Best of LinkedIn: Smart Manufacturing CW 41/ 42

00:00:00: Welcome back to the deep dive.

00:00:02: If you need a shortcut to understanding where smart manufacturing is actually investing its time, energy, and, well, most importantly, its capital right now, this is probably it.

00:00:10: We've spent the last couple weeks really deep diving into the most influential conversations happening over on LinkedIn, pulling out the critical trends, the ones that are really moving the needle for industry professionals like you.

00:00:21: Before we jump in, just a reminder.

00:00:44: Yeah, and looking across all the sources, all the posts we reviewed, a few distinct patterns really do jump out.

00:00:51: It feels like we're clearly past that you know, purely experimental phase, the whole conversation seems to have shifted.

00:00:57: It's less about theoretical pilots now and more about pragmatic AI use cases, stuff that directly impacts the bottom line in your P&L.

00:01:05: We're also seeing, I think, a necessary tightening of human machine workflows, and those ecosystem plays the platforms, the standards.

00:01:13: they seem to be finally maturing.

00:01:14: Absolutely.

00:01:15: I agree.

00:01:15: It feels like we're done debating if this tech is coming down.

00:01:18: The only real question is, how do we actually make money with it?

00:01:22: and how do we keep it secure?

00:01:23: So for our deep dive today, we're going to unpack these key themes.

00:01:26: We'll start with AI and analytics, then move into automation, then operations, and wrap up with that really crucial topic, the human element.

00:01:34: Okay, let's unpack this then.

00:01:35: Right, let's definitely start with AI and analytics, because this is the theme where we saw the most Well, definitive evidence that manufacturers are really focused now on scaling for strategic advantage.

00:01:45: Yeah, the sources this week really show that executives, particularly in regions like the Middle East and Turkey, they aren't just looking for one-off AI projects anymore.

00:01:53: They are demanding actual P&L impact.

00:01:56: Uh, Mohana Nablusi observed, they're achieving this by scaling by pattern, not just project by project.

00:02:02: That phrase, scaling by pattern, I think that's critical.

00:02:06: It means... Basically, when they find a successful AI deployment like, say, predictive quality control on one machine, they aren't completely re-engineering the solution for the next machine.

00:02:17: No, they just clone that successful configuration across similar processes, maybe globally.

00:02:23: It's an approach built for rapid, consistent

00:02:26: ROI.

00:02:27: I love that idea.

00:02:28: But scaling like that, it requires the data to be clean and accessible first, doesn't it?

00:02:32: Oh, absolutely.

00:02:33: Rory de Gaulle's stressed that successful AI must always be company specific, and it has to start with robust data governance.

00:02:39: You just can't point some fancy AI tool at like unmanaged internal files and expect reliable results.

00:02:45: You need that accessible quality data foundation.

00:02:48: Precisely.

00:02:48: It's still garbage in, garbage out, even with the smartest algorithms.

00:02:52: And speaking of advanced algorithms, we did see some fascinating, uh, pragmatic applications that leverage existing tacit knowledge.

00:03:00: You did.

00:03:00: Yeah.

00:03:01: Meton Kaplan highlighted a great use case, video to instruction AI.

00:03:04: Right.

00:03:04: This is where an AI literally observes an expert performing a complex task, maybe like a specific maintenance procedure on a unique machine.

00:03:12: Right.

00:03:13: And it automatically translates that, you know, tacit on the floor know-how into structured electronic work instructions, EWIs.

00:03:20: That is

00:03:21: huge.

00:03:21: Just think about speeding up knowledge transfer and codifying standard operating procedures, especially in industries like with dynamic changes or high turnover, maybe like pharma.

00:03:31: It just accelerates that time from the expert knows it to, okay, now everyone can follow it.

00:03:36: much faster.

00:03:37: And we're also seeing where the major platform players are heading next with this.

00:03:41: Siemens, for instance, is clearly pushing the envelope.

00:03:43: They're working with generative and egenic AI in engineering processes.

00:03:49: Yeah, Rainer-Brem mentioned their ambition and it's pretty high.

00:03:52: They want to unlock productivity gains potentially over fifty percent through what they're calling agentic AI.

00:03:58: Fifty percent is a massive number.

00:04:00: Maybe quickly define agentic AI.

00:04:02: Sure.

00:04:02: It's essentially AI that's designed to autonomously execute and solve complex multi-step industrial problems.

00:04:10: So without constant human interaction needed, think of it maybe like a virtual self-directing engineer.

00:04:15: Okay,

00:04:16: got it.

00:04:16: And we're seeing that show up in actual product releases already.

00:04:19: Yeah.

00:04:19: Jethyn Tellis highlighted Siemens new AI-powered co-pilot.

00:04:23: Apparently, it specifically accelerates engineering design by simplifying complex simulation processes.

00:04:29: So it kind of offloads the heavy computational work from the engineer speeding up design iterations.

