1 00:00:09,900 --> 00:00:16,039 Hello, everyone. Welcome to another installment of MHI Supply Chain tech talk. 2 00:00:16,039 --> 00:00:20,100 We're looking forward to starting this live panel by discussing an emerging topic, 3 00:00:20,100 --> 00:00:23,300 how AI is empowering warehouse automation. 4 00:00:23,300 --> 00:00:26,479 I'm Bruce Muir, Logistics Industry Manager for SICK, 5 00:00:26,479 --> 00:00:30,160 and I have the pleasure of being joined by three of my esteemed colleagues, 6 00:00:30,160 --> 00:00:34,559 Dave Gustafson, Nick Kern, and Divya Prakash. 7 00:00:34,559 --> 00:00:36,999 Dave is the digital Consulting Manager for 8 00:00:36,999 --> 00:00:40,099 mobile Automation and has 20+ years in the industry. 9 00:00:40,099 --> 00:00:43,880 I'm sure he'll have lots to share on this panel topic. 10 00:00:43,880 --> 00:00:49,640 H, good morning. Thanks, Bruce for inviting me and thanks MHI for hosting the event. 11 00:00:49,640 --> 00:00:52,180 Our next guest is Nick Kern, 12 00:00:52,180 --> 00:00:55,960 Systems application Engineer for Industrial Robotics, at SICK. 13 00:00:55,960 --> 00:00:58,860 I look forward to hearing about all the insights he can provide. 14 00:00:58,860 --> 00:01:01,299 Yeah, thanks, Bruce. I'm excited to be here. 15 00:01:01,299 --> 00:01:05,339 Should be interesting to learn more about AI and how it's affecting the warehouse. 16 00:01:05,339 --> 00:01:08,499 Excellent. Our last guest is Divya Prakash, 17 00:01:08,499 --> 00:01:11,600 Director of Business Consulting, Industry 4.0. 18 00:01:11,600 --> 00:01:14,340 If anyone is going to go to impart knowledge on 19 00:01:14,340 --> 00:01:18,439 innovation and impactful solutions in this industry, it's m. 20 00:01:18,439 --> 00:01:23,420 Thank you, Bruce. Glad to be here on 21 00:01:23,420 --> 00:01:28,319 the panel to learn and discuss about AI enabled automation. 22 00:01:28,319 --> 00:01:31,639 Alright, great. Gentlemen, thank you for joining me today to discuss 23 00:01:31,639 --> 00:01:34,659 how AI empowers warehouse automation in the supply chain. 24 00:01:34,659 --> 00:01:36,919 Before we start, I just want to remind 25 00:01:36,919 --> 00:01:40,860 the audience that a Q&A box is available beneath the video player. 26 00:01:40,860 --> 00:01:43,700 Feel free to submit your questions regarding our topic. 27 00:01:43,700 --> 00:01:45,580 We'll discuss them at the end of the talk. 28 00:01:45,580 --> 00:01:48,400 And if any questions don't get responded to, 29 00:01:48,400 --> 00:01:52,479 we'll take that directly to the person that asks. 30 00:01:52,479 --> 00:01:54,079 So, let's get started. 31 00:01:54,079 --> 00:01:56,939 In recent years, we've seen AI make its way into 32 00:01:56,939 --> 00:01:59,840 supply chain operations, specifically with automation. 33 00:01:59,840 --> 00:02:03,359 Divya, can you tell us how AI driven automation has transformed 34 00:02:03,359 --> 00:02:07,620 traditional warehouse operations in terms of efficiency and productivity? 35 00:02:07,620 --> 00:02:10,759 Sure, Bruce, Automation itself is 36 00:02:10,759 --> 00:02:16,500 do efficiency and productivity is always there. That's the main game. 37 00:02:16,500 --> 00:02:18,320 Now with new tools with AI, 38 00:02:18,320 --> 00:02:20,559 AI enabled automation, as we call it. 39 00:02:20,559 --> 00:02:23,580 We are definitely enhancing the efficiency and productivity. 40 00:02:23,580 --> 00:02:27,240 Now, when you talk about AI enabled tools, 41 00:02:27,240 --> 00:02:29,940 they are different tools and is pervasive 42 00:02:29,940 --> 00:02:33,539 across different areas of the warehouse and the whole supply chain. 43 00:02:33,539 --> 00:02:39,039 It is optimizing various aspects from inventory management to order full impairment, 44 00:02:39,039 --> 00:02:42,140 you know, different areas having seen the applications. 45 00:02:42,140 --> 00:02:46,139 Now, machine learning algorithms are improving the demand forecasting, 46 00:02:46,139 --> 00:02:49,765 allowing warehouses to adapt to changing customer needs. 47 00:02:49,765 --> 00:02:53,070 And then when you come down to computer vision technologies, 48 00:02:53,070 --> 00:02:55,509 those are enhancing the robotic visions for 49 00:02:55,509 --> 00:02:58,850 tasks like item recognition and quality control. 50 00:02:58,850 --> 00:03:00,890 So there's I mean pervasive. 51 00:03:00,890 --> 00:03:06,410 M AI tools are slowly making their way across the entire aros operations. 52 00:03:06,410 --> 00:03:09,369 There's certainly been some major transformation. 53 00:03:09,369 --> 00:03:15,149 Can you tell us more about AI about specific AI powered solutions from SICK? 54 00:03:15,149 --> 00:03:19,570 Sure. Sure. SICK, is a sensor solution provider, 55 00:03:19,570 --> 00:03:24,189 but we offer AI powered solutions that leverage the true power of automation, right? 56 00:03:24,189 --> 00:03:25,489 So what do I mean by that? 57 00:03:25,489 --> 00:03:29,349 A intelligence, which is very special to us. 58 00:03:29,349 --> 00:03:34,209 No. It represents the start of a new era of sensor intelligence. 59 00:03:34,209 --> 00:03:36,210 It makes it possible to solve 60 00:03:36,210 --> 00:03:40,390 any demanding tasks to quickly adapt to changing conditions, 61 00:03:40,390 --> 00:03:45,724 and to recognize patterns much more quickly easily and reliably than before. 62 00:03:45,724 --> 00:03:47,919 To capitalize on these benefits, 63 00:03:47,919 --> 00:03:50,480 our sensors, our smart sensors, 64 00:03:50,480 --> 00:03:51,980 collect the valuable data, 65 00:03:51,980 --> 00:03:54,719 which is then interpreted by our algorithms, 66 00:03:54,719 --> 00:03:56,940 so you can focus on the big picture, 67 00:03:56,940 --> 00:04:01,280 optimize your workflows, and make efficient use of your resources. 68 00:04:01,280 --> 00:04:03,419 Now, AI is a broad topic, 69 00:04:03,419 --> 00:04:05,539 that's got different subsets within it. 70 00:04:05,539 --> 00:04:07,500 More specifically, for example, 71 00:04:07,500 --> 00:04:10,680 machine learning, large language models, you know, 72 00:04:10,680 --> 00:04:13,040 generative AI we hear a lot of those words, 73 00:04:13,040 --> 00:04:15,060 but more specifically in the automation, 74 00:04:15,060 --> 00:04:20,539 the subset of the AI that we are using a lot in our smart sensors is deep learning. 