Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
The Currency of Trust: Building an AI Consultancy on Practicality with Myles Harrison
32 Plays1 month ago
AI is everywhere.
Execution isn’t.
We sat down with Myles Harrison to talk about what works and why most companies are lighting money on fire pretending they “have an AI strategy.”
We get into:
- Why bad AI consulting feels like getting your watch stolen then sold back to you
- What has to be true before AI does anything useful
- How to build workflows that actually perform (not just look good in a deck)
And also… serious question:
Has Mark Zuckerberg replaced himself with a very calm, very consistent AI agent?
If your AI isn’t driving outcomes, it’s not strategy. It’s cosplay.
Recommended
Transcript
Introduction to Miles Harrison and PractiKai
00:00:09
Speaker
and this so of frequency we are super fired up to chat with Miles Harrison, founder and principal of PractiKai, an AI consultancy that is focused on real-world adoption. And so we'll find out the story behind PractiKai and get into what actually works.
00:00:32
Speaker
as people are navigating this new age of technology.
Building Human-AI Workflows
00:00:36
Speaker
and We'll talk about where you should start, what needs to be true first, and how to build a human and AI workflow that delivers real impact.
00:00:46
Speaker
Welcome, Myles.
00:00:49
Speaker
Thanks. Great to be here. Awesome. tell us, um what does Practicai mean? um and tell us a little bit about how you got started.
00:01:02
Speaker
Yeah, certainly. So it's not practic AI, which some people think it is. Practicae is a Greek word ah that means practicality or putting things into practice.
00:01:17
Speaker
So it's similar in the root to like the English word practicum that people do in in school.
Pragmatism in AI Adoption
00:01:23
Speaker
um And I really adopted this just because I'm sure probably aware and most listeners are aware if you read the news, there's a lot of hype and confusion around ai right now. We're still very much in the hype cycle, definitely for agentive AI, agentic AI, excuse me, as they call it.
00:01:45
Speaker
um And so, you know, I joke that my team were kind of like AI hipsters. We were doing AI before it was cool. ah It's hard sometimes for me to convince people that we are are data scientists and that we were doing ai before it was the new ai um But we see the work that we do driven by pragmatism and practicality and you know kind of an antidote to a lot of the hype and confusion that's going around. um And I'm sure we'll talk about that more.
00:02:19
Speaker
And then in terms of how I got started, i have a background In the big five, kind of a career consultant, I started in the agency world, Publesys, in digital analytics.
Miles Harrison's Background and Entrepreneurship
00:02:33
Speaker
And I was kind of lucky enough to ride a different hype cycle of of big data into to AI and machine learning.
00:02:42
Speaker
um And then I also worked in analytics at PricewaterhouseCoopers, PwC, and at Accenture. um and then I briefly took some time away from that to to teach and that as well.
00:02:58
Speaker
um But really i started freelancing because I've reached a point in my career where I i felt that i know the consulting world well enough that it's something I can set out and do on my own and it's also something I want to do differently because I've seen good parts and the bad parts of of the world of consulting. um And so a lot of what we're trying to do with with my team is set ourselves apart from that. And that was really the motivation for me setting out and and surprisingly finding myself being an entrepreneur. Although you know, the past me, probably never thought this is something I would do.
00:03:40
Speaker
That's awesome. Interesting.
Common Mistakes in AI Integration
00:03:43
Speaker
You have, we were talking earlier, you've worked across so many different industries in your consulting practice, manufacturing, transportation, retail.
00:03:52
Speaker
And tell us a little bit about the themes that you see across those sectors. What would you say is um maybe the most common or most expensive mistake that's happening in organizations or across organizations trying to integrate AI right now? Yeah, well, the the most expensive mistake ah in trying to integrate ai across a lot of these disciplines. And i like to say to folks that ah that don't know Practicae, you can tell these are very important disciplines because they are the least sexy disciplines that work with real objects in the physical world.
00:04:32
Speaker
Things like supply chain and and trade and things like that. We do a lot of work in manufacturing. But that typically means that they're also not very good at at technology, and so they're the folks we can work with and help.
00:04:46
Speaker
um I guess the worst mistake would be it's difficult to know If you don't know about a domain and and someone starts talking to you, it's difficult to know whether to believe them, right?
The Role of Quality Data in AI
00:04:59
Speaker
And whether they actually are offering you what ah what uh what they're claiming and this is something i have an issue with in the world of sales and business and i think one of the ways that we try to differentiate ourselves is one of the other core values we have is clarity um and i spent a lot of time teaching and i believe that if you can't explain something very technical to someone that is not technical and have them understand it then you don't really understand
00:05:31
Speaker
what you're doing right and that should be part of the sales process and helping businesses problems as well but assuming that you don't you know buy a bridge that someone is trying to sell you uh i think the the number one mistake that a lot of organizations make is getting ahead of themselves and i'm guilty of this too because i'm pretty impatient but you cannot make a house you know without bricks and you cannot do the ai without data, right? And most organizations, I don't believe all this stuff that says that like white collar work and programming and all this is gonna go away and be replaced by AI because there's so much complexity and technical debt ah and the world is so complex that most organizations
00:06:23
Speaker
They need to walk before they can run and they need to step back and take a good look at what they have to understand what their real business problems are first and whether their house is in kind of disarray or complete disarray before they start trying to do AI.
