Latest / MaxValue Podcast / How AI Scales Your Business: From Chief Everything Officer to CEO

Transcript

How AI Scales Your Business: From Chief Everything Officer to CEO · MaxValue Podcast

• 0:12
  SPEAKER_01: Welcome to the Max Value Podcast, a show for business owners and the trusted advisors who are on a mission to maximize their value through valuation, exit planning, and maybe even cloning yourself.
• 0:21
  SPEAKER_01: I'm your host, Ben Town, and today we are joined by Dr.
• 0:25
  SPEAKER_01: Steve Kirch, and we're going to discuss how AI moves leaders from the chief everything officer to the chief executive officer, all while protecting IP and boosting engagement.
• 0:37
  SPEAKER_01: Steve, welcome.
• 0:40
  SPEAKER_00: Thanks, Ben.
• 0:41
  SPEAKER_00: It's great to be here.
• 0:43
  SPEAKER_01: Awesome.
• 0:44
  SPEAKER_01: Great to have you.
• 0:45
  SPEAKER_01: It feels like we just talked to each other.
• 0:47
  SPEAKER_01: I don't know why.
• 0:48
  SPEAKER_01: Uh you because you have a podcast and we went on that.
• 0:53
  SPEAKER_00: And I'm I'm looking forward to publishing that episode.
• 0:55
  SPEAKER_00: It was a great conversation.
• 0:57
  SPEAKER_01: Very soon, very soon.
• 0:58
  SPEAKER_01: So let's see if we can uh do do at least as good for for this one this time.
• 1:03
  SPEAKER_01: So before we jump into the content, give us your origin story.
• 1:07
  SPEAKER_01: How did you get here?
• 1:08
  SPEAKER_01: What are you doing now?
• 1:11
  SPEAKER_00: Yeah, so uh it's kind of an interesting story, and and one of the reasons in in my podcast I do this as well is because everybody has got a really interesting story about how they got to where they are.
• 1:23
  SPEAKER_00: And I started out, you know, I have a PhD in physics, and I worked in high-tech for about 30 years.
• 1:29
  SPEAKER_00: Um, along the way, discovered a passion for teaching people how to be more efficient and effective in their jobs.
• 1:37
  SPEAKER_00: Um, and uh became a certified facilitator of the Covey 7 habits of highly effective people.
• 1:43
  SPEAKER_00: I've been doing that for almost 30 years now, um, along with a number of other courses of that ilk personal, professional, or you know, organizational development, productivity, effectiveness, all that kind of stuff.
• 1:56
  SPEAKER_00: Um, when I gave up my corporate gig, um now a little more than nine years ago, I decided to hang out a shingle as a productivity coach.
• 2:06
  SPEAKER_00: Uh and what I discovered was that small business owners in particular have absolutely no clue what their time is worth.
• 2:16
  SPEAKER_00: And that created a challenge for me as a productivity coach.
• 2:20
  SPEAKER_00: I was trying to sell them an increase in the value of something they didn't understand.
• 2:26
  SPEAKER_00: And and as a consequence, I got, you know, I had some clients, but I got a lot of pushback.
• 2:32
  SPEAKER_00: And I'm I'm constantly learning and growing and understanding and learning new things and putting new things in my toolbox.
• 2:40
  SPEAKER_00: And I came across a system for profit growth.
• 2:43
  SPEAKER_00: Um, and then um through the nonprofit where I'm on the board, uh, developed a program for scaling businesses, uh particularly nonprofits, uh, in in by using the Rockefeller habits.
• 2:58
  SPEAKER_00: Some people know this as scaling up, other people know it as EOS.
• 3:02
  SPEAKER_00: I actually use a yet another platform, but it's all based on the same principles that John Rockefeller discovered as he was running Standard Oil, what, 100 plus years ago?
• 3:11
  SPEAKER_00: Um, was reasonably effective as a uh um as a corporate leader.
• 3:19
  SPEAKER_00: Um but uh so so we've adapted some of those.
• 3:24
  SPEAKER_00: 15 months ago or so, again, as I'm continuing to learn and grow and and discover new things, I came across the system that enables me to help small businesses really leverage AI.
• 3:38
  SPEAKER_00: Um, and we can enable every employee in every role to do everything that they do faster and easier than it with by using AI.
• 3:47
  SPEAKER_00: And this includes the chief everything officer.
• 3:50
  SPEAKER_00: I mean, that's that's one of the beauties of the scaling up process, right?
• 3:57
  SPEAKER_00: One of the things that I go through with my clients is we have a list of all of the organizational duties, all of the different boxes that have to be checked, all of the different things that need to be done to run an organization.
• 4:12
  SPEAKER_00: And I ask who's responsible, and how do you know when they're successful?
• 4:17
  SPEAKER_00: What's the metric that matters in that role?
• 4:20
  SPEAKER_00: And the thing that I discover so often when I talk to small business owners, to um you know, people that that run their business, and and it's you know, typically it's not a publicly traded company, right?
• 4:36
  SPEAKER_00: Is that their name is on every box and they're literally responsible for every decision that's taking place in that company, and that's not scalable.
• 4:50
  SPEAKER_00: So we help figure out well, what can you offload?
• 4:54
  SPEAKER_00: How do you do that?
• 4:58
  SPEAKER_01: That totally makes sense.
• 4:59
  SPEAKER_01: And I think the part of that comes from when especially when you're first starting out, maybe as a solopreneur, you you literally are the chief everything officer.
