WEBVTT 00:00:00.000 --> 00:00:02.779 Welcome back to the Deep Dive. You know, everywhere 00:00:02.779 --> 00:00:05.400 you look right now, AI is just dominating the 00:00:05.400 --> 00:00:07.559 conversation. It really is. Hard to escape it. 00:00:07.740 --> 00:00:09.240 Yeah, we're kind of drowning in headlines, aren't 00:00:09.240 --> 00:00:12.039 we? Generative models, machine learning, these 00:00:12.039 --> 00:00:15.080 fully autonomous systems promising massive efficiencies. 00:00:15.160 --> 00:00:20.260 Right. But if you, our listener, are, let's say, 00:00:20.820 --> 00:00:23.429 deeply committed to lean methodology, If your 00:00:23.429 --> 00:00:25.750 whole operating system is built on things like 00:00:25.750 --> 00:00:29.109 PDCA cycles continuous improvement and crucially 00:00:29.109 --> 00:00:31.750 respect for people that foundational principle 00:00:31.750 --> 00:00:34.649 Exactly, then you are likely asking a really 00:00:34.649 --> 00:00:37.630 critical question How does this you know cutting 00:00:37.630 --> 00:00:40.969 -edge tech actually integrate is AI meant to 00:00:40.969 --> 00:00:43.429 replace lean? Or is it just another tool in the 00:00:43.429 --> 00:00:45.229 toolbox? Well, that integration is precisely 00:00:45.229 --> 00:00:47.049 what we want to focus on today. Our sources, 00:00:47.070 --> 00:00:49.350 they give a pretty definitive answer here. OK. 00:00:49.549 --> 00:00:52.570 AI is not some magic wand. It won't fix a lean 00:00:52.570 --> 00:00:54.329 transformation that's already struggling. It 00:00:54.329 --> 00:00:56.810 definitely can't compensate for, say, a lack 00:00:56.810 --> 00:00:59.390 of leadership or if there's no real problem -solving 00:00:59.390 --> 00:01:02.270 culture. OK, but, and this sounds like the important 00:01:02.270 --> 00:01:05.510 distinction, it can absolutely supercharge a 00:01:05.510 --> 00:01:07.609 lean transformation that's already got some momentum. 00:01:07.670 --> 00:01:09.469 That is it, exactly. Supercharge is the right 00:01:09.469 --> 00:01:12.840 word. So, our mission in this deep dive is really 00:01:12.840 --> 00:01:16.060 to show you how AI can act more like a strategic 00:01:16.060 --> 00:01:18.640 thought partner, how it accelerates continuous 00:01:18.640 --> 00:01:21.700 improvement, and, well... enables leaders to 00:01:21.700 --> 00:01:24.599 be better lean practitioners. Yes, the core idea 00:01:24.599 --> 00:01:27.340 really is that AI's biggest strength, especially 00:01:27.340 --> 00:01:30.060 when you pair it with lean, is its ability to 00:01:30.060 --> 00:01:34.079 just cut through massive amounts of organizational 00:01:34.079 --> 00:01:36.719 friction, all that information noise. Okay, so 00:01:36.719 --> 00:01:38.340 it clears the decks. It clears the decks, yeah. 00:01:38.819 --> 00:01:41.019 So human leaders can actually focus on the truly 00:01:41.019 --> 00:01:44.849 value -added stuff. coaching, strategy, getting 00:01:44.849 --> 00:01:47.329 to the root cause of problems. That makes perfect 00:01:47.329 --> 00:01:50.349 sense. Let AI handle the complexities so people 00:01:50.349 --> 00:01:52.530 can handle the creativity and the critical thinking. 00:01:52.829 --> 00:01:55.010 Precisely. We absolutely have to lead with that 00:01:55.010 --> 00:01:56.849 foundational warning though, the one the sources 00:01:56.849 --> 00:02:00.030 really stress. Oh, absolutely critical. AI integration 00:02:00.030 --> 00:02:03.950 will fail like... immediately, without that bedrock 00:02:03.950 --> 00:02:06.769 of respect for people. And you need that deep 00:02:06.769 --> 00:02:08.909 -seated culture where teams are empowered, where 00:02:08.909 --> 00:02:11.389 they're expected to actually engage with the 00:02:11.389 --> 00:02:13.270 data. Yeah, if the culture isn't right, you're 00:02:13.270 --> 00:02:16.530 basically just automating chaos. It's garbage 00:02:16.530 --> 00:02:19.030 in, gospel out, unless people are actively questioning 00:02:19.030 --> 00:02:21.509 and challenging what the AI is producing. Right. 