00:06:41
Speaker
Because data is the raw material that machine learning uses. And if you don't have it's garbage in, garbage out. If you don't have the right material, you can't build.
00:06:52
Speaker
so it's not ah It's not the sexiest sales pitch as I say to people, but sometimes I talk to potential clients and they say, we want to do AI and they're very excited.
00:07:03
Speaker
And you know it it's not exactly the most salesy thing to be a little bit of a wet blanket and say, okay, you know I know you're excited, but like what are you trying to do? And like who told you this word? And what makes you think that it's what you need right now?
00:07:17
Speaker
And then if you talk to them, like you you start to learn that it's like, okay, they they know that AI can do these things and they want that done. But in some cases, they're not ready for that.
00:07:28
Speaker
So it's like, we need to get your data in order first. Or ah there may be a simpler way to do what they're trying to do with that that it's not really an AI problem, right? But I think it's like,
00:07:41
Speaker
It's not fun work, but you know, you have to kind of eat your vegetables. I mean, I personally don't, but um you have to eat your vegetables ah and and you have to kind of get your house in order in terms of data and data quality and having all the information and being ready to do that before, you know, you go about doing it.
00:08:07
Speaker
Makes a lot of sense having to lay some foundations and do the grunt work before you can do the really sexy, exciting part of things.
Critique of Consulting Practices
00:08:17
Speaker
And when when we first spoke, Miles, you you were talking a little bit about the market of consultants or experts interacting with prospective partners or clients like these.
00:08:30
Speaker
And we wrote down what you said because it really stuck with us, and and that was They stole our watch, told us the time, and then charged us $45,000 for the privilege.
00:08:42
Speaker
So tell us a little bit about what does that refer to in the context of organizations seeking consulting advice? Yeah. So, I mean, $45,000 you're lucky, right?
00:08:56
Speaker
Time is expensive and and it it people's time is money, but I mean, I didn't come up with that saying. That's an old saying about management consultants, right? Steal your watch and tell you the time.
00:09:10
Speaker
ah You know, the the one of the kind of anti-heroes of consulting, think it's Marty Kinn, you know, wrote that book called House of Lies about the business of management consulting and his issues with it and had similar challenges. i think they made that into a show on HBO with Don Cheadle.
00:09:29
Speaker
But yeah, I mean, the problem is that
Challenges in Management Consulting
00:09:34
Speaker
and And I think it can be good and it can be bad, but I feel that because of people that do consulting badly and position it badly, ah that management consulting, whether it be in data and AI or whether it be in strategy or whether it be in marketing or any discipline has a bad name um because you're an outsider and then there's issues of trust and then people will come in and they say they'll deliver one thing
00:10:04
Speaker
and then they'll deliver something else or they'll come in and you're an outsider. You always have to do discovery work. We still have to do discovery because every client is different. As I like to politely and positively say, every business is different. They're all uniquely messed up in different ways.
00:10:20
Speaker
um So you do have to go up, you you do have to go in and do some discovery. But I think the problem is that why people get this feeling is hindsight is 20-20, right? like consultants come in they say, we want to help, we're going to fix these problems. But in order to fix the problems, we need to do, you know, some detective work and understand what the problems are. And then the temptation, especially for clients that I would say, don't do really good work.
00:10:47
Speaker
That's all they'll do. and they'll say, here's your 128 page deck. That's really expensive. And the president of the company, especially if it's, not a big enterprise. If it's somebody like some of our clients who, you know, wear a plaid shirt, built business in their garage and have run it for 25 years as the president of the company, they're just like, you told me stuff I already knew and then you just charge me for right?
00:11:14
Speaker
I think about that a lot because I try to have empathy and put myself in my client's position And it's difficult for me because I have to say, look, we have to come in and talk to your people and take their time. And our time is expensive too, but there should always be something after the discovery part, which actually is the execution.
00:11:38
Speaker
And even if our clients say, Hey, you don't have the budget or you did the discovery and things have changed. Like I talked to my team and push them very, very hard. And I say like,
00:11:48
Speaker
This is not an exercise of just gathering information and repeating it back to them. It's like, you need to look at what is there, tell them something they don't know, and then also tell them something.
00:11:59
Speaker
and this this word is misused so much that become a cliche, but I mean it this time. And tell them something that is actionable, right? Like not just like, hey, this is the state of your business.
00:12:11
Speaker
and this is things that are broken, the CEO is going to be like, I'm well aware, I fight these fires every day. You want to say, this is the state of your business. This is broken. This is our perspective.
00:12:22
Speaker
And this is what we think you should do. And this is how you should do it. And we will help you if you want our help. And if you don't, this is how your team would do it. I have a we partnering with a couple companies.
00:12:35
Speaker
a couple other small startups very closely and I guess a little embarrassing because he's not here but Ryan had this thing he said on a call to me that might be bad sales but I believe it very much so as well. e He said to a client on a call, he said, look, he said, I want to work with you and get you to a point where you don't need us anymore, but where you want to work with us because we have fixed your problems to the point and like upskilled your team to the point that you are self-sufficient, but we provide enough value that you want to keep working with us because you see the value, right? As opposed to, I'm going to go in, you know, do some work. Drag it out so that I can invite you more and more.