• 5:11
  SPEAKER_01: And some things are just ignored, some things are an afterthought.
• 5:18
  SPEAKER_01: But there is this impression that you are a free resource in a certain sense.
• 5:25
  SPEAKER_01: Not that you want to take so much time doing it, but if I just need to work one more hour, I will just go do that in order to accomplish all the things until at some point you get burned out and you can't do it anymore.
• 5:36
  SPEAKER_01: You know, I always think of it like in kind of like in a geometry sense where you know, if all the things that a CEO needs to do are a circle, as it scales outward, the volume of things that you need to do as the owner of that circle gets bigger even than the than the perimeter um on a proportional base.
• 6:01
  SPEAKER_01: On a proportional basis, you just can't keep up.
• 6:04
  SPEAKER_01: Um it goes like a square law, you know, pi r squared.
• 6:08
  SPEAKER_01: Now there you go.
• 6:09
  SPEAKER_01: And that's probably more thank you, Dr.
• 6:11
  SPEAKER_01: Steve.
• 6:12
  SPEAKER_01: Uh that probably a more elegant way of saying that.
• 6:14
  SPEAKER_01: But uh I think the uh the the visual holds true.
• 6:17
  SPEAKER_01: There's just too much stuff, and very quickly as as everything uh expands in every direction at once.
• 6:23
  SPEAKER_00: The the other piece of it, Ben, is very often nobody will do it as well as I do.
• 6:31
  SPEAKER_00: Nobody cares as much as I do about the outcome of the decision.
• 6:36
  SPEAKER_00: There's there's a certain amount of ownership that they're somehow somewhat reluctant to give up.
• 6:42
  SPEAKER_00: And I get it, right?
• 6:43
  SPEAKER_00: I'm I'm I I run a small business.
• 6:45
  SPEAKER_00: I I I know, right?
• 6:47
  SPEAKER_00: It's me, my daughter, and I have a couple of assistants now.
• 6:51
  SPEAKER_00: So I understand why giving up that that authority, that final check is is so hard.
• 7:03
  SPEAKER_01: I I I I will say, and you know, we're we're kind of going off in in a in a little bit of a different direction for a second, but I will say that when Alan Yang, my business partner, joined, it really changed my perspective very much.
• 7:16
  SPEAKER_01: And I think this might be what a lot of other solopreneurs go through at some point.
• 7:22
  SPEAKER_01: It it everything changes from a literal I to a we, first off.
• 7:27
  SPEAKER_01: And things like, well, the bank account, which is kind of how a lot of small businesses keep track of their success, is their money in the bank account.
• 7:38
  SPEAKER_01: Uh, that's not always the best way to do it, but it is it's a very tangible way.
• 7:43
  SPEAKER_01: Yeah, it goes from well, that's my money in a different account, a non-personal account, to that's the business's money as a separate entity.
• 7:54
  SPEAKER_01: And so the concept of having this separate entity running uh with you, but in in some some ways having a life apart from you is is super important.
• 8:05
  SPEAKER_01: And so when we're then I think that can feed into the idea of well, uh I'm not just the everything, I need to be a part of this organization, this entity that lives without me, outside of me.
• 8:22
  SPEAKER_01: And therefore, I need to optimize my own usage of myself in that case.
• 8:28
  SPEAKER_01: It's a little bit of a weird, you know, duality of thinking, but I think that's what we kind of have to do in order to get the best out of ourselves.
• 8:37
  SPEAKER_00: Yeah.
• 8:38
  SPEAKER_00: I think one one one of my coaches said, you know, if you were gonna hire yourself to be the CEO of the company, what what tasks would you give yourself?
• 8:52
  SPEAKER_00: What what responsibilities would you give to the CEO?
• 8:57
  SPEAKER_00: Not to you, but to the CEO, just happens to also be you.
• 9:02
  SPEAKER_00: Um and think about that it in that way.
• 9:08
  SPEAKER_01: Well, being introspective is hard.
• 9:10
  SPEAKER_01: Let's just let's just say that.
• 9:11
  SPEAKER_01: I mean, gosh, if if people could do it easily, you wouldn't have a job.
• 9:15
  SPEAKER_01: Uh, and and maybe neither would I.
• 9:17
  SPEAKER_01: At the end of the day, though, you know, now we have uh better tools than ever to do some of these things, including, and kind of one of the central focus points for the uh conversation today is AI.
• 9:31
  SPEAKER_01: So if if you had to bake it down into a sentence, uh what could a small business expect to have happen if they actually do implement AI and do it correctly?
• 9:47
  SPEAKER_00: Yeah, so one of the things that we have found is that when everybody in the company adopts AI in the in the way in which we try to infuse so that everybody is thinking about AI first, or I'll say AI forward, right?
• 10:02
  SPEAKER_00: Becoming AI primary, um we can we can basically somewhere between 40 and 60 percent reduction in time spent on the things that are on your list, the things that you do on a daily basis.
• 10:17
  SPEAKER_00: Um and and we tell the CEOs, you know, you've got to get out in front of this because some people are afraid.
• 10:24
  SPEAKER_00: But basically you have three choices, right?
• 10:27
  SPEAKER_00: You can fire half your staff.
• 10:30
  SPEAKER_00: We don't recommend that, but you can do it.
• 10:32
  SPEAKER_00: It's an option, right?
• 10:34
  SPEAKER_00: You can double the output, just have them produce twice as much of what they're currently doing, or you can free them to do the stuff that they say, oh gosh, I wish I had, I wish I could do that, I wish I had the time to do that.