00:02:21.969 --> 00:02:24.490 But for anyone still maybe a bit skeptical about 00:02:24.490 --> 00:02:27.469 marrying this, you know, old, lean discipline 00:02:27.469 --> 00:02:31.159 with new AI tech, We already see proof points, 00:02:31.199 --> 00:02:33.319 don't we? We do. Look at Toyota. They rolled 00:02:33.319 --> 00:02:36.280 out an internal AI platform, analyzed their processes. 00:02:36.580 --> 00:02:39.419 And found what? Found thousands of hours of non 00:02:39.419 --> 00:02:42.819 -value -added work. Just gone. Mostly manual 00:02:42.819 --> 00:02:45.379 data crunching and analysis that teams were bogged 00:02:45.379 --> 00:02:47.400 down in. Wow. I can just picture the relief. 00:02:47.629 --> 00:02:49.389 teams getting that time back. It's not about 00:02:49.389 --> 00:02:51.430 cutting heads, it's about freeing up capacity, 00:02:51.490 --> 00:02:53.610 right? Exactly. So those teams could speed up 00:02:53.610 --> 00:02:56.449 their PDCA cycles, dedicate real time to meaningful 00:02:56.449 --> 00:02:59.590 casein. That's the goal. And GE appliances, you 00:02:59.590 --> 00:03:02.129 mentioned them too. Another example. Yeah, similar 00:03:02.129 --> 00:03:05.270 story. They integrated robotics with AI -driven 00:03:05.270 --> 00:03:07.639 metrology. And again, it wasn't just about going 00:03:07.639 --> 00:03:10.020 faster. What was the benefit then? It improved 00:03:10.020 --> 00:03:13.000 the consistency of flow, the accuracy of assembly, 00:03:13.340 --> 00:03:15.800 and importantly, enhanced safety, letting AI 00:03:15.800 --> 00:03:18.879 handle tricky or repetitive measurements. So 00:03:18.879 --> 00:03:20.620 the organizations that are winning aren't asking 00:03:20.620 --> 00:03:23.360 AI or Lean. No, they're designing systems where 00:03:23.360 --> 00:03:26.919 AI enables Lean principles, makes them work even 00:03:26.919 --> 00:03:29.639 better than before. That distinction, AI as an 00:03:29.639 --> 00:03:32.780 enabler, that feels really key. OK, let's dive 00:03:32.780 --> 00:03:36.250 into our first section then. Strategic leadership. 00:03:37.229 --> 00:03:39.430 Strategy deployment. Hoshinkanri, it's often 00:03:39.430 --> 00:03:42.229 the toughest part, isn't it? Slow, really human 00:03:42.229 --> 00:03:44.710 intensive. It is. And what's really fascinating 00:03:44.710 --> 00:03:47.030 here is how some thinkers like Jeff Woods and 00:03:47.030 --> 00:03:49.870 the AI -driven leader are shifting the perspective. 00:03:50.240 --> 00:03:53.199 AI isn't just an operational tool down on the 00:03:53.199 --> 00:03:55.659 floor. It's more. It becomes a genuine co -pilot, 00:03:56.099 --> 00:03:58.360 a thinking partner for strategy, for C -suite 00:03:58.360 --> 00:04:00.599 decisions. OK, so let's unpack that. The first 00:04:00.599 --> 00:04:03.139 concept you mentioned is AI personas. I find 00:04:03.139 --> 00:04:05.240 this idea really compelling. Seems like a powerful 00:04:05.240 --> 00:04:08.539 way to break down groupthink. It really is. Groupthink 00:04:08.539 --> 00:04:10.900 can kill strategy before it even gets off the 00:04:10.900 --> 00:04:14.819 ground, right? AI personas accelerate that perspective 00:04:14.819 --> 00:04:17.639 gathering at a strategic level dramatically. 