00:13:22
Speaker
Rack up the billable hours and then be like, okay, the client's happy. The project is done. We didn't get sued. know Ride off into the sunset in your boat or whatever. i am oh larryie aon just real quick i had that same feeling when we set up is that i've seen it so many times the big firms coming in charging not $45,000, that's for sure. And just coming in and providing a million decks and nothing, no way to move forward. And I was like, people can't like that.
00:14:03
Speaker
People can't be like, yeah, this is the way for that. But I saw them making that same choice over and over again. thank really they want impact. They want action, they want solutions, they want somebody to go in and like roll up their sleeves and figure out what's really going on
Client Expectations from Consultants
00:14:20
Speaker
under the hood.
00:14:21
Speaker
And i think the I definitely feel like people know the value now because everything has changed. But there was crazy money flying out of big companies or any company. to get to know results. I was just amazed how how that industry like stood up for so long. because seeing it and Of course you feel 10 times worse if you're in in the company because they would suggest things that you might have suggested six months ago.
00:14:52
Speaker
But because the consultant said it, then it's something that we might do. You're just like, you're like a lot. I've heard that before as well, right?
00:15:02
Speaker
Yeah. The other thing I want pick your brain about, Miles, and that is aside from the walk before you run concept and companies getting over their skis around AI in general, what is your perspective or what are some of your observations around how companies are addressing their people with respect
AI's Impact on Employee Management
00:15:26
Speaker
to AI? So their employee population, what is the industry getting wrong there? Yeah.
00:15:32
Speaker
um i have I have an opinion on this and I have observations. um And I'll just say that. ah It's my my opinion, what I've seen and our perspective from from my company.
00:15:47
Speaker
So it will depend on, I guess, your management style. and whether or not you've read The Prince and whether or not you believe it's better to be feared than loved ah or not.
00:15:59
Speaker
um But people is is ah your your business, so this would be of interest to you. But I see, and this has been well published in a lot of newspapers, um there's so much momentum around AI and so much focus on it that you see organizations
00:16:22
Speaker
I would say taking more of a stick approach than a carrot approach to AI. um Things like tech companies saying, you must use AI and we will, you know, obviously we can track that and then that's somehow tied to your performance review or something, right? um Or saying, you know, I was literally just talking about this with someone, there's this you You consume different parts of language and your metered utility that you're using with these AI companies is tokens, right?
00:16:57
Speaker
And so there are these companies and they have this whole thing called like token maxing, where it's like how many tokens, you know, who is burning the most tokens, billions of tokens. And it's a total quantity over quality thing.
00:17:09
Speaker
um Because just like in the old days, as somebody was saying to me, ah it used to be like, how many lines of code did you write? it's like and And there was a time, I remember watching this this documentary, Computer Science, my computer science course, about how they used to pay programmers by the line of code, right?
00:17:29
Speaker
Which, when you think about it, is really illogical. Because that means if you code a very inefficient solution, you get paid more, right? So we're seeing that kind of thing now where people are getting lost in this quantity over quality.
00:17:41
Speaker
And i I am a big believer in the, like, both the carrot approach for all kinds of things in organizations. Like don't, we're all on the, your your employees should be all on the same team, right?
00:17:56
Speaker
So you should be going, like someone used this term, I don't know if this is common in your domain, but about like, they use it as a cliche as well, but people that actually do it, this this idea of of servile leadership, where the the the purpose of a leader is to serve their employees Right.
00:18:15
Speaker
They should be the one. Right. And I've met people in positions of leadership in government, too. And they go to a forum. You think they're ask some questions. And the they they said to me, yeah, I'm here to take marching orders.
00:18:27
Speaker
Right. Because I'm in charge and I have to make changes. um And I think it that should be on the leaders to to take the carrot approach and say, look, you know, everybody has challenges in their job.
00:18:39
Speaker
Everybody has parts of their jobs they don't like. If you can use AI to be more effective and it's a useful tool and it can make your life easier. Here it is. Oh, I don't believe you. I'm anti-AI. You know, i am too busy already.
00:18:54
Speaker
OK, we're going to do training. Let me help. Help me understand the parts of your job that suck. And these are the things that I can do. And like, you know, don't don't dictate to your people. Hey, here's this new tool. You must use it.
00:19:11
Speaker
Right. yeah Show your people how the things that are wrong with the tool and the things that are good about the tool and say, this is how we are going to empower you. Right. Or your work is not going to be it's not going to make everything all better and work will be different.
00:19:28
Speaker
But it is a lot. You can do a lot now much, much, much faster ah than you ever could. And I think some people don't know that. And I think telling people, hey, here's this thing, use it.
00:19:40
Speaker
It's not empowering. You need to take the time or spend the money or both to to make, to incentivize people and show them that that it it is something that they choose.
