WEBVTT 1 00:00:00.171 --> 00:00:03.253 Agentic AI, you've heard the term, but what is it really? 2 00:00:03.493 --> 00:00:06.495 And how is it already changing the way that energy companies are working? 3 00:00:07.075 --> 00:00:13.239 Our guest today on Power Perspectives is AI strategist, Jeff Wilser, and he's here to answer those questions for us. 4 00:00:13.559 --> 00:00:15.020 Jeff, welcome to the show today. 5 00:00:21.913 --> 00:00:23.377 Jeff, let's start with the basics. 6 00:00:23.477 --> 00:00:27.105 When you say agentic AI workflow, what exactly does that mean? 7 00:00:27.145 --> 00:00:31.796 And how is that different than say, someone at a utility company using a chat bot to ask a question? 8 00:00:32.605 --> 00:00:32.965 Sure. 9 00:00:32.985 --> 00:00:36.068 I know you've had a lot of recent great episodes on AI. 10 00:00:36.128 --> 00:00:38.770 So I'm sure many of your listeners have some basics here. 11 00:00:38.890 --> 00:00:50.618 But for those who are new to the agent world, I think when most people think of something like ChatGPT or Chatbot, it's mostly about just asking a question and getting a response, right? 12 00:00:50.718 --> 00:00:56.662 So simple analogy, hey, I want to go to Tokyo for a vacation. 13 00:00:57.163 --> 00:00:57.523 What do you 14 00:01:00.445 --> 00:01:02.588 would actually do the things for you. 15 00:01:02.748 --> 00:01:09.476 So instead of just getting an answer, like here is a good itinerary in Tokyo, you can say, hey, I want to buy plane tickets. 16 00:01:09.696 --> 00:01:12.379 Please buy me plane tickets, book me hotels. 17 00:01:12.860 --> 00:01:18.567 So you can think of it as actually the arms and legs of doing things, not just the brain itself. 18 00:01:19.063 --> 00:01:19.363 Gotcha. 19 00:01:19.383 --> 00:01:20.643 No, that's definitely helpful. 20 00:01:20.703 --> 00:01:22.584 It's a little bit more action-based. 21 00:01:23.004 --> 00:01:33.246 And for our audience that's in the utility and energy world, what's an example of something that maybe one of their employees spends hours doing today that an AI agent could realistically handle? 22 00:01:33.786 --> 00:01:34.367 Absolutely. 23 00:01:34.547 --> 00:01:39.988 And Matt, I think a lot of folks hear about AI agents and they immediately race to 24 00:01:40.888 --> 00:01:46.452 the most provocative, scary, crazy thing that can be done. 25 00:01:46.572 --> 00:01:51.335 Like some agents who has full autonomy of a control room. 26 00:01:51.915 --> 00:02:01.702 But there is a lot of middle grounds and a lot of far less glamorous, more basic things that can be done before you get to that point. 27 00:02:02.202 --> 00:02:09.469 I think the number one thing, and this is true, so many industries and top of utilities and energy is paperwork. 28 00:02:09.969 --> 00:02:16.816 Virtually every industry, including you guys, we all have more paperwork than we want to do. 29 00:02:17.516 --> 00:02:34.165 And that ranging from compliance documents to trying to find what your rates were in some niche area from 12 years ago, there are buckets and buckets of paperwork that no one really has time to do. 30 00:02:34.745 --> 00:02:41.969 And AI agents can be built to actually sift through the paperwork and summarize and actually draft documents. 31 00:02:42.189 --> 00:02:46.151 So kind of the two-year-ago and one-year-ago AI thing was, 32 00:02:46.671 --> 00:02:57.779 was to manually, let's say, feed a PDF into something like ChatDBT or Claude or Copilot and say, hey, summarize this for me. 33 00:02:58.520 --> 00:03:11.950 Now that could be done with far more efficiency where agents can actually romp through stacks and stacks of paperwork and glean insights and help you create summary documents. 34 00:03:12.697 --> 00:03:16.278 And is the benefit just that it's able to do it that much quicker and more efficiently? 35 00:03:16.358 --> 00:03:19.699 Or do you find that the AI agent can actually do it better too? 36 00:03:20.319 --> 00:03:35.922 So in your world, think about a compliance issue that no one analyst could possibly have time to read the hundreds of precedents and documents behind this, but an agent can go through. 