00:04:17.699 --> 00:04:19.680 How does it work in practice? Give us an example. 00:04:19.800 --> 00:04:22.399 OK, imagine your leadership team is looking at 00:04:22.399 --> 00:04:25.420 a major change. Maybe altering your global value 00:04:25.420 --> 00:04:28.699 stream. Perhaps automating a key quality inspection 00:04:28.699 --> 00:04:31.800 point. Big decision. Huge capital cost. Lots 00:04:31.800 --> 00:04:34.019 of implications. So how do we use the personas? 00:04:34.180 --> 00:04:37.420 You prompt the AI, tell it to review that proposal 00:04:37.420 --> 00:04:40.339 through, say, four specific, often conflicting 00:04:40.339 --> 00:04:43.240 lenses. First, run it by the customer persona. 00:04:43.439 --> 00:04:46.529 What's the risk to quality? Lead time. the customer 00:04:46.529 --> 00:04:49.170 experience. Okay, pure customer focus, then. 00:04:49.269 --> 00:04:51.370 Then switch hats. Run it through the finance 00:04:51.370 --> 00:04:53.990 leader persona. Purely ROI, capital cost, depreciation 00:04:53.990 --> 00:04:56.089 schedules, the hard numbers. Makes sense. And 00:04:56.089 --> 00:04:57.990 you could add others like a supply chain leader 00:04:57.990 --> 00:05:00.290 persona focusing on constraints, bottlenecks, 00:05:00.490 --> 00:05:02.970 risk. Exactly. Material constraints, capacity, 00:05:03.149 --> 00:05:05.870 resilience, and maybe top it off with a board 00:05:05.870 --> 00:05:08.790 member persona looking at long term market position 00:05:08.790 --> 00:05:12.449 growth strategy. So the AI rapidly surfaces potential 00:05:12.449 --> 00:05:14.589 conflicts or risks from all these different angles 00:05:14.589 --> 00:05:17.470 before the humans even get in a room. Precisely. 00:05:17.550 --> 00:05:20.579 You get this rapid deep insight. forces the human 00:05:20.579 --> 00:05:22.980 team to address those friction points up front. 00:05:23.540 --> 00:05:25.620 You don't wait for the actual finance leader 00:05:25.620 --> 00:05:27.779 to poke holes in the plan weeks later. The AI 00:05:27.779 --> 00:05:29.519 has already done it. Leading to sharper questions, 00:05:29.720 --> 00:05:32.160 better preparation, a much higher quality decision 00:05:32.160 --> 00:05:34.560 process from the start. Absolutely, better preparation 00:05:34.560 --> 00:05:37.279 before you even engage the full human team. That 00:05:37.279 --> 00:05:40.040 leads nicely into the next idea, digital catchball. 00:05:40.860 --> 00:05:43.699 Now in Lean, catchball is that vital back and 00:05:43.699 --> 00:05:46.279 forth dialogue for strategy deployment, right? 00:05:46.490 --> 00:05:49.470 Hushen Connery, making sure strategy flows down 00:05:49.470 --> 00:05:52.670 and ideas flow back up. Correct. It ensures alignment 00:05:52.670 --> 00:05:55.310 and refinement. AI enhances this in a couple 00:05:55.310 --> 00:05:58.029 of ways. First, it just obliterates the noise 00:05:58.029 --> 00:06:01.610 in the data gathering phase. How so? It can synthesize 00:06:01.610 --> 00:06:03.550 vast amounts of external intelligence market 00:06:03.550 --> 00:06:06.149 trends, competitor moves, geopolitical stuff, 00:06:06.589 --> 00:06:08.589 and structure it instantly. It can build you 00:06:08.589 --> 00:06:11.350 a comprehensive SWOT analysis or run a Porter's 00:06:11.350 --> 00:06:14.180 Five Forces review like immediately. OK, so the 00:06:14.180 --> 00:06:17.360 human leaders start that catchball dialogue grounded 00:06:17.360 --> 00:06:20.339 in solid synthesized data, not spending weeks 00:06:20.339 --> 00:06:22.800 just compiling reports. Exactly. Cuts out huge 00:06:22.800 --> 00:06:26.279 amounts of non -value added work. And then the 00:06:26.279 --> 00:06:28.540 AI can actually participate in the catchball 00:06:28.540 --> 00:06:31.920 itself. Participate? How? Well, it can challenge 00:06:31.920 --> 00:06:34.800 a proposed objective based on the data it synthesized. 