00:19:53
Speaker
And then if it's not, then, you know, don't give them a bad performance review just because they don't use a certain tool, right? I know the measuring of usage is just crazy to me i i think it it's exposing some leadership gap when this conversation happens because everyone's saying that they're implementing AI or, you know, I'm getting up to speed with ai
00:20:24
Speaker
But if you try and like click into that, it's pretty pretty easy to see the people who have like no idea like what the direction is that they want to go down. Whereas if you can have a conversation with a leader and say, do you know where your team is spending the most time? Or do you and like what parts of the process of really challenging for your team that sometimes can cause a bottleneck in the process?
00:20:50
Speaker
If you're they're able to have like those conversations, then that's a better start than just we're going to implement it because someone said we should. Like we're going to create some efficiencies so that we can handle when work volumes go up um more effectively than if we just say, well, everyone's in the office till nine o'clock tonight. You know, we can manage the the wave a little bit better.
00:21:17
Speaker
um So I do think it's from from a people perspective, it's it's definitely exposing vulnerabilities in leadership. um And I'm just I'm not sure if um if there's much room anymore for leaders who that's their whole personality is just I'm in charge and you're here to follow my orders.
Trust in Consulting for Small Firms
00:21:40
Speaker
because there's accountability at every level now and there's output expectations um and you need to to be comfortable and competent in discussing what your team is doing to enable you to get them to the next level and if you just kind of sat back because you're in charge for the last five years, um i think that that person is going to struggle.
00:22:08
Speaker
But that's just that's just my opinion again. So um now something interesting that also came up when we talked last was you bill in USD, but your currency is trust.
00:22:23
Speaker
Why is trust so important for you and your business?
00:22:31
Speaker
Yeah, so Trust is, I just gave a presentation about this to a class of students about the business of consulting.
00:22:43
Speaker
And now that I'm in a position of leadership, I mean, I still think of myself as a data scientist. I'm a technical person, but I'm coming to terms with the fact that in in ah in a in a human way, your job when you are in a position of leadership is to sell.
00:23:04
Speaker
because you have to feed your people and you have to ensure that there is like pipeline of that there's business to to make the machine and the business. check um So why is trust important in our business?
00:23:20
Speaker
I spent a couple years as a new entrepreneur getting a lot of very, very bad advice from very, very well-meaning people.
00:23:30
Speaker
And I'm trying to be positive and saying they're trying to help. but they're telling you what they know, right? And it wasn't until I spoke to, you know, I got all this advice. It's like, oh, do marketing and this type of social media and like have this type of lead magnet and all this nonsense. um And that wasn't- Did you ever do a dance to a Taylor Swift song for TikTok? Did you ever do that?
00:23:56
Speaker
As far as this interview is concerned, no.
00:24:05
Speaker
um fun hey
00:24:09
Speaker
I knew i knew I should have. I was, I had a call with my lawyer before this podcast. knew I should have. I couldn't get in touch with him. I knew I should have had that conversation. There's always a tricky question in there, right? There's always some tough questions about the TikTok and stuff like that.
00:24:31
Speaker
Yeah. Um, No, I mean, I think certain types of business and and the simplest example I say to people now is think about the difference between a B2C company or even a D2C company, a direct consumer, that makes a product that has a low price point.
00:24:50
Speaker
Right. I like to drink Coca-Cola. I don't need to know who made that Coca-Cola or who the CEO of Coca-Cola is to go into a store, be like, Hey, I'm thirsty. I want something sugary.
00:25:03
Speaker
a can of Coke. Right. Cause it's a low price point. It's a commodity. Right. But for us, like I, a lot of consulting firms say this, some live it, some don't, but I firmly believe, and sometimes it's hard to, to actually get clients to understand how fully I believe this, that,
00:25:23
Speaker
we are essentially working side by side with our clients and it's not just like hey we're a vendor that person's a customer it's like they're a partner we are helping them solve a problem together and maybe we'll stick around in the long term maybe we'll have a long-term relationship and that that's the ideal and maybe we won't but you are working side by side with that person and so they need to trust you right because regardless of everything else and things that some people say, like people want to work with people that they know and trust and like, or that someone that they know and trust and like knows and trust and likes. Right.
00:26:03
Speaker
And this is the same when you're looking for a full-time job. And then also just the nature of the the price point of the work, like time is money, you know, projects that you deliver in consulting take a lot of people and a lot of time.
00:26:17
Speaker
And so they're very expensive and until you have kind of a body of like social proof and a brand behind it which i also have thoughts about just because you've done a bunch of other things and you know communicated them in ah in a marketing way does that mean someone should trust you don't know but if someone is going to come to me and say miles i want your team to deliver this project and i'm going to say hey that's going to be half a million dollars they better be very confident that I mean what I say and I'm gonna deliver what I say I'm gonna deliver and that they can trust me.
00:26:52
Speaker
um And so like it it's it's not negotiable and it is the most important thing. and especially when companies are small, someone else told me as uncomfortable with it as I am, when companies are small, the founder, just like the lead singer of a band, right, is essentially the brand Right. And they're the person that the the clients in leadership have the most face time with. And then as the company scales, it starts to develop its own kind of abstract brand and culture.
00:27:27
Speaker
um But that trust in the in the leadership team and in management is still important.
00:27:38
Speaker
So just to go back to a little bit of practical use cases.