• 10:51
  SPEAKER_00: Um and so that's really, you know, in in one sentence, we find a 40 to 60 percent reduction, if properly implemented, across the board, all knowledge work.
• 11:03
  SPEAKER_01: Right.
• 11:03
  SPEAKER_01: And and I think that that's what ends up leading to this.
• 11:06
  SPEAKER_01: Well, gosh, if I just had two of myself, everything would be easier.
• 11:09
  SPEAKER_01: Well, you can't quite do that, but through leverage, in this case, uh technological leverage, we're able to get twice as much done as well as we would before in you know the same amount of time or the same amount in half a time.
• 11:26
  SPEAKER_01: Uh the uh the ability uh to have uh a shift in the focus of your productivity is awesome.
• 11:39
  SPEAKER_01: I mean, even just thinking about, and I'll bring this back to boring old accounting for a second, uh between when I first started, when everything wasn't computerized, not quite, there were still a couple of clients that I worked with that used column paper to do their accounting.
• 11:58
  SPEAKER_01: And it took hours even just to check whether they had they totaled it, totaled it upright.
• 12:05
  SPEAKER_01: Whereas obviously now you get it in QuickBooks and you have the or whatever, uh you you can import things right away, and you know, somebody at a you know fairly reasonable level of knowledge can you know create a reasonable set of books in minutes.
• 12:22
  SPEAKER_01: Uh I'll say minutes, you might even be able to count it in seconds or something pretty easy, but but certainly minutes, and that's a uh you know, just what's happened in the last say 20 years or so, and now we're at you know a new cusp.
• 12:36
  SPEAKER_01: Um and being at a new cusp, that's scary, or can be.
• 12:44
  SPEAKER_01: And AI definitely does scare some people, and some people, you know, you you kind of kind of joked about it for a second.
• 12:51
  SPEAKER_01: Some people are worried about losing their jobs over it.
• 12:53
  SPEAKER_01: Uh, I'm going to be replaced by AI, is it's a reasonable fear, whether or not right.
• 13:00
  SPEAKER_01: It's yeah, no, no, you're not.
• 13:02
  SPEAKER_01: But and I'm certainly not, and you're not.
• 13:05
  SPEAKER_01: Um even the regular workers.
• 13:07
  SPEAKER_00: Yeah.
• 13:08
  SPEAKER_01: Absolutely.
• 13:09
  SPEAKER_01: So, but but you do make it sound like, hey, this is a blue sky, unalloyed good for for people.
• 13:18
  SPEAKER_01: Um, but what what do you see at kind of the ground level for staff morale when when AI is actually implemented well?
• 13:28
  SPEAKER_00: Well, what we find is people are more efficient than effective, yes, but they're also more engaged.
• 13:36
  SPEAKER_00: And why are they more engaged?
• 13:38
  SPEAKER_00: Because they have this really long list of stuff that they wish they could get to.
• 13:43
  SPEAKER_00: They just don't have the time to do it because they got all this piddly little stuff that they have to get done, right?
• 13:50
  SPEAKER_00: These are the have-to's, that's the stuff that gets done.
• 13:53
  SPEAKER_00: And if they run out of time, well, the stuff that they would like to do or would be fun for them to do, or is more interesting about the job, that's stuff they don't have time for.
• 14:04
  SPEAKER_00: Well, if you can reduce the have to to two or three days a week, then you have two or three days a week of stuff that you want to do, that you're excited about doing.
• 14:18
  SPEAKER_00: And so again, assuming that the that the leader comes up, gets out in front and says, look, we're not learning AI so we can get rid of half of you.
• 14:29
  SPEAKER_00: We're learning AI so we can double the output of the company.
• 14:36
  SPEAKER_00: And really take it to the next level.
• 14:38
  SPEAKER_00: And of course, what does that do to the what does that do to the valuation, Ben?
• 14:41
  SPEAKER_00: That's that's your bailiwick, right?
• 14:44
  SPEAKER_01: Oh, absolutely.
• 14:44
  SPEAKER_01: I mean, then you start looking at things like uh you know, revenue per full-time equivalent.
• 14:50
  SPEAKER_01: It obviously drops right down to the the bottom line.
• 14:54
  SPEAKER_01: Maybe if you have especially effective personnel, you can pay afford to pay them more and have more profitability at the same time.
• 15:00
  SPEAKER_01: You paying them more tends to lead to more uh uh retention of your staff, which is excellent because not retaining your staff is very expensive, yada yada yada.
• 15:12
  SPEAKER_01: So at the end of the day, whenever whenever you're valuing something, you want to see reduced risk and increased cash flow and possibly even growth.
• 15:20
  SPEAKER_01: And whenever you can properly leverage something, including uh technology, including AI, that can help get you there.
• 15:29
  SPEAKER_00: So also things, there's also things like like um being really clear on who owns what, the outcome of meetings.
• 15:39
  SPEAKER_00: I mean, I use tools to record my conversations with everybody, can follow up easily with, oh yeah, here's what we talked about, here's your action items from this meeting.
• 15:51
  SPEAKER_00: Some of the note takers even tell you what they are, so you don't even have to do that.
• 15:57
  SPEAKER_01: Uh I I absolutely use that for almost almost every prospect call.
• 16:03
  SPEAKER_01: Uh, obviously, there there are exceptions to this, but what happens is uh you you get those transcripts and you get those recordings, and you're able to share it with the rest of your team, even though they weren't there in the first place.