00:06:35.120 --> 00:06:37.540 Or, using natural language processing, it can 00:06:37.540 --> 00:06:39.480 almost interview the leader. What are your real 00:06:39.480 --> 00:06:41.970 goals here? What are the constraints? Then it 00:06:41.970 --> 00:06:44.250 can draft a summary action plan. Forcing clarity, 00:06:44.589 --> 00:06:47.009 closing the loop faster. Right. Keeping humans 00:06:47.009 --> 00:06:49.329 firmly in control, but accelerating the whole 00:06:49.329 --> 00:06:52.389 process. That really does minimize that non -value 00:06:52.389 --> 00:06:55.300 added. strategy churn. It frees up leaders to 00:06:55.300 --> 00:06:57.939 reinvest time where it matters most in lean coaching 00:06:57.939 --> 00:06:59.800 their teams. Which is the core activity, isn't 00:06:59.800 --> 00:07:02.000 it? Okay, let's pivot now. Hard turn to the GEMBA, 00:07:02.160 --> 00:07:04.800 the shop floor. What are the practical hands 00:07:04.800 --> 00:07:06.560 -on applications? Let's start with strengthening 00:07:06.560 --> 00:07:11.040 problem -solving. AI as a sort of sensei. An 00:07:11.040 --> 00:07:12.620 AI sensei. Okay, that one's interesting. How 00:07:12.620 --> 00:07:14.639 does that work? This is maybe the most direct 00:07:14.639 --> 00:07:17.620 application we're seeing. AI can act as this 00:07:17.620 --> 00:07:19.899 always available sensei. It dramatically improves 00:07:19.899 --> 00:07:22.480 the discipline, the rigor of the problem. solving 00:07:22.480 --> 00:07:25.959 process itself. I have to push back a bit here, 00:07:26.100 --> 00:07:29.620 though. If teams have an AI reviewing their work, 00:07:30.000 --> 00:07:32.519 isn't there a risk they become dependent? Do 00:07:32.519 --> 00:07:34.720 they stop learning the critical thinking needed 00:07:34.720 --> 00:07:37.360 if the AI is always prompting them? That's a 00:07:37.360 --> 00:07:39.579 really crucial caution. And it brings us right 00:07:39.579 --> 00:07:42.480 back to that respect for people foundation. The 00:07:42.480 --> 00:07:45.339 goal has to be augmentation, not replacement. 00:07:45.740 --> 00:07:47.600 So the AI isn't solving the problem for them. 00:07:47.759 --> 00:07:51.259 No, absolutely not. It reviews things like A3s 00:07:51.259 --> 00:07:53.920 or 8D reports, specifically looking for gaps 00:07:53.920 --> 00:07:56.439 in logic, maybe missing evidence. It critiques 00:07:56.439 --> 00:07:59.339 the process of solving the problem. Ah, I see. 00:07:59.379 --> 00:08:02.019 So it enforces rigor. It might ask structured 00:08:02.019 --> 00:08:04.779 5 why questions, or help brainstorm potential 00:08:04.779 --> 00:08:08.319 causes using Ishikawa Categories, manpower, machines, 00:08:08.579 --> 00:08:11.300 methods. Exactly. But the team owns the analysis, 00:08:11.500 --> 00:08:13.300 the countermeasure selection, the implementation, 00:08:13.519 --> 00:08:16.000 the verification. The AI just helps ensure they've 00:08:16.000 --> 00:08:19.199 done their homework thoroughly. So, counter -intuitively 00:08:19.199 --> 00:08:22.459 perhaps, it actually strengthens the human analytical 00:08:22.459 --> 00:08:26.040 skill by demanding higher quality output. That's 00:08:26.040 --> 00:08:28.500 the idea. Better inputs, better questions leads 00:08:28.500 --> 00:08:31.139 to stronger human problem solvers. Okay, application 00:08:31.139 --> 00:08:34.120 two. Standard work creation and retrieval. Oh, 00:08:34.159 --> 00:08:36.980 this is a classic lean pain point. Capturing 00:08:36.980 --> 00:08:39.200 that tribal knowledge after a case in burst. 