37 00:03:36.602 --> 00:03:46.840 And not just do it faster, but have access to this sprawling amount of paperwork to turn up insights you wouldn't otherwise have known to even look for. 38 00:03:47.262 --> 00:03:52.365 Yeah, that's definitely, I think, something that the utility executives would be eager to learn more about. 39 00:03:52.626 --> 00:03:57.629 And you mentioned that people jumped to the idea of like, oh, the AI is running the control room. 40 00:03:57.649 --> 00:03:59.790 And then there's a little bit of an aversion to that. 41 00:04:00.090 --> 00:04:05.074 But even when you're just looking at the paperwork and the regulatory aspect of it, it's still important. 42 00:04:05.194 --> 00:04:14.380 And I'm curious how much you find that there's oversight or fact checking still needed from the human side of the equation when the AI agents are doing these tasks. 43 00:04:14.832 --> 00:04:24.035 For sure, there's a cliche that I'm sure you've heard, but it's one of those cliches that happens to be true, and that's keeping a human in the loop. 44 00:04:24.395 --> 00:04:27.877 And that can happen in a couple of different levels, right? 45 00:04:27.917 --> 00:04:32.498 You could have the human spot check and look for errors. 46 00:04:33.178 --> 00:04:37.580 Obviously, you want someone to be sanity checking it before it goes anywhere sensitive. 47 00:04:38.200 --> 00:04:45.663 But one thing I think that one thing I share when I work with companies and workshops is, yes, human loop. 48 00:04:45.743 --> 00:04:53.787 But one additional thing you can do to help with control and help guardrails is actually using AI agents and AI models. 49 00:04:54.327 --> 00:04:59.833 kind of to cross-check each other in an adversarial, constructive way, right? 50 00:04:59.873 --> 00:05:13.649 So imagine that you, to keep it very simple, imagine you're working with something like Claude to create some agentic workflow and your agent says, aha, I am now going to create this compliance document. 51 00:05:14.310 --> 00:05:17.471 You can then feed it to a different system entirely. 52 00:05:17.671 --> 00:05:32.277 Let's say in ChachiPT, give it the rules of the road, give it the context and say, hey, I want you with fresh eyes and to get the nice splash cold water here, cold water, I want you to evaluate this and find blind spots. 53 00:05:32.777 --> 00:05:38.460 So by using AI system to check the other, that can help actually determine 54 00:05:39.240 --> 00:05:51.515 reduce errors and give yourself more of a more um again i wouldn't use the word certainty a human should always be in a loop but it's helpful to have ais check each other and that can reduce errors as well 55 00:05:52.713 --> 00:06:00.478 Is that one of the go to solutions that we hear about hallucinations or just, you know, faulty conclusions sometimes come from the AI? 56 00:06:00.518 --> 00:06:05.281 And, you know, the natural inclination is to have the humans be the one checking that, like you said, spot checking. 57 00:06:05.341 --> 00:06:12.686 But do you find that it's a better practice or almost time saver to have that, you know, AI checking its own work? 58 00:06:13.138 --> 00:06:13.839 For sure. 59 00:06:14.059 --> 00:06:20.083 It's, again, not to take the human out of this, but it's like one additional level of protection before the human, right? 60 00:06:20.303 --> 00:06:31.230 And it's also, it's kind of a lot of fun too when you have these two different models and you almost surprise them and say, hey, by the way, I asked Claude the same question. 61 00:06:31.830 --> 00:06:47.296 here's what it found, and you give it its answer, that will almost, is a way to spur the first model to make sure it's buttoned up, and to actually second and triple and quadruple check itself. 62 00:06:47.756 --> 00:06:59.065 I know in the utility sector, we're in an environment where these companies are seeing massive liabilities with things like wildfire management, outages that impact vulnerable customers. 63 00:07:00.026 --> 00:07:03.929 The utility leaders don't really feel like they can afford to let a mistake through. 64 00:07:04.189 --> 00:07:08.492 And this is great advice for how to capture and minimize those mistakes. 65 00:07:08.592 --> 00:07:14.256 But on the chance that there is an AI agent involved and a mistake does get through, you 66 00:07:15.017 --> 00:07:18.038 Who ultimately is held accountable when something goes wrong? 