← Artificial Intelligence Podcast: ChatGPT, Claude, Midjourney and all other AI Tools
Artificial Intelligence Podcast: ChatGPT, Claude, Midjourney and all other AI Tools
Is AI Going to Take Over For Outsourcing With Thomas Doherty
00:28:55
Loading summary
Transcript32 lines
- [00:03]
A
Is AI going to take over for outsourcing? Let's find out. Today's amazing special guest, Tom Daugherty welcome.
- [00:09]
B
To the Artificial Intelligence Podcast where we make AI simple, practical and accessible for small business owners and leaders. Forget the complicated tech talk or expensive consultants. This is where you'll learn how to implement AI strategies that are easy to understand and can make a big impact for your business. The Artificial Intelligence Podcast is brought to you by Fraction aio, the trusted partner.
- [00:29]
A
For AI Digital Transformation.
- [00:31]
B
At Fraction aio, we help small and medium sized businesses boost revenue by eliminating time wasting non revenue generating tasks that frustrate your team. With our custom AI bots, tools and automations, we make it easy to shift your team's focus to the tasks that matter most, driving growth and results. We guide you through a smooth, seamless transition to AI, ensuring you avoid costly mistakes and invest in the tools that truly deliver value. Don't get left behind. Let Fraction AIO help you stay ahead in today's AI driven world. Learn more and get started.
- [01:02]
A
Fraction AIO so Tom, I'm really excited to have you here because this is a topic that I've been very interested in. I've noticed that we've already shifted from if someone reads your social media post and it's a little weird or there's a misspelling, they used to say, oh, did a VA write that? And now it shifted to did AI write that? So we've already shifted in our mindset this idea that outsourcing is replacing it, being replaced by AI. But my original perspective was that AI is the great equalizer because if you take, let's say a person whose second language is English, that's their biggest weakness in competing in the American market. If they can get rid of all those artifacts that reveal they're not a native English speaker, now they can compete at the same level as a native born va, I see that as an advantage. So I'm very curious how this is starting to change the outsourcing market. We've already seen AIs that will remove your accent and do other things to make you sound more American or what are some of the things you're noticing and which direction are things going?
- [02:03]
C
Yeah, I think you're going to continue to see that big shift in that direction. I recently somebody reached out to me too about outsourcing growing in Africa and African outsourcing agents doing this type of work because it's a lower cost structure than in the Philippines or India. And they're talking about how they're Going to use AI, agentic AI to continue that same model where, yeah, we're going to help get rid of your accent. We're going to help get through that processing. And if you remember, the basis of outsourcing really is to get to business process outsourcing. Right. So how do you get to the main source of the process and then let the AI roll through that? Right. That's all the AI is doing. It's just going through process step by step. Where it gets curveballs is the nuance.
- [03:00]
A
Yeah. So that's what I want to talk about. What are those particular nuances that are making that difference? We've had these generations of outsourcing. Like first we outsourced to Central and South America because it was same time zone. Then we shifted to India because it was even cheaper and India has very strong English. And then we started shifting towards Philippines or Eastern Europe or South Africa. And now you're saying we're shifting towards Africa. And what do you think is the direction? Like what's going to happen? We've seen what happens when like factories get moved from China to India. There's like an increase in a decrease. Like what do you think is going to happen to the market? And for a company that's thinking of expanding operations, I've definitely seen a lot of you should hire coders from Turkey or Romania. We've seen a lot of that. Like Eastern European coders. And I'm just wondering what the future is at the larger company or enterprise level. Like I can understand VA is jumping country to country. I kind of shift where I hire from. For a while I only hire from South Africa. Now I mostly hire from Serbia. It kind of changes which is the best country for me. But I'm a smaller operation.
- [04:06]
C
Yeah. And I think that's where you're seeing the adoption of AI tools. The ones that are being successful are the ones that are actually paying for the processes to be broken down completely and, and starting the enterprise with small wins. Where you're seeing it fail is people doing in house adoption across the board and they didn't do the work. Right. This goes back to the original parts of outsourcing is you hire an expert, you have somebody go through all the deep dives, all the process mapping and then being able to flush out where those nuances come up where you'll have some challenges or errors. Right. We look back at when we did software transitions. Right. Somebody didn't do all the mapping. And then you find out when you go live, oh, people aren't Going to get paid on Friday. Nobody did that process mapping to go all the way through the system. It just got stuck somewhere. So I think where people can really be successful is when they start off the process the right way and they focus on the low hanging fruit first. Get some successes, work on some of the processes that you can easily manipulate and get them to a positive outcome and then build off of that once you get past those so you can get to the harder stuff.