00:08:39.980 --> 00:08:42.600 Transcribing sticky notes in flick charts. It's 00:08:42.600 --> 00:08:45.169 pure waste. And it's a massive barrier to actually 00:08:45.169 --> 00:08:46.990 sustaining the games you make. This is where 00:08:46.990 --> 00:08:49.590 computer vision and AI capture tools can really 00:08:49.590 --> 00:08:52.269 shine. Like using your phone? Kind of. Imagine 00:08:52.269 --> 00:08:54.990 using a Google Lens type app. You snap a photo 00:08:54.990 --> 00:08:57.450 of the whiteboard flow chart or the cluster post 00:08:57.450 --> 00:08:59.629 -its from a brainstorm. And the AI understands 00:08:59.629 --> 00:09:01.889 it. It converts it instantly into searchable 00:09:01.889 --> 00:09:05.639 digital text. Documentation. No more manual transcription 00:09:05.639 --> 00:09:08.379 needed. That alone frees up huge amounts of team 00:09:08.379 --> 00:09:10.700 time for actual improvement work, not admin. 00:09:10.919 --> 00:09:13.320 Correct. And the AI can help format that raw 00:09:13.320 --> 00:09:16.019 content into clear, consistent, standard work 00:09:16.019 --> 00:09:18.299 instructions. Then you can link those instructions 00:09:18.299 --> 00:09:20.559 to a QR code placed right at the workstation. 00:09:21.019 --> 00:09:24.480 Instant access. Very slick. But the caveat remains, 00:09:24.700 --> 00:09:28.139 I assume. Absolutely mandatory. Because AI generated 00:09:28.139 --> 00:09:32.129 it, leaders must trust But verify. Always. Check 00:09:32.129 --> 00:09:35.350 everything for accuracy, for relevance. The human 00:09:35.350 --> 00:09:38.169 is still the ultimate quality gatekeeper. AI 00:09:38.169 --> 00:09:40.169 is the thought partner, not the final word. Got 00:09:40.169 --> 00:09:43.750 it. Application three. Daily management and metrics 00:09:43.750 --> 00:09:47.370 flow. AI pulling data automatically from MES, 00:09:47.529 --> 00:09:51.370 ERP quality systems, populating electronic boards. 00:09:52.129 --> 00:09:54.110 Sounds efficient. It sounds like efficiency heaven, 00:09:54.269 --> 00:09:56.429 doesn't it? But this is where we hit that really 00:09:56.429 --> 00:09:59.769 crucial paradox of automation in Lean. Paradox, 00:09:59.870 --> 00:10:02.100 what do you mean? Well, we have decades of lean 00:10:02.100 --> 00:10:04.159 practice and research showing that the physical 00:10:04.159 --> 00:10:06.360 -human interaction with visual controls is what 00:10:06.360 --> 00:10:08.179 drives ownership. Actually, writing the metric 00:10:08.179 --> 00:10:10.200 on the board, moving the magnet, updating the 00:10:10.200 --> 00:10:12.200 status. That physical act makes a difference. 00:10:12.440 --> 00:10:14.620 It connects the brain differently than just passively 00:10:14.620 --> 00:10:16.500 looking at a screen. There's research backing 00:10:16.500 --> 00:10:19.299 this up. Oh, yeah. Yeah, that 2014 study, Mueller 00:10:19.299 --> 00:10:22.179 and Oppenheimer. The pen is mightier than the 00:10:22.179 --> 00:10:25.340 keyboard. It showed students taking notes longhand. 00:10:25.580 --> 00:10:27.960 They engaged more deeply. They synthesized. They 00:10:27.960 --> 00:10:30.659 reframed. There is typing. Typing often led to 00:10:30.659 --> 00:10:33.379 just verbatim transcription, shallower processing, 00:10:33.940 --> 00:10:36.659 less real learning. Wow. Okay, so if the metrics 00:10:36.659 --> 00:10:39.159 board is just automatically populated by AI, 00:10:39.799 --> 00:10:42.919 the risk is it becomes just visual noise. Data 00:10:42.919 --> 00:10:45.200 that's done to the team, not by the team. You 00:10:45.200 --> 00:10:47.669 nailed it. That's a powerful synthesis. The takeaway 00:10:47.669 --> 00:10:50.529 isn't no screens allowed. It's that if you automate 00:10:50.529 --> 00:10:53.529 the data pool, you absolutely must redesign the 00:10:53.529 --> 00:10:56.129 process to include a human synthesis step. Meaning? 