67 00:07:18.478 --> 00:07:24.900 There's one group that is definitely without question will benefit from AI, and that is lawyers. 68 00:07:25.280 --> 00:07:30.102 There are going to be so many gray area legal questions about who ultimately is accountable. 69 00:07:30.662 --> 00:07:34.663 That is a legal question beyond my pay grade as far as the actual liability. 70 00:07:35.123 --> 00:07:40.365 That said, I think the North Star, the way I view all of this for your industry, 71 00:07:41.986 --> 00:07:47.850 is something that I call the red line theory and where you are on the red line. 72 00:07:48.571 --> 00:07:59.999 And what I mean by that is what I found is the most successful way companies can use AI, whether in utilities or in publishing or in healthcare, 73 00:08:00.719 --> 00:08:14.113 is to think clearly about where is the red line of what you hold sacred that AI should not touch, that humans, to your point, Matt, ultimately must be accountable for. 74 00:08:14.634 --> 00:08:22.182 So for one example, in media and editing and journalism, Time magazine, they said, okay, 75 00:08:22.662 --> 00:08:28.927 We're not going to use AI to write AI slop to write our articles and do our research. 76 00:08:29.548 --> 00:08:31.809 That is still the work of humans. 77 00:08:32.650 --> 00:08:41.678 But the chief operating officer of Time, who I interviewed recently, what he told me was, yes, the AI is not going to replace the human writers. 78 00:08:42.258 --> 00:08:43.499 However, that's a red line. 79 00:08:43.539 --> 00:08:45.060 That's what time is sacred for. 80 00:08:45.140 --> 00:08:48.263 That is their sacred editorial judgment and credibility. 81 00:08:48.803 --> 00:08:50.405 On the other side of the red line, 82 00:08:50.985 --> 00:08:57.610 They want to use AI ambitiously, and they created AI agents to do far more things that cannot be done with time. 83 00:08:57.750 --> 00:09:07.717 Now they have AI agents and AI bots that let you interact with Time Magazine and explore its 103 year history in ways you couldn't do before. 84 00:09:07.797 --> 00:09:09.098 They added new functionality. 85 00:09:09.378 --> 00:09:19.826 The goal is not to take humans out of the loop of using judgment with respect to the wildfires and all the tricky questions you're facing. 86 00:09:20.326 --> 00:09:34.109 The goal is to give humans more time to think deeply about edge cases, about the sensitive issues, by automating the more drudgery stuff, the mundane stuff, to give humans more time to do what they are really good at. 87 00:09:34.732 --> 00:09:35.712 Yeah, that's interesting. 88 00:09:35.732 --> 00:09:45.177 It makes me think of, we ran a survey a little while back asking utility executives, what do they think AI will not replace in the utility sector? 89 00:09:45.277 --> 00:09:52.800 And the main tranche of answers came into, one, there's the physical work of repairing a line and that much was obvious. 90 00:09:52.820 --> 00:09:54.060 Oh, the robots are coming. 91 00:09:54.120 --> 00:09:56.361 I'm sure the robots are coming imminently and they'll do all of that. 92 00:09:56.381 --> 00:09:57.262 Eventually, yeah. 93 00:09:58.342 --> 00:10:06.067 But the other aspects that were cited multiple times were empathy for the customers and, you know, having kind of that human connection. 94 00:10:06.207 --> 00:10:12.271 And then the sort of judgment that leaders have to make based on all the information that AI can help collect for them. 95 00:10:12.631 --> 00:10:18.635 But ultimately, a human has to make, you know, weigh the pros and cons and the different tradeoffs and make that ultimate decision. 96 00:10:19.221 --> 00:10:20.462 Yeah, well said. 97 00:10:20.642 --> 00:10:25.066 And along those lines, think about flying and aviation. 98 00:10:26.027 --> 00:10:27.048 That's similar to your world, right? 99 00:10:27.068 --> 00:10:28.589 The stakes just as utilities. 100 00:10:29.030 --> 00:10:32.593 When you're flying a plane, human lives are quite literally at stake. 101 00:10:33.233 --> 00:10:41.140 When autopilots, when that technology developed years and decades ago, 102 00:10:41.901 --> 00:10:44.063 pilots were understandably skittish about it. 103 00:10:44.223 --> 00:10:47.146 They did not want to trust flying to a computer. 