• 16:14
  SPEAKER_01: So, to the extent that you know I need the project manager to come up to speed on something on a on a new project that we have, boom, there you go.
• 16:23
  SPEAKER_01: And uh yeah, I've taken I take notes too, because I'm gonna think about it maybe in a little bit of a different way, but 90% of it already is gonna be there just from the conversation I have with the client, and and they can they can say it, it's it's good stuff.
• 16:35
  SPEAKER_01: Well, it's a re it's definitely a a chance to reinvest in your in your staff.
• 16:43
  SPEAKER_01: And there are, of course, I think more truly scary and more truly challenging parts.
• 16:50
  SPEAKER_01: Like, what about keeping your intellectual property, keeping proprietary confidential data safe from some of the I'll say some of the more popular or within some of the more popular AI platforms?
• 17:07
  SPEAKER_00: Yeah, so uh we actually, you know, we run into this all the time, where people they're using a free account.
• 17:16
  SPEAKER_00: And if you use a free account, you know who has who gets the value out of that, and you know what the value is, right?
• 17:24
  SPEAKER_00: It's your data.
• 17:26
  SPEAKER_00: So you have to be very careful when you're using a free account, just like you know, Facebook or any of those, right?
• 17:34
  SPEAKER_00: Your data is the product, right?
• 17:38
  SPEAKER_00: So please don't do that.
• 17:41
  SPEAKER_00: Create an IP policy, pay for your employees to use a platform, make sure you turn off all of the data sharing options.
• 17:51
  SPEAKER_00: And of course, if somebody wants to talk to me about that, I'm happy to show them how that happens.
• 17:57
  SPEAKER_00: But even better, we have a platform through a partner that creates an IP vault for your data.
• 18:04
  SPEAKER_00: So when you upload it to this platform that's an interface to large language models, um, your data is safe and secure inside that IP vault.
• 18:17
  SPEAKER_00: And you can communicate with, well, open AI, even up through GPT-5, which just got released last week, um, with Claude, um, Anthropics Claude, several different models of that, um, through uh Amazon Nova Pro, uh, Google's Gemini platform.
• 18:38
  SPEAKER_00: Again, lots of different choices there, lots of different models.
• 18:41
  SPEAKER_00: You can interact with all of those models, and your data remains safe and secure.
• 18:49
  SPEAKER_01: Gotcha.
• 18:50
  SPEAKER_01: So, I mean, I I kind of gotcha.
• 18:54
  SPEAKER_01: That that maybe maybe maybe you can explain what that means, maybe in more lay person English, because I I already can't sleep at night about some of the security issues.
• 19:05
  SPEAKER_01: A lot of people can't.
• 19:07
  SPEAKER_01: And I want to be able to understand those nightmares a little bit better.
• 19:11
  SPEAKER_00: Yeah.
• 19:12
  SPEAKER_01: So, how how can how can you make that more understandable for me?
• 19:16
  SPEAKER_00: So um my daughter can explain this much better than I can in all the technical lingo, but the way she explains it in ways that I think most people understand, it's like when you put your credit card into one of those shredders and you get all those little bits and pieces of the credit card, right?
• 19:35
  SPEAKER_00: The large language, the interface to the large language model figures out which pieces of the credit card it needs to share with the large language model and does it in a mathematically, well, it's vectorized, but it's it's a mathematical formulation, military grade.
• 19:57
  SPEAKER_00: You know, I think I think they said 1048 um or 1024 uh bit encryption.
• 20:06
  SPEAKER_01: Okay.
• 20:06
  SPEAKER_00: Um, so so it's really encrypted, and only the pieces that are required get shared with the large language models.
• 20:16
  SPEAKER_00: So there's no way for Humpty Dumpty to put that back together again.
• 20:20
  SPEAKER_00: The danger in taking information to the large language model is not that it understands, because it doesn't understand anything, but it's looking for frequency of words and which words go together.
• 20:34
  SPEAKER_00: And if so, if you break this up into little pieces, they can't put the words back together.
• 20:40
  SPEAKER_00: In other words, there's no way to put Humpty Dumpty back together again.
• 20:45
  SPEAKER_00: Does that make sense?
• 20:47
  SPEAKER_01: Gotcha.
• 20:48
  SPEAKER_00: So it's the what bits and pieces, little pieces of Humpty Dumpty shell, but you can't create a whole shell because you don't have all of the pieces.
• 20:58
  SPEAKER_00: And furthermore, you don't know which one goes where.
• 21:02
  SPEAKER_01: I mean, it's that sounds like a lot like a tabletop puzzle where all the pieces are the same shape or something, something evil like that.
• 21:13
  SPEAKER_01: Uh, but your but your uh but the encryption software is able to track each and every piece of that data so it can understand and parse out where all those pieces ought to be and can reinterpret uh everything.
• 21:31
  SPEAKER_01: So okay, I think that makes sense.
• 21:35
  SPEAKER_01: Boy, I mean, frankly, that's a lot for for me.
• 21:39
  SPEAKER_00: It it it it is.
• 21:40
  SPEAKER_00: It's it's it's very complicated mathematics.
• 21:43
  SPEAKER_00: There's a patent pending on the on the platform, on the technology.
• 21:48
  SPEAKER_00: Um that you know, it's again, it's it's a special kind of encryption uh that that is, you know, really keeps your data.