75 00:04:20,539 --> 00:04:24,959 Now, deep learning is a subset of artificial intelligence that enables 76 00:04:24,959 --> 00:04:29,699 computers to mimic human decision making and problem solving capabilities. 77 00:04:29,699 --> 00:04:32,760 Deep learning uses neural networks to 78 00:04:32,760 --> 00:04:35,839 teach computers to process data inspired by human brain, 79 00:04:35,839 --> 00:04:39,540 and our deep learning solutions help you to be more precise, 80 00:04:39,540 --> 00:04:42,324 more flexible, and ultimately more successful. 81 00:04:42,324 --> 00:04:44,969 With our scalable product offerings, 82 00:04:44,969 --> 00:04:47,810 you can easily grow your business together with us. 83 00:04:47,810 --> 00:04:52,229 So we support your businesses in a simple and a flexible way, 84 00:04:52,229 --> 00:04:53,870 regardless of his size. 85 00:04:53,870 --> 00:04:55,390 No one thing to keep in mind, 86 00:04:55,390 --> 00:05:01,289 AI tools are a lot of them are being implemented in the Cloud and others. 87 00:05:01,289 --> 00:05:04,329 What we have seen is that causes some latency issues, 88 00:05:04,329 --> 00:05:07,309 and our focus is go where the data is generated, 89 00:05:07,309 --> 00:05:10,749 which is right at the sensor level and to reduce the low latency. 90 00:05:10,749 --> 00:05:16,090 And implementing an AI enabled automation solution needs some special skills, 91 00:05:16,090 --> 00:05:19,309 especially some consulting because not everybody has experience. 92 00:05:19,309 --> 00:05:22,249 Nick, can you talk a little bit about our consulting services? 93 00:05:22,249 --> 00:05:25,070 Yeah. So like Divya mentioned it. 94 00:05:25,070 --> 00:05:26,710 We're focused on the deep learning, 95 00:05:26,710 --> 00:05:28,730 we're focused on AI powered solutions. 96 00:05:28,730 --> 00:05:31,109 But how do we actually implement them into your warehouse? 97 00:05:31,109 --> 00:05:33,449 That's where our consulting services come into play. 98 00:05:33,449 --> 00:05:36,149 So we are here to help integrate into your warehouse. 99 00:05:36,149 --> 00:05:39,250 We serve to analyze that data, build up a solution, 100 00:05:39,250 --> 00:05:40,749 kind of develop what's needed, 101 00:05:40,749 --> 00:05:44,989 and then support the customer throughout and post integration. 102 00:05:44,989 --> 00:05:51,469 Yeah, our goal is to make it as easy as possible for implementing AI into your warehouse. 103 00:05:51,469 --> 00:05:54,350 So, Divya, can you speak a little bit more about how 104 00:05:54,350 --> 00:05:58,420 our sensors and applications will enable digital transformation? 105 00:05:58,420 --> 00:06:02,329 Absolutely. Everybody is on a journey right now of digital transformation, 106 00:06:02,329 --> 00:06:05,110 taking advantage of all the new digital tools come in. 107 00:06:05,110 --> 00:06:08,470 So are we and this is how we are helping our customers. 108 00:06:08,470 --> 00:06:10,609 Now, there's a huge, huge I mean, 109 00:06:10,609 --> 00:06:14,090 we have over 40,000 products in our portfolio, you know, 110 00:06:14,090 --> 00:06:17,080 But foundation, as I mentioned earlier, 111 00:06:17,080 --> 00:06:21,060 the data is what is the basis for the old digital transformation. 112 00:06:21,060 --> 00:06:23,739 The data starts right at the sensor level. 113 00:06:23,739 --> 00:06:29,280 You know, I can talk about specific applications a lot from different industries. 114 00:06:29,280 --> 00:06:35,319 We have intelligent cameras with inspection sensor apps that allows, 115 00:06:35,319 --> 00:06:40,100 you know, to inspect for bottle washer machines to see bottles are clean or not. 116 00:06:40,100 --> 00:06:43,420 This different industries, different applications using the same hardware, 117 00:06:43,420 --> 00:06:47,259 but different AI enabled or smart int inspection app. 118 00:06:47,259 --> 00:06:49,810 So let me take one specific example. 119 00:06:49,810 --> 00:06:53,260 Of cameras, how vision solutions are done. 120 00:06:53,260 --> 00:06:56,079 Now, if you look at it from a high level, 121 00:06:56,079 --> 00:06:59,859 there are three parts that really goes to putting a vision solution. 122 00:06:59,859 --> 00:07:01,580 There is the vision camera, right? 123 00:07:01,580 --> 00:07:06,480 It offers SICK has a lot of 2D and 3D cameras for industrial environment. 124 00:07:06,480 --> 00:07:11,140 Now, these cameras capture images and depth information, 125 00:07:11,140 --> 00:07:14,360 enabling precise measurements, inspections and identification. 126 00:07:14,360 --> 00:07:15,579 That is the camera itself, 127 00:07:15,579 --> 00:07:18,400 and our cameras are really good at doing that. 128 00:07:18,400 --> 00:07:20,640 And then the next step comes in is, 129 00:07:20,640 --> 00:07:23,059 Okay, we got the pictures. 130 00:07:23,059 --> 00:07:27,809 So how So how do we take advantage of OS inspection app? 131 00:07:27,809 --> 00:07:31,229 So we have onboard apps that simplify creating 132 00:07:31,229 --> 00:07:33,090 custom quality inspections for 133 00:07:33,090 --> 00:07:37,370 complex or irregular shaped goods for packaging, and for assemblies. 134 00:07:37,370 --> 00:07:40,710 So that is the second intelligence that goes on the camera out there. 135 00:07:40,710 --> 00:07:44,349 And the third one, which is more important is the real time processing. 136 00:07:44,349 --> 00:07:48,980 And this is goes back to the latency problem that issue that I talked about earlier. 137 00:07:48,980 --> 00:07:52,810 Our cameras can handle tastic object recognition, 138 00:07:52,810 --> 00:07:55,269 defect detection, and quality control. 139 00:07:55,269 --> 00:07:58,869 Now, if you combine all three down in the sensor level itself, 140 00:07:58,869 --> 00:08:02,189 that's where you get the big big advantage of these AI enables. 141 00:08:02,189 --> 00:08:05,249 So I mean, great timing. 142 00:08:05,249 --> 00:08:09,910 We just recently launched our AI enabled Inspector 83 X. 143 00:08:09,910 --> 00:08:14,049 It is designed to make quality control easy and efficient. 144 00:08:14,049 --> 00:08:16,669 The cutting AI technology enables 145 00:08:16,669 --> 00:08:21,530 non expert users to rapidly solve vision applications such as quality assurance, 146 00:08:21,530 --> 00:08:27,769 defect detection, and sorting directly on the device by simply teaching it by example. 147 00:08:27,769 --> 00:08:30,450 So no more building large models and everything else. 148 00:08:30,450 --> 00:08:32,350 You can teach it directly on the camera. 