104 00:10:47.286 --> 00:10:47.966 Like, fair. 105 00:10:48.086 --> 00:10:49.548 That is an understandable reaction. 106 00:10:50.108 --> 00:10:55.573 They also worried they could be replaced by robots and by self-flying planes. 107 00:10:56.234 --> 00:11:02.259 And what happened was not that pilots were replaced or that pilots stopped using their judgment. 108 00:11:02.459 --> 00:11:05.622 What happened instead was pilots, they still had to know how to fly, but 109 00:11:06.102 --> 00:11:16.895 They still made the judgment calls, but they're actually able to make their role more about the judgments as opposed to doing more of the manual things. 110 00:11:16.915 --> 00:11:20.139 I think utilities can be looked at the same way, right? 111 00:11:20.339 --> 00:11:23.083 Ultimately, it's the human's job to apply the judgments. 112 00:11:23.663 --> 00:11:27.266 And if we can automate some systems to make that job easier, great. 113 00:11:27.766 --> 00:11:38.434 But I would also encourage utilities to think about it from the lens, not just of automating and cutting, but even growing. 114 00:11:38.794 --> 00:11:42.497 And by growing, I don't necessarily mean doubling revenue. 115 00:11:43.077 --> 00:11:48.782 But rather, what else could you do that you couldn't do beforehand? 116 00:11:48.902 --> 00:12:07.438 That's a big light bulb I see when I work with CEOs about this across industries is most companies and most organizations are thinking very narrowly about, okay, maybe I can use AI to cut this one thing or to automate this. 117 00:12:08.178 --> 00:12:13.982 And sure, automation is great, but it's also just scratching the surface of what you can do. 118 00:12:14.302 --> 00:12:28.392 I'd encourage everyone listening to your podcast to do is to take a moment and imagine, let's say theoretically, your headcount has doubled without spending an extra nickel. 119 00:12:29.492 --> 00:12:31.735 What then would you be able to do? 120 00:12:32.455 --> 00:12:34.718 What additional safeguards could you create? 121 00:12:35.358 --> 00:12:36.539 What additional services? 122 00:12:37.360 --> 00:12:40.964 Where are you able to add more value if your head got doubled? 123 00:12:41.545 --> 00:12:49.673 And then think, okay, how can AI allow us to actually get some of that new resource and capability? 124 00:12:49.693 --> 00:12:49.953 Okay. 125 00:12:50.374 --> 00:13:03.330 So I think that's one way that people in utilities can use AI, not just for like saving off a bit of cost at the margin, but to more ambitiously get more done in your organization. 126 00:13:03.830 --> 00:13:07.154 Yeah, that's a terrific thought experiment and how to approach it. 127 00:13:07.275 --> 00:13:07.475 Yeah. 128 00:13:07.815 --> 00:13:18.465 kind of going back to the treachery of the paperwork we were talking about earlier, you know, I know a longstanding challenge in the power sector is siloing, you know, information processes. 129 00:13:18.786 --> 00:13:23.250 They're always separated between different departments and the right hand doesn't always know what the left hand is doing. 130 00:13:23.730 --> 00:13:26.093 So given that part of that problem is that utilities have 131 00:13:26.733 --> 00:13:31.516 So much siloed data buried in PDFs, filings, engineering documents, old records. 132 00:13:32.036 --> 00:13:39.980 How close do you think we are to a world where, you know, through an agentic AI, an employee can essentially have a conversation with all of that institutional history? 133 00:13:40.983 --> 00:13:42.304 I think it's as close as you want to be. 134 00:13:42.704 --> 00:13:50.408 The technology exists right now, and it's not something that requires billions of dollars of investment. 135 00:13:50.588 --> 00:13:51.729 It's mostly about that. 136 00:13:51.749 --> 00:13:55.831 The irony, Matt, is I think this is one thing that's kind of lost in the conversation. 137 00:13:55.851 --> 00:13:58.853 There's so much AI hype out there and all these buzzwords. 138 00:13:59.773 --> 00:14:05.418 Even the word agentic AI can sound kind of magic wandy. 139 00:14:05.738 --> 00:14:12.444 Ultimately, an AI agent is mostly about context and context files. 140 00:14:13.104 --> 00:14:19.489 And by context files, kind of fancy way of saying really just like, it could be as simple as text documents. 