00:10:56.470 --> 00:10:59.070 The team has to actively engage with that automated 00:10:59.070 --> 00:11:02.250 data, discuss the trends, plot the countermeasures 00:11:02.250 --> 00:11:04.669 together, make sure the information flow drives 00:11:04.669 --> 00:11:07.730 action, not just passive observation. Keep the 00:11:07.730 --> 00:11:09.769 human thinking at the center. Okay. Application 00:11:09.769 --> 00:11:14.730 four, quality at source. AI enabling a sort of 00:11:14.730 --> 00:11:16.970 digital poke -o -yoke or mistake -proofing. Yeah, 00:11:17.049 --> 00:11:18.870 taking traditional mistake -proofing concepts 00:11:18.870 --> 00:11:20.750 and giving them a digital boost. We're seeing 00:11:20.750 --> 00:11:23.269 computer vision systems, for instance, that can 00:11:23.269 --> 00:11:26.049 spot minute abnormalities or defects in real 00:11:26.049 --> 00:11:28.590 time. Instant feedback. Instantly alerting the 00:11:28.590 --> 00:11:30.629 operator before a bad part moves downstream. 00:11:30.789 --> 00:11:32.870 That's the lean ideal, right? Catch it at the 00:11:32.870 --> 00:11:36.120 source. And beyond just the local station, AI 00:11:36.120 --> 00:11:38.740 can analyze defect patterns across thousands 00:11:38.740 --> 00:11:41.299 of units, multiple lines, even different factories, 00:11:41.879 --> 00:11:44.059 looking for correlations humans would never spot. 00:11:44.279 --> 00:11:46.600 So they can identify systemic issues. Exactly. 00:11:46.860 --> 00:11:48.840 And predictive models can then suggest design 00:11:48.840 --> 00:11:51.919 changes back to engineering, ways to actually 00:11:51.919 --> 00:11:54.700 engineer out the possibility of human error in 00:11:54.700 --> 00:11:56.799 the first place. That's the ultimate poke -yoke, 00:11:57.259 --> 00:11:59.980 designing the error out completely. That's powerful 00:11:59.980 --> 00:12:02.220 stuff. Okay, that covers the first four. Let's 00:12:02.220 --> 00:12:04.200 make sure we hit the final three. Application 00:12:04.200 --> 00:12:07.279 five, predictive maintenance within a TPM framework. 00:12:07.659 --> 00:12:10.019 Right, total productive maintenance. This is 00:12:10.019 --> 00:12:12.559 a huge win for stabilizing flow and reducing 00:12:12.559 --> 00:12:15.940 waste. TPM rightly emphasizes operator ownership 00:12:15.940 --> 00:12:19.000 for basic care. Cleaning, inspection, lubrication. 00:12:19.289 --> 00:12:23.309 The basics. Exactly. But AI is the ideal augmentation 00:12:23.309 --> 00:12:25.289 tool here. It integrates sensors, vibration, 00:12:25.750 --> 00:12:27.470 temperature, current draw, analyzing those signals 00:12:27.470 --> 00:12:29.990 constantly. Two hundred four seven. Looking for 00:12:29.990 --> 00:12:32.090 signs of trouble. Looking for the faint signals 00:12:32.090 --> 00:12:34.429 that predict an impending breakdown, often days 00:12:34.429 --> 00:12:37.230 or weeks before it happens. This lets maintenance 00:12:37.230 --> 00:12:40.350 teams shift entirely away from reactive firefighting. 00:12:40.610 --> 00:12:43.610 Which is pure waste waiting, unexpected downtime. 