149 00:08:32,350 --> 00:08:35,750 This high speed and high resolution vision sensor 150 00:08:35,750 --> 00:08:39,829 provides an easy way to inspect every single part, 151 00:08:39,829 --> 00:08:41,269 every detail every time, 152 00:08:41,269 --> 00:08:45,090 reliably directly on the camera itself. 153 00:08:45,930 --> 00:08:48,269 Let's, Yeah. 154 00:08:48,269 --> 00:08:51,990 Go ahead. 155 00:08:51,990 --> 00:08:55,170 Oh, no. That sounds like a fascinating product via. 156 00:08:55,170 --> 00:08:58,210 With all the AI powered solutions entering the discussion, 157 00:08:58,210 --> 00:09:01,210 can you give us some more real cases in which AI enabled 158 00:09:01,210 --> 00:09:05,410 automation in the warehouse has significantly impacted supply chain operations? 159 00:09:05,410 --> 00:09:08,349 Absolutely. You know, as we say, 160 00:09:08,349 --> 00:09:11,609 AI is giving a lot more efficiency, 161 00:09:11,609 --> 00:09:15,870 adding to the efficiency and productivity as such, you know. 162 00:09:15,870 --> 00:09:18,689 Now, a lot of the task, 163 00:09:18,689 --> 00:09:20,950 the routine tasks that we done earlier, 164 00:09:20,950 --> 00:09:22,929 you know, repeatable tasks and others, 165 00:09:22,929 --> 00:09:25,730 with the AI enabled solutions, 166 00:09:25,730 --> 00:09:27,749 those tasks can be automated, 167 00:09:27,749 --> 00:09:30,370 ping employees to do a higher level tasks. 168 00:09:30,370 --> 00:09:33,129 So you'll see a lot more, 169 00:09:33,129 --> 00:09:36,889 Utilization of the resources more effectively to do 170 00:09:36,889 --> 00:09:40,729 some higher level work than just doing the routine repeatable work. 171 00:09:40,729 --> 00:09:45,409 The other areas that we see in the warehouse is logistics optimization, 172 00:09:45,409 --> 00:09:49,869 efficient pick routes, smart batching, to enhance productivity. 173 00:09:49,869 --> 00:09:53,950 Good example. Would be a specific example. 174 00:09:53,950 --> 00:09:56,350 Let's talk about a bag of chips. 175 00:09:56,350 --> 00:09:58,310 You know, I can be crushed, 176 00:09:58,310 --> 00:10:00,030 it can be crumpled. 177 00:10:00,030 --> 00:10:03,110 It can be anything, and it's a highly flexible object. 178 00:10:03,110 --> 00:10:04,669 Now with deep learning, 179 00:10:04,669 --> 00:10:11,610 the chip bag can still be detected and still be located and can be picked up efficiently. 180 00:10:11,610 --> 00:10:14,649 So the human and AIs are, 181 00:10:14,649 --> 00:10:18,189 you know, each other strength complementing each other strength. 182 00:10:18,189 --> 00:10:21,070 You know, AI is assisting in the crunching of the data, 183 00:10:21,070 --> 00:10:23,510 between the routine task and the physical labor, 184 00:10:23,510 --> 00:10:25,709 using robots and other aspects as well, 185 00:10:25,709 --> 00:10:27,890 while humans are providing the judgment, 186 00:10:27,890 --> 00:10:30,670 the creativity, and the leadership. 187 00:10:31,950 --> 00:10:33,430 Right. 188 00:10:33,430 --> 00:10:36,069 Thanks, Divya. So, Dave, now we turn to you. 189 00:10:36,069 --> 00:10:41,750 How can AI driven automation positively impact human work forces in the long run? 190 00:10:41,910 --> 00:10:47,650 Right. First to allow me to just mention my fascination with the speed of innovation, 191 00:10:47,650 --> 00:10:51,155 the enormous gains and productivity that AI is having. 192 00:10:51,155 --> 00:10:55,300 Particularly generative AI, and the investment. 193 00:10:55,300 --> 00:10:56,599 Just a couple of weeks ago, 194 00:10:56,599 --> 00:10:59,980 it was AI that announced $6,000,000,000 195 00:10:59,980 --> 00:11:04,700 around of funding to compete with the likes of Open AI and entropic. 196 00:11:04,700 --> 00:11:07,599 And of the fortune 500 companies, apparently, 197 00:11:07,599 --> 00:11:10,540 it's 99% are using AI. 198 00:11:10,540 --> 00:11:12,299 And Bruce, I wonder, what is that, 199 00:11:12,299 --> 00:11:14,039 you know, remaining 1%? 200 00:11:14,039 --> 00:11:17,339 Because, well, even with Apple intelligence on the way out, 201 00:11:17,339 --> 00:11:18,859 it's going to be 100% of us. 202 00:11:18,859 --> 00:11:21,560 And granted, this is a little outside of our discussion today. 203 00:11:21,560 --> 00:11:23,139 But back to your point, right, 204 00:11:23,139 --> 00:11:27,390 how do we use AI driven automation to impact our human workforce? 205 00:11:27,390 --> 00:11:30,959 It's the insight and the inference that's possible on 206 00:11:30,959 --> 00:11:35,319 our industrial data that is truly step function next level stuff. 207 00:11:35,319 --> 00:11:40,819 This data in our industrial world is more reliable and accessible than ever, 208 00:11:40,819 --> 00:11:44,139 data generated by new sensor technologies, 209 00:11:44,139 --> 00:11:46,920 new edge and distributed compute capability, 210 00:11:46,920 --> 00:11:49,599 and new system and workflow integrations. 211 00:11:49,599 --> 00:11:52,680 So AI driven automation will positively impact 212 00:11:52,680 --> 00:11:57,020 human work forces in the long run in three distinct, three key ways. 213 00:11:57,020 --> 00:11:59,040 It'll make workforces safer. 214 00:11:59,040 --> 00:12:01,000 It'll make them more productive, 215 00:12:01,000 --> 00:12:04,780 and subjectively, it'll make it more enjoyable. 216 00:12:04,780 --> 00:12:09,039 Let's look at the example of autonomous forklifts and collaborative robots that 217 00:12:09,039 --> 00:12:13,920 use computer vision for object detection and collision avoidance. 218 00:12:13,920 --> 00:12:15,319 They're able to adjust, 219 00:12:15,319 --> 00:12:18,280 avoid and re route themselves in milliseconds, 220 00:12:18,280 --> 00:12:21,739 considering the position and speed of the objects identified. 221 00:12:21,739 --> 00:12:24,920 Detection of such things as overhanging objects 222 00:12:24,920 --> 00:12:28,554 protruding from the rack or people and machines passing by. 223 00:12:28,554 --> 00:12:31,570 Computer vision alone is super powerful. 224 00:12:31,570 --> 00:12:35,789 And then there's AI for more advanced path planning and routing, 225 00:12:35,789 --> 00:12:39,550 not unlike how Google Maps is considering traffic 226 00:12:39,550 --> 00:12:44,589 and giving us impressively accurate ETAs to our destination, right? 