Agentic AI in Software Engineering
00:27:42
Speaker
um A lot of my clients are very excited about building, and a gent well, agentic AI engineers is the new job title that's going around. So what's what's the practical use case for agentic AI in in the in business?
00:28:05
Speaker
Where have you seen it work really well? Yeah, i think I think most people know that the strongest,
00:28:16
Speaker
the strongest use of agentic AI right now is actually in software. So people that work in AI and software engineering, if you ask some people, they say, oh, we've built the tools that are going to replace us. Right. But I mean, i don't know that's true, but yeah, it definitely something that that generative language models and that agentic ah is really good at is writing code and developing software and and building things.
00:28:52
Speaker
And before, this is an important distinction I want to make on this podcast. um People will use words without defining them.
00:29:04
Speaker
And so you will talk to someone and and they'll say, what do you think about agent AI? And you'll start having like a bar room type conversation and you'll get about half an hour in and people will have opinions. And if there are certain types of engineers and technical people, CTOs, the conversation will get heated.
00:29:21
Speaker
And then you realize that their definition of agent AI is not the same. So it's like, you're not talking about the same thing. That's why you're disagreeing. Right. So there's very, very complex definitions of agent AI. There's all kinds of,
00:29:34
Speaker
schematics on LinkedIn that are also generated by aa AI of that concept. To me, agentic is just a language model that can use tools.
00:29:46
Speaker
That is my definition. Right? If you have a chat model, it's just generating responses. If it can use a tool, search, a Python interpreter, you know, a plugin and MCP skill through claude canva whatever kayak you know that's a tool so it's agent yeah but the strongest use case is by far um is by far just doing software engineering and even my team i mean i am very transparent about the fact that you know my team we are data scientists first but we are able to build amazing product uh
00:30:24
Speaker
that we definitely couldn't build in the past, just in terms of front end development, in terms of platform development, because you have agentic coding and and even below agentic coding, you have AI assisted coding, right?
00:30:38
Speaker
I don't think I'm being negative enough on this podcast. um So. don't think that's true. I think the honesty is good because I just have like five questions pop up like as you were saying that. Yes, I agree. Defining what people are saying, I think, is so important because that is where misunderstandings happen. And then if you if you don't know where you're both coming from, hot like that can cause confusion. And yeah, i I did a little, you know, back in the day, I did a little research myself, like the difference and and like the simplest way to explain it was, yes, it's it's your, you become, it becomes agentic when it takes the idea and does something about it.
00:31:25
Speaker
Like it's doing. But I hear, it you will hear people, you will hear people, some people, they will casually use the word agent to just mean a chat bot. and And they're using these words interchangeably. And that's it leads to a lot of confusion. And that's part of what I think is a lot around the hype.
00:31:42
Speaker
There's an article. I think it they had examples historically. But they said if you have a term that should have a technical definition definition that you can look up in the dictionary, but it doesn't have a hard and fast definition that everyone agrees upon, that's a sign that you're in a period of inflated hype. So it's still ambiguous, and that's part of the problem.
00:32:05
Speaker
um But when I said I wanted to be more negative, I wanted to talk about what I think is a very bad use case is what I wanted to say.
Voice AI in Customer Service
00:32:12
Speaker
So all of you startups out there, fight me. Stop doing this. I know there's money to be made, but I hate you all.
00:32:19
Speaker
nobody wants Nobody wants to talk to a voice AI agent as a customer service cold caller. People already don't like that. They do not. They don't. yeah Right?
00:32:33
Speaker
Like, and there's, I have talked to probably 50 startup like founders and they have the same thing and they say, yeah, we're going to do a voice CSR. And like, they're talking about KPIs and they say, yeah, it's it's it's automated. So we got this this AI cold caller to call 75,000 people in this short time. And I'm like looking at the guy and be like, that's not a good thing.
00:33:00
Speaker
like like Like that is the another example of quantity over quality. And i had another conversation about this and and the founder said to me, he said, yeah, the models are so amazing. The people don't even know they're talking to and a voice ai They're not even aware.
00:33:20
Speaker
And i I like had to do a double take. I said like, your your system does not identify that it is an AI. Like it doesn't tell. So I said, you're deceiving your customer as part of your startup, right?
00:33:34
Speaker
Like these are the sorts of things that I think they're part of people realizing, oh, there's a lot of money to be made here and there's a lot of hype and a lot of spin. But there's very real like there's very real human factors at play that have nothing to do with tech, AI, data, machine learning that have everything to do with human psychology,
00:33:57
Speaker
ah human bias and and emotion that and user experience in a way too, that that these companies don't take into account before they start going down this path. They're putting the cart before the horse. They say, oh, we have voice AI. We can do this. We can automate. It's like, do people want to talk to an AI voice agent?
00:34:19
Speaker
Do they, are are you legally allowed to make ah a robocall that literally is a robot voice? Is that that something that is allowed legally? What are the ramifications of that? 75,000. I was just just repeating how many calls that that robot was able to do. And it's like, that's why I'm getting all those calls all the time, Erin. It's one of the...
00:34:45
Speaker
see I think I'm on their list. Seems familiar. Yeah. maybe Maybe I'm for someone that should work in this business. Maybe I'm too bearish about it.