- [05:29]
A
One of the things that I've seen a lot in my AI projects is that when someone comes to me and they don't have a process clearly defined and then they want to AI it, it's very difficult. Especially when they go, I'll know when I see it if it's working, I'll be able to tell you. I'm like, that means it's going to take years because I have to keep guessing. And this is the most common thing I've seen people do with a new hire, especially a newer business. They'll give a person a goal, but it's vague. Take care of my social media. But that's such a broad definition. And it's how many posts do you expect a week? What's the measure of success? What's the measurement of failure? How many posts should I do on whichever platforms? What's your process? And so I've seen a lot of like whenever I work with a company, sometimes they'll have a really detailed process. They go, we know this is complicated. I go, no, you haven't. You have it recorded. That's easy. If you know what it is and you can show me an example of good and bad input and output, then I can build that. It has the pieces. But we sometimes assume that because we're hiring, it goes back and forth. We're hiring an AI, so it's smart, it'll figure out. Or we're hiring a human, they're smart to figure out. But it's really hard to work a job if you don't know the measurement of success. That means you don't even know if you're doing a bad job until someone tells you. And it's a surprise for both of you. And I see this very common with people's first VA hire. It's like almost, that's why everyone almost universally fires their first VA and you didn't record your thing. You didn't have a measurement of what's the process to replicate. Here's my measure of success, here's the things I'm looking at. It certainly Happened to me with always in social media, where they, like, I had one social media manager who thought like, the measurement was reach. I was like, no, the measurement is money. The coming sales are we making? That's how I pay you. I can't pay you in reach. It's not a real thing. But you get caught up in these pseudometrics.
- [07:08]
C
The point. Right, the pseudo metrics. So it's very interesting. Whenever we win a deal, one of the things I, and being an operations person, I always want to ask the client, why did you choose us? It gives you insight over what really drove the decision. Was it cost savings? Was it efficiency? What. What was it that was going through their heads? And then the second part of that question is, what metrics and KPIs and dashboards today do you have to determine if you're successful or not? And that's either met with one of two things, a very like blank slate of face because they don't have it, or you have somebody that has like KPIs and metrics through the roof. I've had clients that have had over 100 metrics. To me, that's just too much. Right. So we try to dial it back. In my experience, there's usually eight to 12 metrics that really drive the success of that particular group. And it's getting the client to see what those 8 to 12 are that are really driving that success. But again, how can you scorecard an operation that doesn't have any goals to hit?
- [08:22]
A
Yeah, I think that I've seen exactly those two directions where you have so many metrics, it's possible to keep up with them and it means that, like, you just get lost in the sauce and. But I've also seen where we don't have enough. And I think that the big problem is that we try to outsource when we don't have a clear system. Another mistake I see people making specifically with AI is they can, hey, we're going to train this AI and then it's going to replace you. And I've seen this happen at large companies where you're training your replacement and it's like, I'm going to sabotage it. I'm going to make sure it don't work right. And it's really the wrong language. Now, I think there's a lot of value in outsourcing, but I see outsourcing used by large companies in the worst ways. It's very efficient for smaller companies who don't have the agility. But if you're like Fortune 500 company. What are you doing? Like why are you cutting these corners? And one of the things that I do know, and this one, I worked at a Fortune 50 company. They said the reason we outsource to India is that when people call tech support, 10% of people will hang up when they hear an Indian accent and then we don't have to fix their problem. What cool perspective, right? That's so sometimes like it is malice. So I'm wondering like what is the right way to apply these tools? There's certainly an advantage. I think it's great to get people to other countries jobs. I live in another country, so I'm in favor of all of that. But I'm wondering now that we're seeing this competition between AI and outsourcers and it's is it really competition or is it these? I feel like outsourcers people in other countries have the most incentive to become AI masters the fastest.