227 00:12:44,589 --> 00:12:48,890 So this scenario, it's going to make our workforces more productive. 228 00:12:48,890 --> 00:12:55,189 Another example is the ability to simulate future outcomes using digital twins to 229 00:12:55,189 --> 00:12:58,169 identify risks and inefficiencies and make 230 00:12:58,169 --> 00:13:02,860 predictions to optimize or identify failure conditions. 231 00:13:02,860 --> 00:13:07,029 Interestingly, simulating robot fleets for route optimization and 232 00:13:07,029 --> 00:13:12,880 path planning has become a key development stage to scale mobile automation. 233 00:13:12,880 --> 00:13:17,490 And something worth mentioning is that AI isn't limited to language or vision, 234 00:13:17,490 --> 00:13:19,369 but it is truly multimodal. 235 00:13:19,369 --> 00:13:24,309 Consider the various sensors now standard on autonomous forklifts and AMRs. 236 00:13:24,309 --> 00:13:27,889 These are IMUs with accmeters, and gyroscopes. 237 00:13:27,889 --> 00:13:30,049 There's ometry and ultrasonics, 238 00:13:30,049 --> 00:13:32,390 LDR, radar. The list is long. 239 00:13:32,390 --> 00:13:35,490 Now, combine this with other external data, 240 00:13:35,490 --> 00:13:38,710 and sensors and AI is helping predict failures, 241 00:13:38,710 --> 00:13:42,409 suggest optimization, improvements, and so on. It's truly exciting. 242 00:13:42,409 --> 00:13:45,149 And the last subjective kind of element of 243 00:13:45,149 --> 00:13:48,309 this is going to make our work forces more enjoyable. 244 00:13:48,309 --> 00:13:51,130 Just look at what has been done for our dull, 245 00:13:51,130 --> 00:13:52,789 dirty, and dangerous work. 246 00:13:52,789 --> 00:13:57,069 We find tools, and we leverage mechanisms and automation. 247 00:13:57,069 --> 00:13:59,809 As a result, these tools allow us to focus on 248 00:13:59,809 --> 00:14:03,229 the more intellectually interesting and challenging tasks. 249 00:14:03,229 --> 00:14:05,549 So there you have, Bruce, that three ways that 250 00:14:05,549 --> 00:14:10,470 AI driven automation will positively impact human workforces in the long run. 251 00:14:10,470 --> 00:14:16,630 Oh, that's great. I think Mike Rose Dirty Jobs is one of my favorite shows. 252 00:14:16,630 --> 00:14:19,725 It'll be sad to see those go away. 253 00:14:19,725 --> 00:14:24,479 So are there ethical considerations supply chain professionals need 254 00:14:24,479 --> 00:14:29,379 to consider when implementing AI driven automation and warehouses, did you? 255 00:14:29,379 --> 00:14:31,280 Yeah. That's a good question. 256 00:14:31,280 --> 00:14:32,979 That's a great question, right? 257 00:14:32,979 --> 00:14:34,879 Anytime a new technology, 258 00:14:34,879 --> 00:14:36,419 anything is introduced, you know, 259 00:14:36,419 --> 00:14:38,639 there's always a question is, where is it? 260 00:14:38,639 --> 00:14:41,459 Is it ethical to use this and how we use it? 261 00:14:41,459 --> 00:14:43,640 Now, this is a very strong tool. 262 00:14:43,640 --> 00:14:45,979 You know, the AI tools are very strong. 263 00:14:45,979 --> 00:14:49,560 We're hearing about how it can do analyze and all the other aspects, 264 00:14:49,560 --> 00:14:55,424 especially the generative AI is seeing a lot more publicity, as I recall it. 265 00:14:55,424 --> 00:14:59,429 So one of the thing when you talk about ethical concerns, 266 00:14:59,429 --> 00:15:03,769 is to make sure because this is all about modeling, building a model, 267 00:15:03,769 --> 00:15:08,269 building as something that you are running your analysis against, right? 268 00:15:08,269 --> 00:15:11,850 It is important to have fairness, 269 00:15:11,850 --> 00:15:15,350 transparency, and mitigated bias. 270 00:15:15,350 --> 00:15:17,069 Right? When I say that is, 271 00:15:17,069 --> 00:15:19,150 as you're building the model, 272 00:15:19,150 --> 00:15:26,540 make sure the model is not influenced based on your personal biases or personal likings, 273 00:15:26,540 --> 00:15:29,839 or anything else that you can build in as you do it. 274 00:15:29,839 --> 00:15:31,420 Make sure it is fair, 275 00:15:31,420 --> 00:15:34,540 and it is actually data driven, 276 00:15:34,540 --> 00:15:37,439 not personal opinion driven. 277 00:15:37,439 --> 00:15:41,819 You know, so it is very important to be fair, you know, transparent, 278 00:15:41,819 --> 00:15:44,309 and biased, mitigated, you know, 279 00:15:44,309 --> 00:15:48,240 That is the first aspect of when you're building the models. 280 00:15:48,240 --> 00:15:52,480 Then you have to be responsible AI use and accountability. 281 00:15:52,480 --> 00:15:54,959 Where do you use it and for what purpose you use it? 282 00:15:54,959 --> 00:15:57,559 You you can use it for destructive purposes. 283 00:15:57,559 --> 00:15:59,280 You can use it for constructive purposes. 284 00:15:59,280 --> 00:16:02,160 It is very important that you use it for constructive purposes. 285 00:16:02,160 --> 00:16:04,919 You use it ethically and not use it to, 286 00:16:04,919 --> 00:16:09,490 you know, for infarous purposes, also. 287 00:16:09,490 --> 00:16:12,859 But the other neat thing that you do is, 288 00:16:12,859 --> 00:16:18,899 because of this AI driven and it is really freeing up the employees for high level work. 289 00:16:18,899 --> 00:16:22,999 So you are doing a lot of the stuff, the mundane work, 290 00:16:22,999 --> 00:16:25,680 and you're doing it the right way based on data, 291 00:16:25,680 --> 00:16:28,559 not on personal biases or something, you're taking that out. 292 00:16:28,559 --> 00:16:33,059 And the AI AI and robots can handle repetitive work, 293 00:16:33,059 --> 00:16:37,740 such as saving individuals from injury and injury costs like dave mentioned earlier. 294 00:16:37,740 --> 00:16:40,619 So those are some of the aspects that are going to go in. 295 00:16:40,619 --> 00:16:47,880 And if you apply it fairly with fairness and with transparency and with no bias, 296 00:16:47,880 --> 00:16:50,579 you take more advantage of it. 297 00:16:51,170 --> 00:16:53,030 Excellent. 298 00:16:53,030 --> 00:16:55,509 So, Nick, can you tell us what role 299 00:16:55,509 --> 00:16:58,535 machine learning plays in optimization in the warehouse? 300 00:16:58,535 --> 00:17:03,640 Sure. So we've talked about how AI is impacting the human workforce. 