141 00:14:19.510 --> 00:14:20.851 You mentioned PDFs, right? 142 00:14:21.011 --> 00:14:28.217 And so what's making or breaking AI adoption, both in the power industry and most industries at large, 143 00:14:28.897 --> 00:14:34.622 is not about are you using the most cutting edge, powerful frontier model. 144 00:14:35.243 --> 00:14:41.048 It's less about the model's power and more about how well are you feeding it context. 145 00:14:41.568 --> 00:14:48.815 What I mean by that is, as you point out, Matt, there could be PDFs living in, there's somewhere in IT, there's all kinds of PDFs. 146 00:14:48.835 --> 00:14:50.957 Somewhere in procurement, in marketing. 147 00:14:52.398 --> 00:14:59.825 Every different division has their own weird warehouse figuratively or literally of all these different files. 148 00:15:00.946 --> 00:15:13.398 Getting those files, having a smart infrastructure that is feeding the relevance files and takeaways from the files into a system that AI agents read is, 149 00:15:14.239 --> 00:15:15.220 That's the whole ballgame. 150 00:15:15.620 --> 00:15:25.047 The other area where, you know, utility leaders traditionally see knowledge spread out is where it's not written down at all, but it's stored in the minds of longstanding employees. 151 00:15:25.107 --> 00:15:31.892 And this is a common topic in the utility sector as much of the workforce is aging closer and closer to retirement. 152 00:15:32.252 --> 00:15:33.914 Are we risking losing that knowledge? 153 00:15:34.294 --> 00:15:40.899 I'm curious what you think AI agentic tools can do to capture that institutional knowledge and maybe re-reporting. 154 00:15:41.199 --> 00:15:49.366 plug in that gap where, yeah, hey, there might have been that 40, that employee who's been there for 40 years, who knows, oh, this is how this substation really behaves. 155 00:15:49.406 --> 00:15:54.450 How can we set up the agentic AI to be able to understand that as well? 156 00:15:55.000 --> 00:15:56.381 Yeah, I'm so glad you brought that up. 157 00:15:56.401 --> 00:15:57.142 That's a great one. 158 00:15:57.182 --> 00:15:58.002 I think you're right. 159 00:15:58.102 --> 00:16:01.024 It's super relevant to the industry here. 160 00:16:01.305 --> 00:16:05.548 And some might call it tribal knowledge that is a danger of being lost. 161 00:16:06.188 --> 00:16:11.952 And I think it's one often overlooked use case of AI agents. 162 00:16:12.233 --> 00:16:14.074 And one of my favorites is technology. 163 00:16:14.374 --> 00:16:21.779 interviews and actually having... Now, this could be done a couple of ways. 164 00:16:22.319 --> 00:16:36.127 You can set up an interview for almost like a series of prompts that this salty engineer about to retire can then answer, crucially, answer verbally, right? 165 00:16:36.307 --> 00:16:36.808 If it's... 166 00:16:37.388 --> 00:16:41.890 If we're asking for type response, that's not going to be nearly as effective. 167 00:16:42.070 --> 00:16:45.292 No one likes to type a 5,000-word essay, right? 168 00:16:45.872 --> 00:16:51.595 But if you ask them questions about things that, frankly, they'll probably be kind of fluttered. 169 00:16:51.635 --> 00:16:52.916 They're asked about, like, oh... 170 00:16:53.636 --> 00:17:00.142 Yeah, I recall well that ice storm of 1987, and no one's thought to ask me that in years, right? 171 00:17:00.542 --> 00:17:12.032 And have them go through and give verbally detailed responses that can create this library, this corpus of knowledge. 172 00:17:12.292 --> 00:17:14.914 And here's where agents really come into play. 173 00:17:16.035 --> 00:17:25.359 In a world before AI, if you were to, let's say, spend, and I'm making this roadmap up, but bear with me here. 174 00:17:25.819 --> 00:17:34.782 Let's say you have maybe two hours a week you devote to interviewing this, the salty veteran engineering knowledge, right? 175 00:17:34.802 --> 00:17:43.206 So over the course of a year, you have a hundred hours of them talking about all kinds of esoteric things they've seen. 176 00:17:44.126 --> 00:17:44.967 And do that at scale. 177 00:17:45.267 --> 00:17:51.612 So you've now interviewed, let's say, let's say there's thousands of hours of conversation of these folks. 