ACE Panel Discussion: How AI Is Transforming the Shopfloor

On 25 August 2026, Gradion hosted a panel discussion at ACE that brought manufacturing leaders together to answer a question most companies are still working through: where does AI actually stand on the shop floor today?

Three voices from the industry joined the stage: Mr. Dung Nguyen - Director of Technology and Innovation, Ms. Phuong Bui - Country Manager from INFOR, and Mr Hung Nguyen - Chief Operations Officers from Viet Dynamic to share insights about the global AI adoption curve, what blocks deployment at the machinery level, and what a production-ready AI system actually delivers. Just a direct conversation grounded in what's actually happening in factories right now.

The session opened by zooming out. Mr. Dung Nguyen, Director of Technology and Innovation at Gradion, framed the bigger picture before the conversation moved to the shop floor. The landscape is clear: AI adoption in office and software work is no longer an experiment but it is used daily. End users and developers have moved ahead of the institutions they work for, adopting AI bottom-up, faster than most companies are governing it.

Manufacturing sits behind that curve, but it's moving. The reasons it lagged are the structures: Manufacturing is regulated, physical, safety-critical, and expensive to get wrong. What's changed is the governance side of AI. As safety and oversight mechanisms mature, tightly regulated segments like manufacturing can now seriously consider adoption.

The session also named the real challenge directly: these are transformation projects, not installations. They require training, upskilling, and people who can actually work with AI. Having AI is not a purpose, AI has to serve one. The most common failure pattern discussed was a familiar one: a tool gets bought, nobody's job changes, and no number anyone is accountable for actually moves.

Three blockers came up as the ones that matter most:

• Strategy - a clear answer to what the AI is for, and which decision it improves.
• Safety and trust - governance and human oversight, which is not optional in a manufacturing environment.
• Data - the one blocker that cannot be shortcut. No usable data means no AI, however good the software is.

Mr. Hung Nguyen, COO at Viet Dynamic, brought the conversation down to the factory level - the state of the floor, not the theory.

The discussion covered how much of the equipment running in factories today is actually smart enough to feed into an AI-driven system, and whether market demand for that capability is real or still being educated into existence. One of the recurring insights throughout the session is that hardware and data readiness don't always arrive together. When hardware comes before a plan for the data it produces, companies risk losing the very value the automation was meant to create.

One sharper exchange of the evening tackled a direct question: is AI reasoning across a business (top floor) and AI making decisions in real time at the machine level (shop floor) really two separate journeys - or one? The answer that emerged: they're distinct today, but the connection between them is where most projects actually stall. Not because the AI is wrong, but because nobody owns the handshake between the two layers.

Ms. Phuong Bui, Country Manager Vietnam at Infor, closed the product side of the conversation - going past automation into what AI in ERP and WMS is actually capable of today.

The discussion addressed some of the sharper questions leaders are asking right now: can AI actually make decisions in ERP and WMS, or is it still decision support? How should companies measure AI ROI? And as systems become more autonomous, how do businesses maintain human oversight and governance? The session was direct about the fact that these aren't hypothetical concerns - they the questions companies should be asking before they buy.

Across all three perspectives, one thread tied the evening together: the pace of a company's AI journey is rarely set by technology. It's set by the strength of five factors - budget, machine age, data quality, leadership, and people and skills.

What sets apart a company that sees real value from AI within twelve months from one still stuck in pilot mode after the same amount of time? It's not about which vendor they picked. It comes down to two things: Did leadership set a clear strategy and governance plan early on, and was the data actually usable from the start of the project. 

The panel closed with an open floor, where attendees pushed the speakers further on what companies get wrong when hardware comes before data, and where the connection between the shop floor and business leadership actually breaks down. These are the questions worth asking before any deployment decision. The evening wrapped with a tour of ACE, where guests walked the floor alongside our engineers - seeing the systems discussed on stage running in practice.

Thank you to our panelist - Mr. Dung Nguyen (Gradion), Ms. Phuong Bui (Infor), and Mr. Hung Nguyen (Viet Dynamics) - for the depth, preparation, and candor they brought to the stage. And thank you to everyone who joined us at ACE and stayed engaged through to the last question. This is the kind of conversation we intend to keep having. If you missed this session, stay close - more ACE sessions are coming.