301 00:17:03,640 --> 00:17:06,419 But how is it actually optimizing it? 302 00:17:06,419 --> 00:17:08,680 When we talk about that optimization, 303 00:17:08,680 --> 00:17:11,659 we're talking about optimizing warehouse processes. 304 00:17:11,659 --> 00:17:14,099 Processes like demand forecasting, 305 00:17:14,099 --> 00:17:17,539 is essentially analyzing trends, that sale history, 306 00:17:17,539 --> 00:17:20,140 which then might lead to inventory management, 307 00:17:20,140 --> 00:17:22,839 kind of predicting when products might need replenishing 308 00:17:22,839 --> 00:17:25,820 or ensuring that products are always in stock. 309 00:17:25,820 --> 00:17:29,520 One of the latest trends we're seeing is predictive maintenance, 310 00:17:29,520 --> 00:17:35,040 analyzing some sensor data to predict a fault before it actually fatally occurs, 311 00:17:35,040 --> 00:17:38,259 All three of these processes are continuously 312 00:17:38,259 --> 00:17:41,500 approved upon through that reinforcement learning. 313 00:17:41,500 --> 00:17:44,539 The same neural networks that were used for training can 314 00:17:44,539 --> 00:17:47,859 always be continuously approved upon and excuse me, 315 00:17:47,859 --> 00:17:50,439 improved upon, with more data, 316 00:17:50,439 --> 00:17:53,479 and therefore driving efficiency in warehouse processes. 317 00:17:53,479 --> 00:17:55,619 With robotic systems specifically, 318 00:17:55,619 --> 00:17:58,959 machine learning can adapt to a live environment that's always changing. 319 00:17:58,959 --> 00:18:03,239 There's no need to stop production when a static environment changes. 320 00:18:03,239 --> 00:18:08,100 And by that, I mean, if a bin moves or boxes don't come stacked the same, 321 00:18:08,100 --> 00:18:11,459 Diva had mentioned the flexible potato chip bags that you know, 322 00:18:11,459 --> 00:18:14,209 they come in thousands of different orientations. 323 00:18:14,209 --> 00:18:16,839 Machine learning can adapt to that. 324 00:18:16,839 --> 00:18:18,940 One of the largest movements we're seeing, 325 00:18:18,940 --> 00:18:22,899 like I mentioned before, is the predictive maintenance for robots specifically. 326 00:18:22,899 --> 00:18:26,259 We're actually in the process of developing a solution to predict 327 00:18:26,259 --> 00:18:31,160 faults in G two of any robot brand before that error becomes catastrophic. 328 00:18:31,160 --> 00:18:35,020 All of this is in an effort to reduce down time and production. 329 00:18:35,020 --> 00:18:39,119 The last thing I want to mention about machine learning is the quick commissioning. 330 00:18:39,119 --> 00:18:43,579 There is no need for highly skilled professionals to install that system, 331 00:18:43,579 --> 00:18:46,759 even though the software may seem a bit complex, 332 00:18:46,759 --> 00:18:48,900 especially with some of those models. 333 00:18:48,900 --> 00:18:51,920 It is fairly simple to commission. 334 00:18:52,410 --> 00:18:55,590 That's great. With all that, to keep in mind, 335 00:18:55,590 --> 00:18:58,090 what challenges do supply chain professionals face 336 00:18:58,090 --> 00:19:01,129 when implementing AI driven automation and warehouses? 337 00:19:01,129 --> 00:19:03,250 Yeah, I can speak to that as well. 338 00:19:03,250 --> 00:19:05,809 Change management, it's a big concern. 339 00:19:05,809 --> 00:19:07,709 You have a lot of work flow changes, 340 00:19:07,709 --> 00:19:09,770 new tools that are becoming available, 341 00:19:09,770 --> 00:19:14,489 new trainings that are required to implement some of these AI powered solutions. 342 00:19:14,489 --> 00:19:17,510 So how do professionals actually deal with that change? 343 00:19:17,510 --> 00:19:19,929 The change itself, it is scary. 344 00:19:19,929 --> 00:19:22,089 There's a form of paralysis that's involved. 345 00:19:22,089 --> 00:19:26,650 It's an all in change, and it will affect every aspect of your warehouse operation. 346 00:19:26,650 --> 00:19:32,330 How do we go about introducing these new tools to minimize disruption to production? 347 00:19:32,330 --> 00:19:34,869 Um, you know, another challenge that people have 348 00:19:34,869 --> 00:19:37,209 been facing is it's not lived up excuse me, 349 00:19:37,209 --> 00:19:42,069 AI has not lived up to about 100% of its expectations of, 350 00:19:42,069 --> 00:19:44,929 you know, accurate reliability and whatnot. 351 00:19:44,929 --> 00:19:48,210 The reliability of models still has a lot of room to grow. 352 00:19:48,210 --> 00:19:50,650 It's still relatively new to the industry, 353 00:19:50,650 --> 00:19:53,390 especially with the data that's now available. 354 00:19:53,390 --> 00:19:57,990 Then there's not a lot of expertise on the functionality behind machine learning, 355 00:19:57,990 --> 00:20:02,169 like how the actual neural networks are are phySICKally generated. 356 00:20:02,169 --> 00:20:06,649 One thing for safety specifically is a fuzzy logic problem. 357 00:20:06,649 --> 00:20:09,709 The AI's fuzzy logic is a problem for 358 00:20:09,709 --> 00:20:14,949 the repeatable and certainty requirements of safety applications. 359 00:20:14,949 --> 00:20:17,730 And lastly, vision still has room to grow. 360 00:20:17,730 --> 00:20:22,289 What a camera sees versus what the human eye sees is quite different. 361 00:20:22,289 --> 00:20:25,429 That object classification, border definition, 362 00:20:25,429 --> 00:20:28,609 et cetera, they all have room to improve. 363 00:20:29,810 --> 00:20:33,050 Great. So as professionals in the industry, 364 00:20:33,050 --> 00:20:35,789 how can we address those challenges, do you? 365 00:20:35,789 --> 00:20:39,989 Yeah. It's a good question. Any time with technology is changing, 366 00:20:39,989 --> 00:20:41,670 new stuff is being introduced. 367 00:20:41,670 --> 00:20:43,609 It causes a lot of disruption, right? 368 00:20:43,609 --> 00:20:45,409 So how do we smoothen that out? 369 00:20:45,409 --> 00:20:49,230 That is the challenge that the supply chain professionals are facing right now? 370 00:20:49,230 --> 00:20:52,229 You know, as we bring in new tools, 371 00:20:52,229 --> 00:20:55,169 is how do I transition my workforce from how 372 00:20:55,169 --> 00:20:58,710 they did the work previously to how they have to do it now? 373 00:20:58,710 --> 00:21:01,130 What do they do with all the extra time they're 374 00:21:01,130 --> 00:21:03,249 going to get now with all these tools doing some of 375 00:21:03,249 --> 00:21:08,690 the routine work so that they can do some more higher level activities. 