178 00:17:52.213 --> 00:17:57.978 In a non-AI world, good luck getting anyone to go through that. 179 00:17:58.958 --> 00:18:06.104 It reminds me of the last scene of Raiders of the Lost Ark with the crates and crates and crates in the warehouse and all these things. 180 00:18:06.645 --> 00:18:10.088 No one can go through those crates in the warehouse 181 00:18:10.728 --> 00:18:11.409 and make sense of this. 182 00:18:11.669 --> 00:18:25.900 The agents can go through this, at this point, this massive library of interviews you've conducted with all these folks who are about to retire and make heads and tails of it. 183 00:18:26.180 --> 00:18:29.043 It could synthesize things. 184 00:18:29.423 --> 00:18:32.325 You can actually have a few different cuts of this library. 185 00:18:32.345 --> 00:18:37.770 You can say, hey, I want to create a masterclass about ice storms 186 00:18:38.390 --> 00:18:43.296 or about substations and how they reacted in this kind of emergency. 187 00:18:43.916 --> 00:18:51.245 And you can create an entire documents and dossier around these thousands of interviews. 188 00:18:51.946 --> 00:18:58.893 And then once you have that effectively, this massive database of tribal knowledge in one place, 189 00:18:59.514 --> 00:19:05.437 In the future, you could ask AI agents to search into it, to probe, to ask questions. 190 00:19:05.897 --> 00:19:18.423 A more advanced thing would that mean to create an actual, you can create effectively your own chatbot based off of this database to let you real-time ask questions to it as well. 191 00:19:18.503 --> 00:19:22.305 So I think we're going to see, that's still relatively fast, 192 00:19:22.865 --> 00:19:26.912 fringe and bleeding edge use case right now. 193 00:19:27.132 --> 00:19:34.064 In the future, my prediction, that sort of use case will go mainstream across many industries. 194 00:19:35.047 --> 00:19:36.567 Yeah, that's fascinating. 195 00:19:36.607 --> 00:19:39.708 And I definitely think that the utility sector is primed for that. 196 00:19:39.748 --> 00:19:45.770 Like I said, the aging workforce, this long institutional knowledge, it's too valuable to just let it go. 197 00:19:45.870 --> 00:19:52.072 So maybe the AI agents are coming at the perfect time to fill in that gap that's so needed. 198 00:19:52.472 --> 00:19:58.094 But I guess as we wrap up our conversation, Jeff, I want to ask the advice-based question. 199 00:19:58.314 --> 00:19:58.354 If 200 00:19:59.534 --> 00:20:06.740 somebody's listening in, they're running utility, they want to experiment with agentic AI this week, not two years from now, not those fringe cases. 201 00:20:07.140 --> 00:20:12.645 What is the first workflow you would tell them to try and how would they know whether it's actually working? 202 00:20:13.573 --> 00:20:24.059 One thing that, and if your listeners already tried this, great, move on to the next, but I think something that everyone busy in the utility industry can deal with. 203 00:20:24.080 --> 00:20:32.385 But before you think about how can we inject AI in most sensitive workflows of our power grid, 204 00:20:33.445 --> 00:20:35.926 Whether we like it or not, we all have busy inboxes. 205 00:20:36.207 --> 00:20:39.648 We all struggle to make the meetings on time. 206 00:20:39.889 --> 00:20:42.230 We all have a hard time keeping track of our to-dos. 207 00:20:43.150 --> 00:20:55.557 I would create your own chief of staff, your own AI agent who is able to triage email for you, who is able to hoover up all the projects you have. 208 00:20:56.598 --> 00:21:02.621 Most systems now let you easily connect in your email company Slack. 209 00:21:03.401 --> 00:21:05.082 to-do list systems, and so on. 210 00:21:05.723 --> 00:21:11.146 And crucially, my advice to even be more specific is to give it lots of context. 211 00:21:12.086 --> 00:21:24.754 Spend an hour walking through what's important to you, what your goals are, what are risks you're worried about, and have it then each morning, 212 00:21:26.795 --> 00:21:34.058 give you a briefing for the day and say, hey, what are the three most things I must do today? 213 00:21:34.258 --> 00:21:35.399 What's most important? 214 00:21:36.099 --> 00:21:44.103 This, I think that's a, one reason I like this starting point so much is that one, it is very de-risking, right? 