00:12:43.789 --> 00:12:47.139 To planned proactive interventions. fixing things 00:12:47.139 --> 00:12:49.539 before they break during planned downtime. So 00:12:49.539 --> 00:12:52.320 we gain uptime, obviously, but we're also respecting 00:12:52.320 --> 00:12:54.740 people, aren't we? Freeing up skilled maintenance 00:12:54.740 --> 00:12:56.980 techs from just running around putting out fires. 00:12:57.279 --> 00:12:59.700 Precisely. Their valuable time gets reinvested 00:12:59.700 --> 00:13:02.059 in improving maintenance processes, tackling 00:13:02.059 --> 00:13:04.840 root causes of failures, upgrading skills, true 00:13:04.840 --> 00:13:07.500 value -added work that creates stability. Makes 00:13:07.500 --> 00:13:10.679 total sense. Application 6. Smart scheduling 00:13:10.679 --> 00:13:14.080 and PFEP alignment. Plan for every part. The 00:13:14.080 --> 00:13:16.600 goal in lean scheduling is always hitting tack 00:13:16.600 --> 00:13:19.159 time, matching production to actual customer 00:13:19.159 --> 00:13:21.620 demand. And this is where the sheer computational 00:13:21.620 --> 00:13:24.440 horsepower of AI becomes almost indispensable, 00:13:24.779 --> 00:13:27.440 especially in complex environments. AI can help 00:13:27.440 --> 00:13:30.120 create really sophisticated data -driven production 00:13:30.120 --> 00:13:32.200 wheels or sequences. Factoring in things like 00:13:32.200 --> 00:13:34.600 changeover times, constraints. Exactly. Constraints, 00:13:34.799 --> 00:13:37.620 real -time customer order signals. And crucially, 00:13:38.179 --> 00:13:40.379 it integrates tightly with that PFEP thinking. 00:13:40.679 --> 00:13:43.220 Plan for every part that detailed map of of every 00:13:43.220 --> 00:13:46.240 component, it's demand, storage, replenishment. 00:13:46.519 --> 00:13:49.919 Right. AI takes that massive complex data set 00:13:49.919 --> 00:13:52.879 and helps recommend inventory policies, can -ban 00:13:52.879 --> 00:13:55.679 levels, pull triggers, things that minimize both 00:13:55.679 --> 00:13:58.320 the waste of overproduction and the risk of shortages. 00:13:58.600 --> 00:14:01.399 So it helps ensure the production sequence truly 00:14:01.399 --> 00:14:03.940 reflects what the customer is pulling now. Making 00:14:03.940 --> 00:14:06.480 the whole supply chain more responsive. It keeps 00:14:06.480 --> 00:14:08.659 the human planning team focused on improving 00:14:08.659 --> 00:14:11.059 the system's flow and responsiveness, not just 00:14:11.059 --> 00:14:13.299 drowning in spreadsheets. supporting TAC -driven 00:14:13.299 --> 00:14:18.090 execution. Finally, Application 7. using AI for 00:14:18.090 --> 00:14:20.909 a voice of customer, foosie, and getting deeper 00:14:20.909 --> 00:14:23.029 market insights. Yeah, this connects the GEMBA 00:14:23.029 --> 00:14:25.330 right back to the strategic Hoshin planning process. 00:14:25.590 --> 00:14:28.690 It closes the loop. How does AI help here specifically? 00:14:28.970 --> 00:14:31.850 Its strength is processing huge volumes of unstructured 00:14:31.850 --> 00:14:34.129 data, stuff that's basically meaningless if you 00:14:34.129 --> 00:14:36.190 just stare at it in a spreadsheet. Think customer 00:14:36.190 --> 00:14:38.769 feedback forms, warranty claims, social media 00:14:38.769 --> 00:14:41.210 comments, complaint logs. Okay, all that messy 00:14:41.210 --> 00:14:43.870 text -based stuff. AI can analyze all of it. 00:14:44.039 --> 00:14:47.299 Identify emerging themes, pain points, maybe 00:14:47.299 --> 00:14:49.379 needs that traditional surveys completely miss 00:14:49.379 --> 00:14:51.700 because they ask the wrong questions. Finding 00:14:51.700 --> 00:14:55.279 the unknown unknowns. Kind of, yeah. It clusters 00:14:55.279 --> 00:14:57.679 all that