• 21:58
  SPEAKER_00: In fact, if if you have
• 22:00
  SPEAKER_01: account on our platform i can't see your data either right it's it's only you know people in that account that can see it and the the the large language model interface only puts those pieces up there that are absolutely required for the for the prompt for the the issue at hand right okay and hopefully everybody got that the i'm usually one of the the last people on that adoption curve for anything technological i'm trying not to be that guy this time uh on on this particular wave of technological improvement because everybody always says when something changes like this that's a real ground shift there will be people left behind there will be companies left behind that aren't taking advantage of the new tool etc and i think this is certainly one that's going to be a a tectonic shift that creates a divide between hopefully us on the positive side uh and and maybe people who get left behind so what would tell you what are some of the flags that would say that a company possibly our company for example that is that it is using AI leverage well um so so I think one of the easiest things um is you talked about it earlier revenue per employee um if you double the output of every employee it seems like the revenue ought to go up maybe it's not double but it certainly should um go up dramatically um in fact you could imagine that if you're doing those higher level tasks that it should go up even more than twice right the the things that are you know the next level of task um yeah the response time um it it really depends on the business itself but um looking at um the ability to handle multiple tasks at the same time um because again you have you have an AI agent that can work for you while you're working on something else right and then you you you know feed it that information and you come back to it and gather that um always have to have a human in the loop you can never just take the what's what the AI puts out and just say it's done but but think about what would happen if every employee who's an expert in what they do had world class experts advising them and taking their game to the next level that's an easy way to think about what you can do with AI.
• 25:14
  SPEAKER_00: You have all that stuff at your fingertips.
• 25:17
  SPEAKER_01: That's absolutely true the uh I I used an AI tool to help kind of organize what we're talking about today based on some of some of the discussion you and I had in in prep for this and is it is it better is the result better than what I would have put together myself I don't think so but is it similar is it the same quality and did I do it a lot faster?
• 25:43
  SPEAKER_00: Yes so I I absolutely this is making something that I that I really enjoy doing the podcast here more doable whilst I'm doing my normal job uh on a day-to-day basis because it doesn't it doesn't create that time suck that I well everybody's afraid of but speaking of sorry I just want to jump in here because we did an experiment with a law firm and a paralegal an overseas paralegal um the paralegal is was uh is a lawyer a trained lawyer just not part of the bar in the state where the where where the the law firm was right in the in the Philippines and the result of that you know experiment the data that we got was yeah she was 40 to 60 percent faster in doing stuff but the quality was better and she asked the lawyer fewer questions uh along what dimension or what what told you that the quality was better I mean how how did how did you measure that the lawyer said it took him less time to review it was closer it was closer because the paralegal is preparing legal documents for the lawyer to then you know pass off to the client right the lawyer has to review those because he's gonna sign them um he's going to say this is from our law firm so he has to review it's a he in this case right he had to review all of those things and it's he said it took him less time so the quality that that sounds like and getting it from a a yeah uh line staff to that that high the highest level of you know approval by the the the signer the person who's uh responsible that's that is truly important that sounds like uh something I was reading the other day about uh measuring the progress of in particular llms uh about the time to translate that one of the ways to measure that is if you ask an LLM to translate from one language to another and you have a native speaker come along and edit it I think it maybe it was time to edit I can't recall but time to edit uh the better and better the AI is the less correction and the the less time it'll take to correct a document prepared by that or translated by that AI by that native speaker.
• 28:26
  SPEAKER_00: And so at the end of the day uh that sounds a lot like what that is I mean because frankly law law is pretty uh opaque for for a lot of us I and by the way just as a total aside I think it's really interesting we could probably do a whole episode on on this but that that business model where there's a paralegal offshore supporting a US based lawyer I think is super interesting and I I really would want to know more about that and and the reason why I think it's super interesting is most of my lawyer friends what when I talk about that kind of stuff they they just cringe and say that's an that's an impossible model that doesn't work for you know whatever very various various reasons and they're not completely wrong at all I mean it's their their business they work but also I think it's interesting that this person did it though yeah that was and and she's been working for this law firm for several years this is not a brand new paralegal that we just spun up I mean this was comparison fair for you know a number of years of her work and now we trained her in ai and it stepped up a level so makes sense proves the efficacy of at least in that sense get getting a little bit practical for a second any kind of change in a business in a in a process a way of thinking just operationally that can be risky right if you change something you're going from something known to something lesser known so how do you suggest mitigating that risk in uh implementing AI enabled procedures you know again without putting your whole business at risk at the same time yeah that that's that's a really good question it's really important um and I work with all of my clients to develop a clear AI policy these are the things that you can and can't do with AI this is the kind of data that you can and can cannot upload um these are the kinds of questions um this is how we mitigate the risk of hallucinations all of those things right hallucinations is when the AI goes off the rails and starts making stuff up and it does that they say chat GPT5 is better than the predecessors I don't know I haven't worked with it enough to tell but but it's still it's not zero right um so so mitigating those risks means you know taking into account what's the perspective of the of the LLM what was it what data was it trained on um what was the post training like because that impacts what what happens um and really you know again this is something we work with our clients to understand what's the best kind of what model to use for the different kinds of things that you're trying to do there are so many choices out there.
• 31:37
  SPEAKER_00: I definitely understand the the training piece and I think people have used that terminology quite a bit so hopefully you know loading things into uh the AI so that it can learn off of your information again you absolutely you at least need to have a paid account to even start thinking about that but you said post uh what did you say post chat um post-training post-training yeah yeah so so these models are you know when I talk about AI training I'm not talking about what you and I do in terms of loading our information I'm talking about the basics of the way the model is trained on all of the data in the internet right okay after they do that then there's a wrapper put around it and and for those of you who understand this kind of thing I'll just give you a quick example right maybe you don't understand it but here's a quick example is you know chat GPT and Microsoft Copilot are the same LLM they're the same large language model there's a a post-training interface that goes on top of that so that if you fed the same prompt into both models both chat GPT and to Microsoft Copilot you would get a different response and it's because of that post training.