00:04:34: Right.

00:04:34: And tying all this AI capability together, Iora Berry made a really critical point.

00:04:39: Your manufacturing product data foundation is your strategic advantage.

00:04:43: That governed managed data is what gives the AI the essential context it needs.

00:04:48: Without that context, these new agents and co-pilots are, well, they're just playing guessing games, really.

00:04:53: Okay.

00:04:54: Let's shift gears now.

00:04:55: Let's move into the physical world.

00:04:56: Talk about automation and robotics.

00:04:59: Because, you know, here's where it gets really interesting.

00:05:01: This theme felt heavily driven by almost competitive anxiety these past two weeks.

00:05:07: It's true.

00:05:07: Brian Cooney noted a significant, almost alarming trend.

00:05:11: Apparently, Western executives visiting China are coming back reportedly terrified.

00:05:17: terrified by the speed and the sheer scale of China's high-tech industrial transformation, especially in robotics and AI deployment.

00:05:24: This isn't just about labor costs anymore.

00:05:26: It seems to be about a vast, rapidly scaling technological lead.

00:05:30: Wow, and that urgency, it seems like it's translating into accelerated investment, right?

00:05:35: Right.

00:05:35: Maybe a shift in technical approach towards something being called physical AI.

00:05:39: Exactly.

00:05:40: Demetrios Spiliopoulos and Joe Bowman have been tracking this.

00:05:43: Demetrios pointed to some recent major IoT acquisitions.

00:05:46: specifically soft bank buying ABV Robotics as a signal of rising confidence in this area.

00:05:52: Physical AI, it's the convergence of advanced AI with physical systems.

00:05:56: The intelligence moves out of the cloud and into the machine itself, making devices autonomously intelligent right at the edge.

00:06:04: So the robot doesn't just execute a pre-programmed move anymore, it actually adapts and responds intelligently to its immediate physical environment.

00:06:11: It's self-optimizing, basically.

00:06:13: That's the goal, yeah.

00:06:14: And we also saw concrete evidence of this strategic investment on the trade show floor.

00:06:25: The Speed, the Precision, the repeatability they offer, it feels like they're becoming non-negotiable competitive advantages.

00:06:31: And to give you a real-world example of that precision, Thomas Schmidberger showcased Schnefmann Machine & Bau.

00:06:38: Apparently, they designed a fully automated solution using KUK robots.

00:06:42: And it's for the incredibly complex assembly and testing of supercomputers for eighties.

00:06:49: Sensors and you know advanced driver assistance systems

00:06:51: right high-stakes stuff.

00:06:52: exactly when you're dealing with components for autonomous vehicles that Process stability and like micron level precision are just paramount.

00:07:00: absolutely and Recognizing that the speed of adoption matters here Dario Stajecic announced ABB's new customer portal robotics one.

00:07:09: This sounds like an effort to really streamline the user experience centralizing access to everything from the robot products themselves to spare parts even using a AI-supported search to simplify service requests.

00:07:21: They're trying to make the whole automation lifecycle easier.

00:07:23: Okay, so moving beyond the physical machinery itself, let's talk about the operational side.

00:07:27: Manufacturing, execution, connectivity and platforms, kind of the backbone of the smart factory.

00:07:32: Yeah, and operational excellence, this offers really tangible and compelling ROI right now.

00:07:37: Mahesh Adnani shared a fanpastic anecdote.

00:07:40: It was about a small focused team that achieved two hundred and forty seven manufacturing uptime and saved, get this, twenty five million dollars with a custom high availability system.

00:07:51: That result just shows what highly focused engineering can achieve in terms of, you know, system resilience.

00:07:55: Twenty-five

00:07:56: million, wow.

00:07:57: And on the software and process side, speed seems to translate directly into quality and safety too.

00:08:02: Manuel Costa shared an excellent example of how SAP Digital Manufacturing empowers a quality engineer named Nwemi to generate detailed traceability reports in minutes, not days.

00:08:13: Right.

00:08:13: Which allows them to pinpoint contaminated batches before a product ever shipped.

00:08:17: that massively mitigates risk.

00:08:18: That

00:08:19: ability to react instantly, it's priceless.

00:08:21: But, you know, to scale those wins, they uptime the quality speed.

00:08:23: you can't just rely on bespoke systems everywhere.

00:08:26: Joel Morrell emphasized this.

00:08:27: He said, a portfolio-based MES lifecycle framework is crucial.

00:08:30: You need structure to standardize rollouts, reduce redundancy, and deploy operational systems consistently across maybe dozens of plants.

00:08:37: Right, standardization.

00:08:38: And standardization requires communication, connectivity.

00:08:42: Tom Schneider noted that... interoperability, not just the technology itself, is the absolute prerequisite for turning all that data into actual value.

00:08:51: If your systems can't talk to each other, you've just created more digital islands, right?

00:08:54: Exactly.