00:34:56
Speaker
Like there are situations where I do want to chat with an AI where I know I am. And there are situations where I've been forced into that, like with a certain brand I interacted with. And then they they just, they have this dark pattern that just funnels you into their chat bot. And then my response is just like over and over again, i would like to talk to a human being. They won't let him.
00:35:19
Speaker
They won't let you. You're stuck. You're stuck in the cycle of doom. Yeah. I mean, no one's being fooled by the fake typing in the background either. Like, give me your name.
00:35:31
Speaker
Like, no, its no one's being fooled. I don't mind. I won't say the name. Actually, I don't know that it matters, but I won't. But I don't mind the type of chatbot that like doesn't want to deal with you, so just will like give you a refund out the gate. like that It's been programmed to be like, I don't care. like You don't want this item anymore. That's fine. Drop it off. You want a lazy Oh, a lazy chat bot that's not going to fight me on what I want. It's going to be like, yeah, don't, you don't need to box it. You don't need to label it, do nothing. Just drop it off. Money's going to be in your account in two hours. That's the chat bot I like.
00:36:12
Speaker
Yeah. When you're trying to deal with complex medical questions. No. Yeah, no, it it, that, that there's very, i think the, i think just like the distribution of, a value to be realized from different kinds of,
00:36:29
Speaker
specific use cases that may span many different verticals are unevenly distributed. i think that the value to be realized from generative AI and agents agenticating on all this are also very unevenly distributed across verticals.
Uneven Benefits of AI Across Industries
00:36:49
Speaker
Right. And i was talking about this, someone, someone today as well. we see a huge uptick in usage transforming the business of software development, right? Like Anthropic themselves, they released this chart.
00:37:04
Speaker
They say, what do people, they took a big survey. It's great stuff. They say, what are you actually using, know, Claude for? And it's like 60% of people say software development, right? But then we see other disciplines where I'm sure it could do an okay job, but because of the cultural aspects and the nature of the work,
00:37:24
Speaker
it is very difficult to realize value and safely use AI in those contexts. I'm talking about the world of law. law is the The raw material of law is language, and the difference between a comma or a single word in a legal document can make a lot of difference in a court of law.
00:37:46
Speaker
And then also it's an industry where people are very used, it's very wasteful to the environment, very, very used to working on paper, you know, the like stereotype of lawyers with like bankers boxes full of paper.
00:37:57
Speaker
And so we don't see that adoption um in these types of industries, heavily regulated industries where oh, it's just a little hallucination. It's like that that hallucination could be disastrous.
00:38:11
Speaker
um Millions of dollars has gone. Yeah. Yeah. From a legal standpoint, whereas in in more, you know, in software, you just kind of test and learn, you build something an MVP. It doesn't work. You throw it away. Right. So I think that's why we see more adoption of it in, you know, agentic coding and things like that and less of it in heavily regulated fields and and industry verticals. And that's not a surprising fact. I'm sure other people have made this observation.
00:38:42
Speaker
Just wonder how far we're going to go compromising like human experience, because, you know, if they're building, if you're automating, like building the product, and I know somebody's looking at it afterwards to make sure that it there is some quality control. But is if you've got the product that you're using being built by AI and then you have a problem and you call and you speak to an ai um like agent or chat bot basically, and then you can't get through to your real person because they've got their automated system and that's the process.
00:39:24
Speaker
Like, won't people just stop buying your stuff if the experience is not good for them? will that Is anybody worried about that when you speak to them as clients, that they want to make sure that people actually still value the the product?
00:39:47
Speaker
Yeah, I mean, there's there's definitely a very real concern ah amongst our clients and everyone I talk to, to be quite frank.
00:39:58
Speaker
ah about maintaining, this this is going to sound very grandiose, but you know what I mean, maintaining our humanity in the age of automation and AI, right?
00:40:08
Speaker
So not having the human the human factor go away. And some people that are doing software development are, I think, less concerned. I think the sad fact is that I see it as a kind of really awful arms race where some businesses that do these practices, they have so much data that they're trying to optimize to find out, okay, how much will people actually put up with and how much can we schlep off to them and they'll still buy, right?
00:40:40
Speaker
And there's actually a really good article I just read where it talks about how certain types of roles and certain things companies do from an economic standpoint, they're they're negatively incentivized and their purpose is to do the opposite. So there's a thing called like sludge and ostensibly when you call customer support at some businesses, their job is to help you.
00:41:02
Speaker
But in reality, from an economic standpoint, that is a form of kind of gatekeeping. And if they put you on hold long enough, you'll eventually give up you'll go and then you'll go away and they don't have to deal with your problem and they don't have to pay a person to be on the phone.
00:41:18
Speaker
Right. So there's a balance there. um But i think ah I think I am cautiously optimistic about people and um and about critical thinking and about empathy and a lot of these things.
Value of Authenticity in the AI Era
00:41:39
Speaker
And I think that if anything, because there's so much being generated by AI, um people will start to value that that isn't um more and things like authenticity and the human factor and the human touch are going to become more important. And I've seen silly exam, this is more of a social economic discussion we're having, but I've seen examples of this that are funny, like people are starting to talk about, you know, there there are really anti-AI people that are kind of radical. and I don't agree with that either. i think AI is just a a useful tool and and an amazing technology.