• 33:00
  SPEAKER_00: Another example the post training comes from the interface well the the the the the company that's selling you the LLM or or licensing the LLM to you does training and and and they spend a lot of time trying to fine tune that I don't know if you remember a a while ago there was a a time when a a release got um too too syncophatic right they were it was too uh too sympathetic and and too um compromising in terms of you know all of the description the conversation that it had it was just over the top sort of telling you everything was fine everything was okay you were doing great all of that kind of stuff it was just it was too far that way and they had to retrain it.
• 33:52
  SPEAKER_00: They had to so that's that's part of that post-training and they do a lot of work to test these models and to tune that before they release it to the public that's called post training um and every model has a perspective i'll I'll give you a a concrete example at the risk of being somewhat political um grok grok four early in its release when you asked it a political question it checked to see what Elon Musk would say in his Twitter feeds about that issue and give you Elon Musk's answer because that was the way it had been trained.
• 34:40
  SPEAKER_00: Now whether you agree with Elon Musk or not he has a version of truth other people have a different version of truth and you need to be aware that there are people working on these models that um influence how they respond.
• 35:00
  SPEAKER_00: And I would imagine that's that's true across the board I just think of that as a specific example right oh sure it because everybody has their own context of how how their knowledge and wisdom and and and everything have have developed over time.
• 35:16
  SPEAKER_01: And so I you know it's it would be hard for me to think like somebody who's very who's had a very different say upbringing and career than I have for example and that would absolutely affect how I might subconsciously program something or create a system again is why it's really great to have mostful viewpoints on on something.
• 35:37
  SPEAKER_00: And it's not necessarily intentional but in this particular case you could watch the reasoning of the grok as it as it was saying you know I'm gonna go check this and I'm gonna go look at this and I'm gonna go check this and I'm gonna ask Elon Musk and you know and people went oh that's interesting um but but it just you know that's the way that one got and they pulled that right they changed that but but you just have to be aware that that's uh a danger and you have to be responsible you have to check the references most of these things will now put out you know here's the website that this information came from right and you need to check that that's a real website and that really talks about that thing and it talks about it in that way right it's not uh um just you know take it at face value you can't do that oh I I certainly liken it to when Wikipedia first came out back in the day where it was really good super helpful saved you a trip to the library or at least one you know the first trip to the library if you will doing research and also you had to be careful so you couldn't just rely on it and certainly nobody would let you just cite it in your in your paper of you know whatever kind so this is this is you know an another leap forward in that same direction and has still has some of the same issues and yet it's still better net net yeah okay so just real quick I mean at at the end here what what is one tool or habit that gave you the the best bang for your buck I mean I hate to say a hack but you know what what is one thing that has been transformative for you that AI brings to you brings to you and and I would say be specific if you can sure um one of the things that we do in our quarterly planning we we have a strategy three year strategy where we're going and um we ask the AI to give us each of us as team members OKRs objectives and key results for the quarter based on that strategy so now everybody has things that they need to work on that line up with that strategy and we track those things throughout the quarter so everybody has projects and tasks that are not just sort of the day-to-day grind but they go after that longer term strategy and is that is that something where you because that's strategic level thinking right is that something where you're giving the AI quite a bit of guidance first or kind of using it as a just a a well not just but as as a as a creative thought creative is always a thought partner it is always a you know hey let's think about this um you know and and the the results need to be double checked by the way you know the first couple of times we did this there were conflicts between this OKR and this OKR with different goals and you know so you yes you can't just again you can't just take it at face value but it's a place to start absolutely okay almost always these things are this is a this is an approach this is a place to start I think my favorite hack if you will for using AI is if you don't know how AI can help you ask it so you have a problem you're working on you you have a problem you're working on and you're not you think AI might be able to help give me three ways that AI could help me with this problem and then look at them right right I and I keep hearing that as kind of the the you know in a certain sense the the ultimate hack or the ultimate just simple place to start for everyone again even using uh chat GPT ends up being what what I tend to use but even using that I don't I don't know do they have instructions out there because you basically just ask it how do you use it?
• 40:17
  SPEAKER_00: Yeah give me yeah it's it's a it's a meta thing that we would never I would never say hey Steve how can you help me solve this problem I would never do that but you can do that with chat GPT or or your large language model of choice and it will happily answer.
• 40:32
  SPEAKER_01: Now uh there's there's there's there's a lot more nuance to it than that but but essentially give me three give me three choices three different ways you could help right right so there this is the umpteenth AI podcast out there and that's just talking about AI let alone the fact that there's you know these these these promises in the news that AI will do all these positive things and some negative all these things and on top of that it's the models themselves are multiple multiple multiple models there's so many platforms and even among those there's different iterations that keep moving different versions that keep improving it's really overwhelming even in me saying all these things I get overwhelmed how does somebody not be overwhelmed you know resist the tide of the over AI overwhelm well I think there's a couple of things one is you could probably use a guide just saying um be happy to talk to any of the listeners about you know where do they want to start what do they want to do but start small try try something I mean lots of people use it for marketing and email and that kind of stuff I'm telling you that stuff the challenge is you can almost always tell when AI wrote it.