feedback into meaningful, actionable 00:14:57.679 --> 00:15:00.700 themes. It can quantify sentiment, spot trends 00:15:00.700 --> 00:15:04.039 early. This lets organizations connect real -world 00:15:04.039 --> 00:15:07.019 voce data directly back into their strategic 00:15:07.019 --> 00:15:10.019 objectives. Ensuring the Hushan plan isn't just 00:15:10.019 --> 00:15:12.480 internally focused, but actually aligned with 00:15:12.480 --> 00:15:15.899 what the market wants or needs. Reducing the 00:15:15.899 --> 00:15:18.919 waste of making things nobody values. Exactly. 00:15:18.960 --> 00:15:21.720 It grounds strategy and reality. This has been 00:15:21.720 --> 00:15:24.720 incredibly insightful. Really useful. We kick 00:15:24.720 --> 00:15:26.840 things off with that common fear, is AI going 00:15:26.840 --> 00:15:29.500 to replace lean? Right. The AI versus lean question. 00:15:29.700 --> 00:15:32.259 But the evidence and what you've outlined really 00:15:32.259 --> 00:15:34.720 shows that AI is best used to remove friction. 00:15:35.200 --> 00:15:37.120 Friction from strategy, from daily management, 00:15:37.200 --> 00:15:39.899 from problem solving, and maybe most importantly, 00:15:40.379 --> 00:15:42.179 friction from human thinking. That's a great 00:15:42.179 --> 00:15:43.539 way to put it. And if you look at this whole 00:15:43.539 --> 00:15:45.659 integration through a more philosophical lean 00:15:45.659 --> 00:15:47.940 lens, it really reminds me of the concept of 00:15:47.940 --> 00:15:50.879 Jidoka. Jidoka. Automation with a human touch. 00:15:51.200 --> 00:15:54.200 stopping the line for an abnormality. Exactly. 00:15:54.580 --> 00:15:56.840 It stops the process to highlight a problem, 00:15:57.019 --> 00:15:59.600 which then elevates the human role. It forces 00:15:59.600 --> 00:16:02.500 people to engage, analyze, find the root cause, 00:16:02.700 --> 00:16:05.159 and improve the system. So you're saying AI, 00:16:05.559 --> 00:16:08.340 when used correctly in this context, acts like 00:16:08.340 --> 00:16:11.500 a kind of digital judoka system, but maybe across 00:16:11.500 --> 00:16:13.559 the whole enterprise, not just the line. I think 00:16:13.559 --> 00:16:16.600 that's a powerful analogy, yes. AI shines that 00:16:16.600 --> 00:16:20.529 bright, unbiased light on ambiguity on complexity, 00:16:20.889 --> 00:16:23.090 whether it's a flaw in a strategic plan highlighted 00:16:23.090 --> 00:16:26.289 by AI personas or an impending machine failure 00:16:26.289 --> 00:16:28.429 detected by sensors. It creates the signal. It 00:16:28.429 --> 00:16:30.570 creates the signal and it gives people the cognitive 00:16:30.570 --> 00:16:32.690 space, the bandwidth they need to actually engage 00:16:32.690 --> 00:16:35.269 deeply creatively and solve the underlying problem. 00:16:35.570 --> 00:16:38.070 You know, the heart of Lean is continuous improvement 00:16:38.070 --> 00:16:40.309 driven by people. So the final thought for our 00:16:40.309 --> 00:16:42.629 listener. The provocative question for you, the 00:16:42.629 --> 00:16:45.509 listener, is really this. How can you start leveraging 00:16:45.509 --> 00:16:48.269 AI not to replace your people, but to automate 00:16:48.269 --> 00:16:50.590 a way to tedious, the transactional, the noise 00:16:50.590 --> 00:16:53.269 in your organization? How can you use it specifically 00:16:53.269 --> 00:16:55.789 to elevate the human element, giving your teams 00:16:55.789 --> 00:16:57.889 the dedicated time and focus they truly need 00:16:57.889 --> 00:16:59.929 to become the innovative, relentless problem 00:16:59.929 --> 00:17:01.509 solvers that Lean demands?