376 00:21:08,690 --> 00:21:10,729 So to do that, you know, 377 00:21:10,729 --> 00:21:12,570 and, of course, like you said, 378 00:21:12,570 --> 00:21:14,469 change management, you got to do that with 379 00:21:14,469 --> 00:21:17,389 your line still running without shutting down anything, right? 380 00:21:17,389 --> 00:21:18,970 So you're introducing automation, 381 00:21:18,970 --> 00:21:21,329 you're introducing AI enabled automation, 382 00:21:21,329 --> 00:21:26,089 and then you're changing the job function standard operating procedures with some of the, 383 00:21:26,089 --> 00:21:28,709 you know, with a newer tools and newer way of doing it. 384 00:21:28,709 --> 00:21:30,449 So workforce transition, 385 00:21:30,449 --> 00:21:32,550 Is most important. 386 00:21:32,550 --> 00:21:36,710 Second one is the skill development and adapting to the new roles. 387 00:21:36,710 --> 00:21:39,810 Now the way they did the work was one way. 388 00:21:39,810 --> 00:21:41,869 Now they have new tools and they have to learn how 389 00:21:41,869 --> 00:21:44,149 to use those tools to do the same work that 390 00:21:44,149 --> 00:21:48,770 they were doing earlier manually by hand or whatever process that they were using. 391 00:21:48,770 --> 00:21:50,590 So those are the challenges. 392 00:21:50,590 --> 00:21:56,610 How you transition, from the old way of doing things to the new way of doing 393 00:21:56,610 --> 00:21:58,509 things and bring that 394 00:21:58,509 --> 00:22:04,029 all the training that goes with it without distrupting your production, right? 395 00:22:04,029 --> 00:22:06,055 So without having any downtime. 396 00:22:06,055 --> 00:22:08,359 So the solutions, really, I mean, 397 00:22:08,359 --> 00:22:10,959 not that there's a one easy, you know, 398 00:22:10,959 --> 00:22:16,339 solution, but it involves reskilling, upskilling, cross training, 399 00:22:16,339 --> 00:22:18,759 redesigning of the business processes, 400 00:22:18,759 --> 00:22:26,139 and getting all that in place and lined up so that when you start down this journey, 401 00:22:26,139 --> 00:22:29,459 it follows a logical pattern and 402 00:22:29,459 --> 00:22:33,139 doesn't cause any major disruptions and shutdowns your production. 403 00:22:33,139 --> 00:22:36,110 That those are the challenges that need to be addressed. 404 00:22:36,110 --> 00:22:39,279 Sounds like plenty of work to do as AI rolls out. 405 00:22:39,279 --> 00:22:43,920 So to all three of you, one last question before we move on to audience Q&A. 406 00:22:43,920 --> 00:22:45,559 As we look to the future, 407 00:22:45,559 --> 00:22:48,980 what advancements in AI technology do you foresee having 408 00:22:48,980 --> 00:22:50,919 the most significant impact on 409 00:22:50,919 --> 00:22:55,609 warehouse automation and supply chain optimization? Who wants to go first? 410 00:22:55,609 --> 00:22:59,089 Let me go first, you know. Let me take this one. 411 00:22:59,089 --> 00:23:01,909 As I mentioned earlier, latency, 412 00:23:01,909 --> 00:23:05,430 I talked about latency and how do I get the data? 413 00:23:05,430 --> 00:23:08,729 Because data is the foundation to everything that we're talking about, 414 00:23:08,729 --> 00:23:11,630 and that data is right there at the sensor level. 415 00:23:11,630 --> 00:23:17,710 So if we can bring all the analysis and all the analytical tools right at the edge, 416 00:23:17,710 --> 00:23:21,280 you know, right if not on the sensor at least onto the edge. 417 00:23:21,280 --> 00:23:24,219 You know, edge computing is something that we're going to see, 418 00:23:24,219 --> 00:23:26,059 I see a lot more, you know, 419 00:23:26,059 --> 00:23:30,040 coming down the line, getting more stronger and more powerful. 420 00:23:30,040 --> 00:23:32,039 And then reinforced learning. 421 00:23:32,039 --> 00:23:35,339 That's the other one is, how do I enable operators? 422 00:23:35,339 --> 00:23:38,199 How do I paint them, train them, excuse me, 423 00:23:38,199 --> 00:23:42,279 in real time with all the additional data and background information. 424 00:23:42,279 --> 00:23:46,040 And of course, Nick touched a little bit about it on AI vision. 425 00:23:46,040 --> 00:23:47,899 These are the three things that I see is where 426 00:23:47,899 --> 00:23:50,400 the warehouse automation is going to see a lot more, 427 00:23:50,400 --> 00:23:53,500 you know, areas of impactment. 428 00:23:53,500 --> 00:23:56,199 And then of course, there's advancements in the hardware. 429 00:23:56,199 --> 00:23:58,080 And then try to get the software, 430 00:23:58,080 --> 00:24:00,379 the hardware as close to the human eye as possible. 431 00:24:00,379 --> 00:24:02,480 He talked about the vision side. 432 00:24:02,480 --> 00:24:06,600 The other last thing I want to touch up is the natural language processing. 433 00:24:06,600 --> 00:24:08,379 You know, you're hearing a lot of that. 434 00:24:08,379 --> 00:24:12,739 You'll see a lot of the warehouse workers to able to talk and interact with 435 00:24:12,739 --> 00:24:17,340 the robot using natural language instead of using a whole bunch of programming, 436 00:24:17,340 --> 00:24:19,659 you know, not a script like or something else, 437 00:24:19,659 --> 00:24:21,939 it's making it a more natural conversation. 438 00:24:21,939 --> 00:24:25,220 And that is what I see coming down the line. 439 00:24:25,220 --> 00:24:27,620 That is so true. 440 00:24:27,620 --> 00:24:30,239 Allow me to jump in agree with this. 441 00:24:30,239 --> 00:24:33,740 Computer vision inference at the edge is absolutely hot. 442 00:24:33,740 --> 00:24:37,800 We're getting good at this object classification and detection. 443 00:24:37,800 --> 00:24:42,020 We've got new product announcements and just new technology all the time here, 444 00:24:42,020 --> 00:24:44,740 so that computer vision is key. 445 00:24:44,740 --> 00:24:46,480 But just on your last point, DP, 446 00:24:46,480 --> 00:24:51,099 I'm also excited in this sort of I'm calling it partner integrations, 447 00:24:51,099 --> 00:24:52,959 using natural language models, 448 00:24:52,959 --> 00:24:54,719 but to query the data, 449 00:24:54,719 --> 00:24:58,199 making it far easier to identify outliers, 450 00:24:58,199 --> 00:25:01,899 trends, and and overall new actionable insights, right? 