00:42:19
Speaker
But I've seen people talk about things where like you, they want to have just like a, with a organic food, they would like to start having like a seal that says like a hundred percent AI free. Right.
00:42:30
Speaker
So that you will, so that you will use their business or buy their product. Right. Um, And in all seriousness, some gaming companies, I follow this in the news as well, there are are game studios, independent game studios, and they have pretty pretty rigid standards ah about not using AI coding even, or AI content creation for artistic assets.
00:42:55
Speaker
um I believe the Linux kernel even, I think even though Ms. Torvalds is ah is a big user of AI, they have a policy about not submitting AI generated code just because there's the human oversight element, which is what they're more concerned.
00:43:14
Speaker
So I don't know, I'm cautiously optimistic that we will find a balance. And I hope that as the air slowly is let out of the balloon, not a bubble popping, that we will find a happy medium.
00:43:28
Speaker
and And i believe that I believe that in this kind of invisible hand of what the market, the market being made up of people actually want and companies, including a very well-known company that we work with all their technology, have probably pushed a little too hard on a lot of the AI things and you're starting to see people push back a little and i I think it's going to find we're going to find where the value actually is the market will let the corporations know hey this is what you can get away with and this is not useful and I don't care and this is and I'm i'm cautiously optimistic.
00:44:09
Speaker
I love that. I can't with the AI free sticker. It's giving, have you ever been in the grocery store and there's, you're in the meat department and you've seen like a gluten free sticker on the meat? It's like, well, why would there like why would there be gluten in it? Yeah. that sort I don't they put all salts in everything.
00:44:33
Speaker
This is true. This is true. But yeah, no. it's wild, um all of the the sort of ripple effects of of these conversations and decision points.
00:44:46
Speaker
But we've got a few rapid fire questions to to close out our conversation, Miles. And I'm going to i'm going to kick off with one about leadership.
00:44:56
Speaker
What is one practical leadership tip for responsible AI adoption?
Responsible Leadership and AI Dependency
00:45:03
Speaker
Yeah, so we we talked already about the the carrot and stick approach.
00:45:08
Speaker
So i think ah I think that that in itself is kind of an important point to note. And then in addition to that, i will say that i also just was reading a funny article about people in positions of leadership relying on AI too much.
00:45:29
Speaker
um So maybe I can put this as simply as I can, whether you're a business leader and or whether you are the most junior intern in the company, don't be lazy. Okay.
00:45:44
Speaker
Like there is this term that people use about over reliance on AI adoption for themselves. AI, sorry, not AI, option, yeah, called cognitive offloading, where you just say, I don't have time to worry about this.
00:46:00
Speaker
I don't have time to think about it. I'm just going ask ChatTBT, copy and paste it, right? So when I hear stuff like Mark Zuckerberg, who is the world's most wealthy Android, um saying things like,
00:46:17
Speaker
Like he is going to make an AI avatar of himself to replace himself in meetings. I mean, for him it'll be easier, right? Because it's like there's not lot of loading happening there, right?
00:46:30
Speaker
I mean, maybe he maybe he already has replaced himself. yeah with ah Maybe that is a clone, an AI clone. but But when you hear people like leaders saying, oh, I can automate away my job.
00:46:43
Speaker
and we can adopt AI in the org and it can just run like a machine that's 100% automated. like that's That's not good leadership and that's not good AI adoption.
00:46:54
Speaker
Your job is to care about your people and guide them and help them and solve the problems for them. It's not something that is just like done perfunctorally that you offload and automate, right?
00:47:10
Speaker
so You should be automating the menial stuff. like like People are even at the lower level, not in leadership, when they talk about AI adoption. like I met a guy at an event, he's in finance, and he comes up to me, he's like, oh, you do AI? And he says, I'm so excited, I want to like automate away my like a lot of my team.
00:47:30
Speaker
And I just said, I'm done talking to you. You're not my customer. right like get like Get out of here. right That's thinking about the wrong way. The point of AI is not to take anyone's job. The point of adopting AI is not to do a 6% layoff afterward.
00:47:46
Speaker
It is to free up people's time and make them more efficient so they can focus on more meaningful work. So you still need to do the work. can't be lazy. And especially for people in leadership, your your job is to be a leader and have people trust you and guide them.
00:48:02
Speaker
It's not something you can offload to a language model.
00:48:07
Speaker
No, just. shouldn't. It shouldn't. And there's so much opportunity now for people who have great ideas and who can use all the experience that we have to actually sit back and have time to think.
Long-term Societal Impacts of AI
00:48:22
Speaker
Like, that should be the goal. Let's have time to think about how things could be better. And we have tools to help us get there. I mean, ultimately, will jobs be impacted? Probably, because that's the cycle of life.
00:48:35
Speaker
But it doesn't mean to say that we're all going to be on some kind of, like, government check while the robots run everything and you know, we're just hanging out.
00:48:50
Speaker
Yeah. i think that's I am a, you know, and now your listeners do from this conversation that I'm very contrarian, but ah let me put it this way. I'm still waiting on the paperless office.
00:49:03
Speaker
Right? Like people, and I believe, I believe that AI is amazing and that it can automate things, but There is a human aspect, a human bias here.