00:08:55: And this is why open standards are gaining some serious momentum now.

00:08:58: Frank Seiferth highlighted how standards like OPCUA and UMATI guarantee that plug-and-play functionality and true interoperability.

00:09:05: And maybe more importantly, these standards form the essential basis for future shared digital spaces, things like manufacturing X. Without them, seamless data flow between companies or facilities is just, well, it's a pipe dream.

00:09:17: And the platform maturity seems to be accelerating.

00:09:19: to support this ecosystem need too.

00:09:22: Rajesh Ramachandran celebrated ABB's GenX platform being recognized as a leader in the twenty-twenty-five Gartner Magic Quadrant for global industrial IoT platforms.

00:09:31: He specifically noted its focus on AI-driven operations and, importantly, faster time to value for customers.

00:09:38: And that value, it's almost always magnified when the ecosystem collaborates effectively.

00:09:43: Yatsili described a really powerful example involving NTT data, SAP and data robot.

00:09:49: They combined like two decades of historical SAP data with real-time SCADA data to identify the compound factors causing a specific production waste event.

00:09:57: The result, millions in savings and substantial sustainability improvements.

00:10:02: That's the real power of data convergence across systems.

00:10:04: Okay, finally.

00:10:05: We need to talk about the digital twin and maybe most importantly, the people who make all this work.

00:10:10: Daniel Kepper highlighted a BCG white paper confirming that this industrial metaverse convergence of digital twins, Gen AI and body AI, it's driving measurable impacts right now.

00:10:21: They cited specific examples of a thirty percent increase in planning efficiency and a forty percent reduction in quality costs.

00:10:27: Those are real numbers.

00:10:28: Yeah.

00:10:28: And Kevin O'Donovan confirmed that the whole metaverse idea, it never really died.

00:10:32: It just evolved.

00:10:33: It ditched the consumer hype and became the industrial metaverse.

00:10:36: Strictly focused now on physics based simulation for tangible business outcomes, bridging the virtual and real worlds of the factory floor.

00:10:44: But the core message this week.

00:10:46: It really felt like it consistently came back to the human element, didn't it?

00:10:50: It did.

00:10:51: Maggie Slow expressed that technology alone won't transform manufacturing.

00:10:55: people will.

00:10:56: Workforce involvement and capability development have to be central to any successful adoption strategy.

00:11:02: And that brings us kind of full circle back to the earlier discussion about agentic AI.

00:11:07: While the goal might be those fifty percent productivity gains through autonomy, Dr.

00:11:12: Isil Burkhun argues companies really must focus on augmenting existing workers, giving them superhuman capabilities, as she put it, rather than just replacement.

00:11:22: She cited successful implementations that started small and scaled, showing ROI within just six months.

00:11:27: That seems like the crucial balance, doesn't it?

00:11:29: Yes.

00:11:30: How do we invest in something like agentic AI without creating a massive skills gap or, you know, fostering fear of replacement.

00:11:37: We need to train people to work alongside these intelligent systems.

00:11:40: Absolutely.

00:11:41: And that talent development piece, it's already underway.

00:11:44: Victor M highlighted academic programs for the connected industry, four point zero.

00:11:47: They're actively developing the next generation of industrial professionals through postgraduate courses, covering essential topics like reinforcement learning, predictive maintenance, and digital twin development for complex industrial facilities.

00:11:59: The industry is recognizing that need for a structured talent pipeline.

00:12:03: So okay, boiling it all down.

00:12:05: What does this all mean from the last couple of weeks?

00:12:07: The conversation across LinkedIn seems to confirm the industry is transitioning into a more mature execution focused phase.

00:12:15: That means practical application of Gen AI for very defined problem solving a focus on foundational maturity things like data standards and MES scaling frameworks and Crucially making sure all this tech is built around empowering or maybe augmenting the human worker.

00:12:30: Yeah, it's really a focus on making AI reliable, making data interoperable, and making people more capable.

00:12:36: That feels like the core theme of these two weeks.

00:12:38: That is a phenomenal summary of what sounds like a very busy two weeks in the smart manufacturing world.

00:12:44: If you enjoyed this deep dive, new episodes drop every two weeks.

00:12:47: Also check out our other editions on digital construction and digital power tools.

00:12:51: Thank you for joining us for this deep dive.

00:12:53: And before you go, maybe consider this provocative thought.

00:12:55: It builds on that acceleration of AI into physical systems, something explored by Jared Coyle.

00:13:02: The key question we really need to address now isn't just about maximizing efficiency.

00:13:06: It's about the ethical and moral quagmires that get amplified when intelligence moves into physical autonomous machinery.

00:13:15: We really must proactively ask what physical limitations, what data privacy protections, and what precise levels of human intervention must we establish now for these autonomously intelligent machines before they become ubiquitous and frankly their actions become opaque.

00:13:29: Something to think about.