- [09:53]
C
Right? Yeah. I think that's what you're going to see is the outsources will become the agentic AI. They'll build the tools and build the models for these folks to be able to plug in and use for their organizations. And you know, we've seen this before, right? We've had these labor spikes before. We've had software changes. It's. This isn't new. Right. Who's going to get left behind is the people that put their heads in the sand and pretend like AI doesn't exist. Companies are focused on delivering shareholder value. What's your. One of your biggest costs? It's labor. So that's why I'm always a big believer too in onboarding people the right way. That's again another process. Where do most people leave the organization? Usually two to four weeks in. If it's a bad onboarding process and you have a bad experience, you leave. And it's the same thing with AI, right. If you try to onboard a new AI or agentic AI and you have all kinds of issues with it within the first couple of weeks, what are you going to do? You're going to bail? One time, Jonathan, I showed up at a product meeting and that's not normal operations. Guys usually don't go to product meetings. Right. And I was new at this company and I was listening to this guy talk about their flagship product was going through a major overhaul and change. He was talking about flipping the switch the coming weekend and he said, does anybody have any questions? And I looked around the room and nobody was said anything. So I said, hey, I'M the new guy. I know I probably don't know, but who did you test this workflow through on the client side? Crickets. And then he went to explain Agile to me and I said, I know what agile is, but you do not test a software product upgrade without any kind of client involvement or workflow, especially in the legal vertical. You want to talk about screwing up a workflow? He ignored me. And they flipped the switch that weekend. You can guess who's call center team was getting blown up 9,000 ways to Sunday. And we had top 10 clients threatening to cut the cord because of that switch. So thankfully we flipped back the switch. We went back the other way, and then we went through client workflow to make sure that it didn't impact it.
- [12:15]
A
In the way that, you know, would.
- [12:18]
C
Be detrimental to the client.
- [12:21]
A
As soon as you said weekend, I was like, I don't know why that's everyone's favorite time to flip the switch is when people are the maddest. There's nothing worse. They make a change and then tech support's not there on the weekend, or they only work office hours. I hate all of that. Like, that's one of the things I do think the best of it. To me, the best thing about outsourcing is you could have 247 tech support, like the 247 support line. That's the first, like super useful thing. But yeah, I have seen that. And there's. I'm guilty of it too. When I'm building stuff sometimes, like, we destroy the good in pursuit of the perfect. And we. I do try to explain this a lot. When I was just talking to this client today, he was working on his new book. And I said, listen, books are democracy. The reader gets to vote. And you always have to remember that, like, you can put everything that you think should be a success. And sometimes it just doesn't work. There's certain. We've all seen movies that are amazing that everyone else hated or flopped. And you just, you don't know why. It just happens. It's not. There's just something about it that people don't like that much. And I've seen my favorite book that I wrote to me six months. It's the hardest ever worked on a book. My audience hates it. Of the 50 books I publish on my own name, it's the least popular. And then one I wrote in eight hours in a single day, one of my top five. So you just cannot predict. And I try to explain that just. And that's a really important thing. I think that we assume that AI is like this magic solution. Like I was talking to someone the other day and they were like, the best way to learn AI is tribal learning. And I was like, I don't think that's a word. And he's like, yeah, you just like talk to other people and figure it out by talking to your friends. I was like, oh, so you don't have to pay for top down training. And it's figured out in your own time. And it's. That's an insane process. Like some of the things I see with AI, one of the weirdest things about AI is that you don't version your updates. So there's a version called chatgpt4o that has probably gone through 500 updates that are not published. Like normally you see 04, 0.1, 0.2, 0.3 and they just don't do that, which means things break all the time. Things reliant on that technology. I don't know where this. Which is exactly what you're talking about, right, where they push updates and don't tell anyone. I've had a prompt work and then not work one second later. When I was working on a project and I was generating pictures for different AI agents I was building in chat GPT 2 years ago and I said, make a picture of whatever you think you look like. So I would have a different picture for each agent, helps me separate them and then it builds 10. And then he goes, I'm not allowed to do pictures of real people. And I said, I'm sorry, what did you just say to me? Did you just go sentient? That's. I was like, wait, it just happened. And it was like, but I tell from the prompt. Oh, they just changed the rules. Because a prompt that worked went from working, not working in the middle of a work afternoon for me. And that's one of the problems with AI. And the other thing is that it's such a bad customer experience. They're like, here's a blank page, figure it out. That's the onboarding or the training with AI tools is the worst I've ever seen of any software tool. The first time I used ChatGPT, it's a blank page. And I was like, where's the instructions? And they're like, you don't need them. I was like, I definitely do, please. So we're starting to see that bleed into other parts of the AI culture, which is that we try to automate processes, we have it refined, we try to train people without training Them. And we pick tools without really a logic. I've worked on projects where people buy a tool and I go, why did you buy this tool? And they're like, not really sure. And I'm like, what did it cost? And then the number's insanely high.
- [15:51]
C
Yeah.