00:49:15
Speaker
And there is a, I think it, I forget who it's written by. I i wish I had looked it up before, but there is a well-known essay and it's called something like Reflections on the the Futures of Our Children or something like that. And this is written in like 1895.
00:49:33
Speaker
And the author is talking about industrialization and automation, and he's saying, in the future, it's going to be a really real problem that people will have way too much free time, and there won't be any work to do, and they won't have meaning in their lives.
00:49:46
Speaker
right And and that was that was more than 100 and some years ago. right So there's some bias there about this idealism, about this utopian state where everything is automated, but the reality is just that work changes and we can do things more efficiently.
00:50:05
Speaker
and even though there's a lot that's bad in the world right now, over the decades, like you can look at Hans Rosling's visualizations, like over the decades, life has gotten better on average and nations have gotten wealthier and and the amount of suffering and the amount of hard work that we have to do has gotten to be less and less, right?
00:50:28
Speaker
But that doesn't mean that we're going to reach some state where Everything is perfectly automated. I don't I don't believe that. And i also want to know, are you going to actually send this video to to Mr. Zuckerberg?
00:50:39
Speaker
Are you going to tag him on Facebook? Yeah, directly. I have his post Malimo. We should post it on Facebook. You'll see it.
00:50:54
Speaker
Miles, what do you think? We talked about leaders specifically, but I think just about everybody is a user of AI today, even if they're just dabbling. What question should people ask themselves about their own usage if they haven't paused to think about it yet? Yeah, I think ah i think it's just a question you would ask yourself about about any use of technology, like even your phone. It's very similar.
00:51:22
Speaker
What are you comfortable with? right i know some people that have a dumb phone they say i'm not comfortable with you know having all this data collection and things like that for me the value isn't equal to my privacy right or the utility that it grants for other people there's it's interesting to me because like i'm technical and you know i would say 99.9 what i use language models for is for either language models for is for either
00:51:53
Speaker
coding, which I do a lot less of now, and my team uses it for that, or it's around business problems or search or information synthesis. Right. But for some people, like a question they ask themselves is like, hey, what's important to you and and how is this tool going to make your life better? And for some people, it's really interesting to me because it's all personal.
00:52:13
Speaker
It's all, oh, this helped me organize my recipes. This helps me, you know, do this part of my hobby. This helps me keep my family life organized. Right. um But yeah, i think you have to ask yourself, and also to to do that, you have to experiment a bit.
00:52:30
Speaker
You have to try different things and say, hey, is this actually useful for me? Is it easier? know am I getting value out of it? Or is it it not? And in some cases, maybe you're better off just you know using a pen and paper.
00:52:46
Speaker
Because on the other side, yeah i should probably lead with the negative stuff first and then have the positive stuff after. But on the other side, there are these examples of people that I am concerned about who you can read about like, and not even, i don't want to talk about the sad, tragic cases, which are related to mental health. Like that is a very real issue that is not, it's being given attention, but it is still not being given enough attention.
00:53:11
Speaker
But there are even like humorous cases of people who are not so good at at at self-control, i think kind of becoming too strong of a user of ai and you see these startup founders and they sort of start to gradually get pulled into the deep end and you can watch their Twitter feed like slowly devolve into to memes and and really weird like tinfoil hat stuff.
00:53:42
Speaker
um So, so yeah, I think the other question you should ask yourself before you start, you know, using these types of tools is, know, do I have an addictive personality?
00:53:57
Speaker
and then also yeah And then also, for a lot of folks I know, am I okay to use it and not start personifying it or start projecting, right? Like you need to use it objectively. It's it's dangerous, I think, unless you have a lot of clear headedness to start going and using you know language models for personal advice or in the place that you know some people- Because it's very fascinating.
00:54:27
Speaker
very validating for your perspective. You can easily... yes yeah you're like... You could easily go down some some sad roads if that's what you what you're using it for. but yeah So yeah, be be it would be good to see some kind of... I'm not i'm not massive on like regulations or frameworks or whatever, but I do think people need something.
Personal Comfort with AI Adoption
00:54:55
Speaker
Something thing to help. with Yeah, it's still it's still a bit of a wild west, right, still. And Amy, if I'm going to go down a rabbit hole, i'm going to do it myself on Wikipedia as God intended. I'm
00:55:11
Speaker
and just going to keep clicking the different blue links. And I started reading an article about the North American economy. And when I'm done, I'm reading an article about Mongolian flour. yeah
00:55:28
Speaker
Yes, I like it. I like it.
Connect with Miles Harrison and PractiKai
00:55:30
Speaker
yeah Miles, where can people find you if they want to know more about you and PractiKai? I'm very active on LinkedIn. You can find me. I'm Miles, M-Y-L-E-S Harrison, H-A-R-R-I-S-O-N.
00:55:44
Speaker
PractiKai is difficult to spell for some people, so you can go to P-R-K-T-K dot A-I or PractiKai dot com, and there's two K's in PractiKai.
00:55:56
Speaker
often. Fabulous. It's been so good to talk to you again, Miles. Thank you so much for spending your Tuesday evening with us. We'll let you get back to conferencing. Likewise. It's been fun. Thanks so much for having me.