• 42:13
  SPEAKER_00: So you need to make sure that you rewrite it with your own voice.
• 42:17
  SPEAKER_00: Right but but that's an easy place to start um but um there are and you're right there are so many tools how do you pick one well you know what again one of the advantages that we have is we are completely platform agnostic and the things that we teach the things that we help people understand they can use in whatever environment they are they are comfortable in whatever one they like um so I would I would just say start start small start with a task remember that context is everything if you want the right kind of output from your AI you need to make sure it has proper kind of background information on you on your problem on your company on what you're trying to do and the more context you give it the better the response that makes sense well that that's how you often need to be communicating with your staff too I don't just have you know a formulaic problem most of the time most of the time it's a problem within a context or a client has a a way that they're thinking about things.
• 43:42
  SPEAKER_01: Whether or not that's how we're gonna think about it is a different question, but we have to know where they're coming from in order to not just perform the the project or perform the work but also to really understand how it fits into the whole solution set of what they're looking for.
• 43:59
  SPEAKER_01: So it That that absolutely makes complete sense to me that you need to tell the AI this is the type of answer I need, not just what is the answer, please.
• 44:10
  unknown: Right.
• 44:11
  SPEAKER_00: And and never ever ask it for the answer.
• 44:16
  SPEAKER_00: I always ask for two or three options or five options.
• 44:23
  SPEAKER_00: Or, you know, give me give me your three best.
• 44:29
  SPEAKER_00: Right.
• 44:30
  SPEAKER_01: Yeah.
• 44:31
  SPEAKER_01: I certainly don't want to have it just say 42 right back at you.
• 44:34
  SPEAKER_01: That's that's not helpful.
• 44:37
  SPEAKER_01: So before we wrap, uh, let's just do a couple of quick personal questions.
• 44:42
  SPEAKER_01: I always like to uh get and get to know you a bit.
• 44:45
  SPEAKER_01: Uh so tell tell me something that's not on your LinkedIn or your website profile.
• 44:50
  SPEAKER_00: Well, I'm gonna cheat a little bit here because this is mentioned on my LinkedIn profile, but not to the extent that it's true, right?
• 44:58
  SPEAKER_00: I sing a cappella.
• 45:00
  SPEAKER_00: Um, and I sing in a very, very fine men's a cappella chorus out of San Francisco called the Fog City Singers.
• 45:10
  SPEAKER_00: Um, it's a world-class a cappella chorus.
• 45:13
  SPEAKER_00: I know it's world class because we just competed at the international competition of the Barbershop Harmony Society.
• 45:20
  SPEAKER_00: And out of, well, there were 38 choruses that qualified to be there, we finished fourth.
• 45:29
  SPEAKER_01: Excellent.
• 45:30
  SPEAKER_00: In the world.
• 45:31
  SPEAKER_01: Congratulations.
• 45:32
  SPEAKER_00: So it's pretty good.
• 45:34
  SPEAKER_00: People want to hear about it or hear us perform.
• 45:38
  SPEAKER_00: Go to YouTube and look for Fog City Singers.
• 45:42
  SPEAKER_01: Okay.
• 45:43
  SPEAKER_01: I that might be pretty interesting to check out.
• 45:46
  SPEAKER_01: I and I will.
• 45:47
  SPEAKER_01: Uh, you definitely do have an excellent voice for podcasting, I will say that.
• 45:51
  SPEAKER_01: Uh I know that from personal experience.
• 45:54
  SPEAKER_01: Uh okay.
• 45:56
  SPEAKER_01: So what is it about in your non-singing daily life?
• 46:04
  SPEAKER_01: What is it about what you do that's rewarding for you?
• 46:09
  SPEAKER_00: I like helping people.
• 46:12
  SPEAKER_00: I like just I love learning and I love helping people figure out how to do new and challenging things.
• 46:22
  SPEAKER_00: Um I spent you know 30 years in high tech figuring out how to solve problems, both for myself, for my teams.
• 46:31
  SPEAKER_00: I became a known as a manager who would come in and stabilize organizations because I care about people more than well, there are some physicists who are not as um personable, I'll say.
• 46:49
  SPEAKER_00: People skills.
• 46:51
  SPEAKER_00: So um I love helping people.
• 46:53
  SPEAKER_00: Yes, they're very technically oriented, very, very, very, very fine people.
• 46:57
  SPEAKER_00: Um, just the soft skills, not always there, and that's okay.
• 47:02
  SPEAKER_01: We're we're in Silicon Valley.
• 47:03
  SPEAKER_01: I don't think we have any of those kind of people here.
• 47:06
  SPEAKER_00: Nope.
• 47:07
  SPEAKER_00: Uh anyway, I I love helping people, I love helping them figure things out, I love solving problems.
• 47:14
  SPEAKER_00: Um, it's it's it's just a lot of fun.
• 47:21
  SPEAKER_01: I do tend to find that the highest value stuff, the highest value tasks and actions that we take often tend to be fun.
• 47:29
  SPEAKER_01: Because if we're high performing people, I'll just use that, man.
• 47:34
  SPEAKER_01: That sounded like a uh ego stroke there.
• 47:36
  SPEAKER_01: But but really, if if we're out there doing things that are valuable in in society, generally speaking, there's a there's a lot of return you get from it, monetary and non-monetary.
• 47:50
  SPEAKER_01: And gosh, it ends up being fun.
• 47:53
  SPEAKER_01: Uh, you you you just mentioned this a little bit, the the learning aspect.
• 47:58
  SPEAKER_01: Uh you know uh we're always on a mission to improve ourselves.