- [15:51]
A
And we've seen some studies come out recently that say 95% of companies are not seeing an ROI on their AI investment. And my belief is that those are the companies that did not do a pilot program or that kind of skipped over the training. The thing I do boutique solutions is I work with smaller companies, like 50, 200 employees. And I. My goal is always to change them as little as possible. I go, what tools using? What's your workflow? Because getting people to change is hard. Right. Try to get someone to change the toothpaste brand. But now you're trying to say, we're going to change the tool using the workflow, all these parts of the process. And it's very hard to generate consistent behavioral change. And I think that's something that we're missing right now when we're implementing AI, especially across large systems or large teams, in that there's always going to be a dip whenever you implement a new tool. And I always say, first figure out how long will it. Like, how long will it take to adopt and retrain your team and get them using this new system and get over that hurdle with, I'm not using it. The old way works. And I'm. I do that all the time. I'm so in the old way. And then actually get to the profitability. Is it worth it? And if it's gonna take eight months and it's a 1% boost in revenue, maybe it's not worth it. And we are AI consultants. Hate when I say that. But I'm like, we have to be strategic with what moves we make. Because you also have to factor in the emotion of the team or the positivity. I forget the right word I'm trying to think of, which is like, morale. That's it. Because if you input these projects, especially with language, like, we're going to outsource you, where everyone's getting replaced with AI and I've killed morale. And you're implementing these projects like, of course the projects are failing. Everyone thinks it's their enemy. Like, I go back to that episode of the Office where they tell them to put in, oh, we have to make it look like all your sales went through the website. And they. What do they do? They sabotage the website to save Their own jobs. Right. No one learned from that episode 20 years ago. It's like, I would. Everyone would do the same thing. If this project succeeds, you're fired. Well, it's a good thing. The project is not going to succeed. I already know how it's going to end.
- [17:48]
C
And you know this too, Jonathan, that when you go through the executives, you go through the purchasing group, they think they know what the process is or what they're doing, and they haven't really talked to the people that do the work. So that part of the. If you miss that part of the analysis and you're trying to do that work now, you're already starting behind the eight ball. The people that really know the job are the people that do the work. And if you don't gain their trust or help the morale from the beginning, it's going to fail miserably. And what you typically find is when you start to uncover these things and start pulling them up, the executive team and the other. I didn't know we did that. I didn't know that was happening. When did that change? All those conversations haven't been had. And then you're trying to slap this AI process on top of it. That just convolutes the whole thing.
- [18:40]
A
So I always. Full of metaphors day. I always think about Dr. House, which is in America, Comm House, England, they call them Dr. House, which is that everyone lies and you have to figure out what the truth is before you build the process. And I always have to. People will always tell me what they want, but it's never the truth. Like, great example of this is, we want more leads. I'm like, okay, let's say if your phone reads a thousand times a day, no one buys anything. Am I getting a bonus? No. Oh, so what do you actually want? More customers. And it's what they really mean is qualified leads. And qualified leads just means leads who turn into customers at a certain rate. And it's very important to understand what the real goal is when you're building a process. And usually the goal of a process is different for the manager, for the employees, and for the boss. Like, everyone sees it differently. And it's amazing to me how often management doesn't know what tools their employees are actually using. And you'll have these. And sometimes it's like they're using the. Like, you've done two software updates, but no one's using that software. You're paying a ton of money. I was working on a project this year. They were paying a hundred small Companies paying a hundred thousand dollars a year for software. And I was like, no one uses it. Why do we pay for this? And the CEOs what we did a major update and switched it over to it. I go, you're, it's not what's happening. No one uses it. I check the logs. I said, I said, here's what you can do is ask everyone to log into it right now and you'll see no one has a password. And it's one of those things. And I always ask this question, is it what problems does this software solve? And the answer that they gave and it's this is all you can tell people buying different ways. When the answer is not what the software does. That's when I always get the most worried. I've run into that a lot. I'm like, this isn't what that software does. It's not even close. Close. Okay, that's amazing if it did that. But it doesn't. And this happens a lot because you get those really cool sales calls or you get excited and you have a lot of emotional buying. But I always try to pull back. And the same thing with hiring an employer, the same thing with outsourcing. It's like, what problem does this solve? And sometimes the only problem it solves is cost, which is a dangerous place to be in because usually there's diminishing returns. Like I've been very fascinated recently with the decreasing quality of cars. It's like there's all, there are so many recalls. Recalls used to be like rare, they happened, but they were rare. Now it's every car in America has been recalled in the last year and we're seeing this same thing. If there's a dip in quality and it's like you're just constantly dipping. Quality has a long term cost. So I try to look at it from that perspective. Same thing with startups. When I go, what problem do you solve? If you can't identify it, you're not going to be in business very long. I'm very interested in what is the right decision making calculus. When someone is deciding, do I want to replace my customer support team with people in another country? Do I want to replace my customer support team with an AI or do I want to. I think the best solution is like really creating AI enhanced employees. Like I really think that you can have a great team in another country that has AI tools to really enhance them. I'm not so much into the accent hiding. That's to me that falls into deception. Like trick. It's like when someone will give you a fake name so you won't realize they're in a different country. I don't like any of that. I think that's when someone figures it out, it creates a really bad starting point because now you've, you have deception is like the first thing you've done. That's why I don't like AI phone calls for sales because it's trickery and it creates a bad like starting point. Once you've lied to someone, you can never recover that trust. But I think there's a lot to be said for having AI tools that can help you create support faster. Like you fill out that phone tree and then you talk to the person, they go, hey, who is this? I'm like, why did I push all those buttons? Why'd I type in my account number? So if you can have an AI and I know there's some companies doing this that keeps that and you, they know which account it is and they can see every person you've talked to that's really useful. That stuff starts to get useful. But I feel like we're just getting distracted by hype, whether it's outsourcing your team or whether it's using AI and we're not being methodical like we should be.