215 00:21:44.123 --> 00:21:46.043 You are not putting the grid at risk. 216 00:21:46.444 --> 00:21:49.365 You are starting small with your own personal workflows. 217 00:21:50.105 --> 00:21:53.587 And two, I think most people will get a ton of value out of it. 218 00:21:53.627 --> 00:22:08.673 When I work with CEOs and show this in workshops, teams are like, oh my God, it is amazing how a little bit of agentic systems in your own workflow can handle so much busy work. 219 00:22:08.813 --> 00:22:18.717 And one more bit of very practical advice that I think listeners will benefit from is when you're working with AI, 220 00:22:19.497 --> 00:22:20.137 in this manner. 221 00:22:20.537 --> 00:22:26.899 The way to speed things up and the way to give it even more context is actually through voice, not through typing. 222 00:22:27.459 --> 00:22:35.441 Reason why is that almost everyone, certainly including me, despite typing for a living, is much faster at talking than typing. 223 00:22:35.741 --> 00:22:37.741 It's also more enjoyable to talk than to type. 224 00:22:38.361 --> 00:22:44.723 And so when we're going back to when we interviewed the salty 35-year veteran about to retire, 225 00:22:45.423 --> 00:22:52.346 they're going to be more likely to share lots of rich details if they're just casually, conversationally chatting. 226 00:22:52.526 --> 00:22:58.089 If they're typing, they cut corners, but the corners are where the good stuff lives. 227 00:22:58.909 --> 00:23:00.490 The corners are the edge cases. 228 00:23:00.830 --> 00:23:01.571 That's the nuance. 229 00:23:02.031 --> 00:23:03.932 The more nuance you give, the better. 230 00:23:04.072 --> 00:23:08.034 When you're talking, you're more likely to just share a random parenthetical. 231 00:23:08.474 --> 00:23:13.818 the AI will actually store that parenthetical and that's useful context in the future, right? 232 00:23:13.959 --> 00:23:27.610 And so, and last, not to be overly mechanical here, but when you hear voice mode, people might be tempted to actually turn on the kind of like voice mode in ChachiPT or Claude, or mainly ChachiPT. 233 00:23:28.250 --> 00:23:37.278 That's not what I mean, because that, when you go back and forth with voice, almost like Siri, and we all know how bad Siri is, you kind of only get the lightest, 234 00:23:37.938 --> 00:23:40.600 most surface level version of the AI. 235 00:23:40.720 --> 00:23:44.363 It's not the more powerful thinking versions of the model. 236 00:23:44.963 --> 00:23:48.946 Instead, what I mean is you can use a dictation service. 237 00:23:49.347 --> 00:24:00.015 And for those who are still not sure exactly where to start, literally after you finish listening, Matt, to your great podcast, they can go in and say, hey, chatbot of choice. 238 00:24:00.755 --> 00:24:04.338 I want to create my own chief of staff AI agents. 239 00:24:04.898 --> 00:24:05.939 Show me how to do it. 240 00:24:06.400 --> 00:24:07.621 They will show you how to do it. 241 00:24:08.121 --> 00:24:15.047 And the last bonus thing they can do too is say, because this is moving so fast, AI is evolving so fast. 242 00:24:15.847 --> 00:24:24.010 they should say in this prompt, also go and do some latest research to make sure you're up to date on what's possible. 243 00:24:24.030 --> 00:24:39.195 Because the irony is, Matt, AI is evolving so fast that a lot of these chatbots don't even know what they're capable of because the training date is cut off and things have happened in the last two months since training data. 244 00:24:39.635 --> 00:24:43.456 So that's where I would start and you will get your eugenic system before you know it. 245 00:24:44.076 --> 00:24:46.840 Jeff, again, this is a topic that's been on the top of everyone's mind. 246 00:24:46.860 --> 00:24:51.445 You mentioned we've been covering A on the podcast a lot, but you definitely gave us some new insights to chew on. 247 00:24:51.505 --> 00:24:53.087 And we thank you for joining us today. 248 00:24:53.408 --> 00:24:53.868 Thanks so much. 249 00:24:53.888 --> 00:24:54.449 Had a great time. 250 00:24:55.170 --> 00:24:59.375 And as always, I also want to thank our audience for being with us today on Power Perspectives. 251 00:24:59.555 --> 00:25:00.617 And we'll catch you next time.