• 48:04
  SPEAKER_01: What is it that you're reading or learning right now?
• 48:09
  SPEAKER_00: You know, I knew you were gonna ask this, and I meant to look up the name of the book.
• 48:13
  SPEAKER_00: Um I'm reading, I'm I'm I'm listening to a new Audible.
• 48:17
  SPEAKER_00: Uh for me, it's and I want to get the title correct.
• 48:21
  SPEAKER_00: It's called The Ruthless Elimination of Hurry.
• 48:27
  SPEAKER_00: Okay.
• 48:28
  SPEAKER_00: Um yeah.
• 48:32
  SPEAKER_00: Uh so it's about how to slow down a little bit.
• 48:39
  SPEAKER_00: We're we're you know, everybody, when you ask them how are they doing, what do they answer?
• 48:46
  SPEAKER_00: Good.
• 48:47
  SPEAKER_00: Busy, but good.
• 48:50
  SPEAKER_00: That's the most common answer that I hear.
• 48:52
  SPEAKER_00: It's probably the most common answer that I give, right?
• 48:55
  SPEAKER_00: And one of the questions he asks in the book, I'm not all the way through it yet.
• 48:59
  SPEAKER_00: Um, but are you constantly when you pull up to a to a red light, are you constantly looking for which lane is has the fewest cars in it and pull into that lane?
• 49:11
  SPEAKER_00: You're in such a hurry that one or two car lengths makes a difference.
• 49:18
  SPEAKER_00: Anyway, fascinating book.
• 49:20
  SPEAKER_00: I'm not through it yet, but but I'm I'm trying to figure out how to slow down.
• 49:25
  SPEAKER_01: Um is it kind of a mindfulness thing?
• 49:30
  SPEAKER_00: Well, no, not the way he's approaching it, at least not yet.
• 49:34
  SPEAKER_00: I mean, I haven't gotten to solution space yet.
• 49:37
  SPEAKER_00: Um, but um, but mindfulness I'm I'm sure will play a role.
• 49:42
  SPEAKER_00: Um, I'm I also, you know, I I am also a meditator and I meditate every morning.
• 49:48
  SPEAKER_00: Um and and uh I think I've missed three mornings since February of 2020.
• 49:58
  SPEAKER_01: What was going on then?
• 50:00
  SPEAKER_01: I'm kidding.
• 50:01
  SPEAKER_01: The uh uh hey, you should put meditation on on your LinkedIn.
• 50:06
  SPEAKER_01: The uh that that that can that could give people a little bit of context around who Dr.
• 50:10
  SPEAKER_01: Steve is.
• 50:11
  SPEAKER_01: Yeah, um okay.
• 50:13
  SPEAKER_01: Speaking of though, Dr.
• 50:14
  SPEAKER_01: Steve, where can people find you if you want to be found?
• 50:18
  SPEAKER_00: Oh yes, absolutely.
• 50:20
  SPEAKER_00: Best place is uh LinkedIn.
• 50:24
  SPEAKER_00: Um reach out.
• 50:26
  SPEAKER_00: Um we'll put the link in the show notes.
• 50:29
  SPEAKER_00: Um and uh the other the other thing is on the website, profitminds.net.
• 50:35
  SPEAKER_00: You can go there.
• 50:36
  SPEAKER_00: There's a place to ask a question if you want.
• 50:39
  SPEAKER_00: Uh there are some quizzes you can take.
• 50:41
  SPEAKER_00: Um if people want, um I have a free assessment, AI readiness assessment for your business that uh I'd be happy to provide for those.
• 50:57
  SPEAKER_01: And is that is that something that uh somebody can just go to the website and download or take or request or behind a email wall?
• 51:05
  SPEAKER_00: I'll give we'll give you a link for the show notes.
• 51:07
  SPEAKER_01: Perfect.
• 51:08
  unknown: Okay.
• 51:09
  SPEAKER_01: Well, awesome.
• 51:10
  SPEAKER_01: Uh you know, is there anything else that you'd want to share about Profit Minds, about you?
• 51:16
  SPEAKER_01: Uh of course, your your podcast, which is out there, uh and well, anything.
• 51:22
  SPEAKER_00: Yeah, so podcast is the Profit Minds podcast, real original title.
• 51:27
  SPEAKER_00: Um, but it's what makes unique Y-O-U-N-I-Q-U-E.
• 51:31
  SPEAKER_00: Um, we look for interesting, unique individuals and uh people like yourself, Ben.
• 51:37
  SPEAKER_00: Glad to have you on the show.
• 51:38
  SPEAKER_01: Somehow I made the made the cut.
• 51:40
  SPEAKER_01: I don't know how.
• 51:41
  SPEAKER_00: Yeah.
• 51:41
  SPEAKER_01: Um, but we'll have we'll we will have links to all that in the uh show notes as you as you said.
• 51:46
  SPEAKER_01: Uh thank you so much, Steve, for joining us.
• 51:51
  SPEAKER_01: And thank you for all the listeners for tuning in.
• 51:54
  SPEAKER_00: Thank you, Ben.
• 51:55
  SPEAKER_00: Listeners, thank you for tuning in.
• 51:57
  SPEAKER_00: This has been an episode of Max Value.
• 51:59
  SPEAKER_00: Find us on YouTube, Apple Podcasts, and other streaming services.
• 52:03
  SPEAKER_00: Please like, subscribe, and follow.
• 52:05
  SPEAKER_00: And we'll see you for the next show.