451 00:25:01,899 --> 00:25:03,860 As many of you have faced in the past. 452 00:25:03,860 --> 00:25:06,835 It's often painfully tedious and time consuming. 453 00:25:06,835 --> 00:25:10,270 To set up filters and constraints and reorganize 454 00:25:10,270 --> 00:25:14,470 the data and just eyeballing tables and to get the meaningful insights. 455 00:25:14,470 --> 00:25:19,149 But now with the natural language technology is available to us, 456 00:25:19,149 --> 00:25:21,944 it means terrific advancements. 457 00:25:21,944 --> 00:25:24,979 Yeah. And to summarize things out, you know, 458 00:25:24,979 --> 00:25:27,400 I'm excited to see where I goes, 459 00:25:27,400 --> 00:25:29,260 but AI driven automation, 460 00:25:29,260 --> 00:25:32,579 it does revolutionize your warehouse operations, 461 00:25:32,579 --> 00:25:37,600 and collaboration between humans and AI is the key to unlocking its full potential. 462 00:25:37,600 --> 00:25:43,479 Keep in mind, we got to stay with the ethical considerations and workforce adaptation. 463 00:25:43,479 --> 00:25:45,859 They're critical for successful implementation. 464 00:25:45,859 --> 00:25:50,220 So looking ahead, I'm excited to see the advancements in AI technology, 465 00:25:50,220 --> 00:25:52,060 and I'm hoping that it'll continue to drive 466 00:25:52,060 --> 00:25:55,900 efficiency and optimization in your supply chain. 467 00:25:56,110 --> 00:25:59,650 Alright. Fantastic. So now, 468 00:25:59,650 --> 00:26:02,870 I'll remind everybody that there is a place to put questions 469 00:26:02,870 --> 00:26:07,870 in in the Slido pole at the bottom of the screen. 470 00:26:08,870 --> 00:26:12,789 We have a question here. So go ahead and put anything in. 471 00:26:12,789 --> 00:26:14,649 Let me just read this one that we have. 472 00:26:14,649 --> 00:26:17,950 So does our expansion into AI enabled 473 00:26:17,950 --> 00:26:22,349 sensors make us more vulnerable to chip shortages moving forward? 474 00:26:22,349 --> 00:26:26,049 I think we're all have a good bit of experience with that, 475 00:26:26,049 --> 00:26:28,729 if you've been in the industry for I don't know, 476 00:26:28,729 --> 00:26:32,725 four or five years now. Who wants to start? 477 00:26:32,725 --> 00:26:35,039 Let me take a stab at it. 478 00:26:35,039 --> 00:26:38,179 So yeah, the chip shortages, as you talk about. 479 00:26:38,179 --> 00:26:40,059 I mean, you have heard in the news, right? 480 00:26:40,059 --> 00:26:42,799 I mean, most of the major governments and everybody is 481 00:26:42,799 --> 00:26:45,360 now ramping up the C chip production, was the Chip Act. 482 00:26:45,360 --> 00:26:47,950 So there's There's definitely 483 00:26:47,950 --> 00:26:50,389 a whole impetus on trying to grow and 484 00:26:50,389 --> 00:26:53,450 expand on the amount of chips that's readily available. 485 00:26:53,450 --> 00:26:56,029 The next thing to keep in mind is, 486 00:26:56,029 --> 00:26:58,989 there's only finite number of chips out there. 487 00:26:58,989 --> 00:27:02,149 What kind of chips can we used to do some of the stuff? 488 00:27:02,149 --> 00:27:05,489 Now, you need some advanced chips to do AI when you're 489 00:27:05,489 --> 00:27:08,890 doing a larger data modeling up at a higher level. 490 00:27:08,890 --> 00:27:12,429 But then if you bring it down to a small I use a lego block concept, 491 00:27:12,429 --> 00:27:14,390 bring it down to at a sensor level, 492 00:27:14,390 --> 00:27:17,269 use a smaller data set and then doing that stuff, 493 00:27:17,269 --> 00:27:18,950 you need a different kind of chips. 494 00:27:18,950 --> 00:27:23,315 So that's where six sensors have the smart sensors going in out there. 495 00:27:23,315 --> 00:27:25,920 So I won't say that we are immune from the shortage, 496 00:27:25,920 --> 00:27:28,340 but we are not that off, 497 00:27:28,340 --> 00:27:29,920 like we used to be in the past, 498 00:27:29,920 --> 00:27:32,159 because the type of chips being used is very 499 00:27:32,159 --> 00:27:34,800 commonly available is used in all kinds of devices, 500 00:27:34,800 --> 00:27:38,659 so that's available, much freely and easily accessible. 501 00:27:38,659 --> 00:27:42,799 But yes, we'll always have a dependency on the chips. 502 00:27:42,960 --> 00:27:45,659 Bruce, if I may comment, 503 00:27:45,659 --> 00:27:48,740 I think we also need to look at the investments being made. 504 00:27:48,740 --> 00:27:52,219 Yes, it's not going to turn on our supply and affect it immediately. 505 00:27:52,219 --> 00:27:55,740 But look at the billions of dollars now even Intel in Ohio, 506 00:27:55,740 --> 00:27:57,260 now breaking new ground, 507 00:27:57,260 --> 00:28:00,360 but all these particularly in North America, 508 00:28:00,360 --> 00:28:01,659 but around the world, too, 509 00:28:01,659 --> 00:28:08,079 this race to improve our supply chain on semiconductors is real, and it's helping. 510 00:28:08,079 --> 00:28:13,480 Hopefully, this AI enablement helps that supply chain as well. 511 00:28:14,710 --> 00:28:17,430 So one more question here. 512 00:28:17,430 --> 00:28:20,949 We talked about, I think you mentioned fuzzy logic, 513 00:28:20,949 --> 00:28:25,690 Nick, and you talked about AI enabled safety. 514 00:28:25,690 --> 00:28:33,989 And I was watching an interview with the now ex CEO of Waymo, 515 00:28:33,989 --> 00:28:36,849 talking about how on the streets of San Francisco, 516 00:28:36,849 --> 00:28:41,369 people were we're trying to fool the system, 517 00:28:41,369 --> 00:28:43,509 stepping in front of moving cars, 518 00:28:43,509 --> 00:28:47,089 wearing T shirts that had a stop sign on them just to see if 519 00:28:47,089 --> 00:28:51,789 they if they could manipulate the cars. 520 00:28:51,789 --> 00:28:59,130 Is there anything that AI is going to do to help as safety gets more and more complex? 521 00:29:00,370 --> 00:29:02,910 Yeah, I think, Dave, 522 00:29:02,910 --> 00:29:09,409 you might be the best one to answer this one for safety. I do have some thoughts. 523 00:29:09,409 --> 00:29:12,170 So you know these edge cases are going to continue to be an issue. 524 00:29:12,170 --> 00:29:16,609 As we said, cones on the hood or the stop signs on our shirts. 525 00:29:16,609 --> 00:29:18,650 Those are real sort of edge cases we hadn't 526 00:29:18,650 --> 00:29:21,369 accounted for or trained the models for, and they didn't know what to do. 527 00:29:21,369 --> 00:29:23,050 And they shut down the vehicle. 