection to prevention. Autonomy will be phased. But vision is what makes physical AI real — not only as a peripheral sensor, but as the layer that turns digital intelligence into physical action.
Tech Briefs: For manufacturers or operations leaders today, what’s the one strategic move they should be making now to stay competitive in an AI-driven automation landscape?
Moschner: The single most important move manufacturers can make today is to build organizational confidence in AI — not just technical capability, but the institutional trust that allows AI to operate where it matters most: in production, at scale, with real consequences. That starts with discipline: use real production data, validate rigorously, and involve operators from the beginning so the system earns its place in the workflow rather than being imposed on it. Then scale within a consistent platform — one that lets what works on one line get replicated across an entire operation without rebuilding from scratch. The companies winning are not only experimenting with AI but operationalizing it — turning localized success into repeatable processes and durable advantage.
Automate Booth #3101
This article was written by Chitra Sethi, Editorial Director, SAE Media Group. For more information, visit www.cognex.com/en and to explore more sessions at Automate 2026, see full agenda .