528 00:29:23,050 --> 00:29:25,730 You could remotely now tie in and control the vehicle. 529 00:29:25,730 --> 00:29:28,190 But it meant that the cars were stuck in the lane, 530 00:29:28,190 --> 00:29:29,784 you know, impeding traffic. 531 00:29:29,784 --> 00:29:35,019 And while our investment has been largely in the machine learning modules. 532 00:29:35,019 --> 00:29:39,179 They're very tightly bound and not in the open world scenarios. 533 00:29:39,179 --> 00:29:40,799 They are different use cases, 534 00:29:40,799 --> 00:29:46,959 but the safety is if we can tightly bind and know where and can predict. 535 00:29:46,959 --> 00:29:51,559 Just look at what anthropic has professed lately with they're able to look at 536 00:29:51,559 --> 00:29:57,399 neuro networks and now understand where are the drift is occurring in bias, as we say. 537 00:29:57,399 --> 00:30:00,800 They're able to now look into the neuro network to identify, oh, 538 00:30:00,800 --> 00:30:06,739 we should re bias this and tweak it to now safeguard you know, 539 00:30:06,739 --> 00:30:09,140 drifting and other conditions of our models. 540 00:30:09,140 --> 00:30:15,500 So real advancements are happening here to even make this for safety applications. 541 00:30:15,780 --> 00:30:18,199 So we have another question here, 542 00:30:18,199 --> 00:30:20,379 kind of the elephant in the room, right? 543 00:30:20,379 --> 00:30:22,219 The question specifically is, 544 00:30:22,219 --> 00:30:25,100 what are the concerns with cybersecurity? 545 00:30:25,100 --> 00:30:28,219 And I think we've seen, you know, 546 00:30:28,219 --> 00:30:34,244 potential vulnerabilities with edge connected smart devices. What do we do about that? 547 00:30:34,244 --> 00:30:37,009 Yeah. I'll jump in here. 548 00:30:37,009 --> 00:30:38,709 There's a lot of directions, 549 00:30:38,709 --> 00:30:40,870 also regulations coming in place, 550 00:30:40,870 --> 00:30:44,670 and then there's of course a strategy for separation with firewalls, 551 00:30:44,670 --> 00:30:46,350 with the de militarized zone. 552 00:30:46,350 --> 00:30:49,429 There's a lot of technical advancements going on trying to separate out 553 00:30:49,429 --> 00:30:52,889 the what we call the shop floor from the top floor, right? 554 00:30:52,889 --> 00:30:56,690 So there's a lot of differentiation going on out there. 555 00:30:56,690 --> 00:30:59,329 There's different technologies going in. 556 00:30:59,329 --> 00:31:02,629 We can spend hours talking about that one, but that is a concern, 557 00:31:02,629 --> 00:31:06,349 but it is being addressed by a lot of different ways by putting data dios, 558 00:31:06,349 --> 00:31:10,439 one way transform transferring of data. 559 00:31:10,439 --> 00:31:12,749 So, yes, it is a concern, 560 00:31:12,749 --> 00:31:15,010 but it is also being worked in the industry, 561 00:31:15,010 --> 00:31:20,110 trying to solve that with some reliable repeatable solutions. 562 00:31:20,110 --> 00:31:23,270 So let's just have one more question here. 563 00:31:23,270 --> 00:31:26,330 There's one more in the slido pole. 564 00:31:26,330 --> 00:31:30,670 Many warehouses today do not use robotics, 565 00:31:30,670 --> 00:31:35,769 and we work with them using Lean Six Sigma methodology to help them optimize. 566 00:31:35,769 --> 00:31:40,249 How do you see Lean Six Sigma integrating with AI? 567 00:31:40,249 --> 00:31:43,929 Yeah. Lean Six Sigma is foundational, right? 568 00:31:43,929 --> 00:31:49,290 I mean, that is you are looking at to try and find how do I optimize my production, 569 00:31:49,290 --> 00:31:54,670 you know, how I take the any activities that is non productive out of out of the process. 570 00:31:54,670 --> 00:31:56,729 So you will do that exercise, 571 00:31:56,729 --> 00:31:59,030 and then you're trying to find out where is the value add. 572 00:31:59,030 --> 00:32:00,490 If you're doing something repeatable, 573 00:32:00,490 --> 00:32:03,809 that's not and it's not these are areas you will 574 00:32:03,809 --> 00:32:08,109 then help you identify put robots in down the line if you need it to. 575 00:32:08,109 --> 00:32:11,470 So six Sigma they both go hand in hand, 576 00:32:11,470 --> 00:32:13,469 I would say, personally, in my opinion. 577 00:32:13,469 --> 00:32:15,709 One is to understand where the weaknesses 578 00:32:15,709 --> 00:32:18,509 and where efficiency can be gained in your process, 579 00:32:18,509 --> 00:32:20,569 and that comes out of the six Sigma process. 580 00:32:20,569 --> 00:32:22,889 And then once you identify those areas, 581 00:32:22,889 --> 00:32:25,090 then you mitigate those areas by using 582 00:32:25,090 --> 00:32:28,809 technology like robots and others AI enabled tools. 583 00:32:29,120 --> 00:32:32,880 All right. So great questions. 584 00:32:32,880 --> 00:32:34,999 We really appreciate it. 585 00:32:36,280 --> 00:32:41,100 So we do have some more questions in the poll. 586 00:32:41,100 --> 00:32:43,919 Please feel free to put your information in. 587 00:32:43,919 --> 00:32:52,159 We're a couple of minutes over the 30 minutes that the tech talk was scheduled for. 588 00:32:52,159 --> 00:33:00,120 So let's throw it back to Dave Nick and Divya, any closing remarks. 589 00:33:00,510 --> 00:33:02,749 Hey, thanks so much, Bruce. 590 00:33:02,749 --> 00:33:06,149 And thank you to MHI for hosting and putting this together. 591 00:33:06,149 --> 00:33:09,509 It's appreciated. I love the thought that I've gone into this, 592 00:33:09,509 --> 00:33:13,610 and it's wonderful, delightful to talk with you guys this morning. Thanks. 593 00:33:13,610 --> 00:33:15,790 Yeah, I'm going to go next. Thank you, Bruce. 594 00:33:15,790 --> 00:33:18,070 And thank you MHI. This was a great discussion. 595 00:33:18,070 --> 00:33:23,390 Thank you for the invitation for us to join and participate in this MHI talk. 596 00:33:23,390 --> 00:33:25,769 Yeah. And I'll reiterate a third time. 597 00:33:25,769 --> 00:33:27,470 Thank you, MHI. This has been great. 598 00:33:27,470 --> 00:33:28,689 I appreciate the invite. 599 00:33:28,689 --> 00:33:32,609 I learned a lot myself, and I hope you all have some takeaways from today as well. 600 00:33:32,609 --> 00:33:35,510 Excellent. Well, big thanks to our audience, 601 00:33:35,510 --> 00:33:39,570 the panelists, and especially our host MHI, for participating today. 602 00:33:39,570 --> 00:33:42,150 I hope everyone has a great day.