- [22:45]
C
Yeah, I think the most successful outsourcers because such a great opportunity, you take your already, you have clients in your pipeline and folks that you, you're already working with and you're already having success with. What a great opportunity to go to them and say, I want to partner with you because you're one of our most valuable clients. How do we partner together to build a solution that works best for you and your organization? That way there is that morale, that buy in that team part. We're going to work together to build this best thing and then you get a training ground over best practices and what you can do to help other customers as you go through the process. But I agree, I think a hybrid solution, whether people are wearing AirPods that translate language or people are wearing smart glasses so that they can hear the language, I think all those things are possibilities. And you take away that human touch and I think that's where you start to get in trouble. But I do think you could automate low level workflow type items like a password reset or you don't need warm touch and fuzzy feelings for those types of things. So it's just thinking along those lines more as a partnership than trying to just. I hate the, I call it the package Solution, right? Oh, hey, pin price A and we'll give you this package. How do you know that package is the right package for the client? That's where you run into all those problems, right? We've seen this for years too. Is people buy an out of the box solution and they wonder why it doesn't work. Because you didn't do any of the work, you didn't do any of the analysis, you didn't do any of the real stuff that makes this tick.
- [24:26]
A
So funny you bring that up. I think about that a lot. There are so many tools I have where I use one of the features and, and they sometimes the companies don't realize that like they keep adding features no one wants and no one uses. Just do one thing really well. Like my favorite image generation AI is now like doing a bunch of movie stuff. I'm like, nobody wants that. Nobody actually wants to use AI video other than for low quality social media posts that people only watch. If you say this is an AI video, there's so much use for AI imagery and so many really good places to use it. But it's not as exciting. I don't know what you're doing but you're really good at one thing. And there's so many tools I talk to because I talk to the developers a lot or I'll talk to the CEO. I'm like, yeah, nobody. I was like, you have this entire feature set that no one cares about. Just work on this. If you audit your users. Everyone's using this one feature, this one feature. And I think that you're exactly right that we grab these tools. Or there's also a problem where sometimes the onboarding is so bad that you don't even know it has the features. And it can be. I've seen it where onboarding is just too long so no one watches it. Like it's a two hour tutorial where you learn everything. I'm like, that's too long. Or it's three clicks and it's, it is hard to balance that because I'm guilty of. I've not watched short and long tutorials. But it, there is something that has to be figured out there which is like what can this do? Like one of my favorite tools, they release the new features every Thursday and they drop like 20 features a week.
- [26:00]
C
I'm like, what?
- [26:01]
A
Can't keep up with that? It's great. It's a great tool, it's super powerful. But there's a lot of tools like that. So I do think that we and this is a great lesson, which is like to really figure out how you want to use the tool, implement with your team so they know which feature to focus on and say, we're going to use this. Here's the problem it's solving, here's how we're using it. Because there's nothing worse than when the CEO buys a tool and says, here, this is solving your problem. You're welcome. And you go, what? I don't have that problem. Now I got to justify this purchase you made.
- [26:28]
C
There's a daily newsletter I get that I really like.
- [26:31]
A
It's.
- [26:31]
C
There's a AI one and then there's a robot one. And I love the format that they use. They tell you what the product is, what the breakthroughs were, and they give you like a little summary and then at the end they really summarize it up, why this is important, what does this actually solve, what does this fix? Why are we breaking down this particular item? I love that format because I think people learn better that way. That you give them the highlights of what it can do, but then the summary of why it's important and then let the reader or the person, the user be able to figure out why that's important to them. Either, yeah, I need this or no, I don't. I'm just going to go to the next tool.
- [27:11]
A
I really like that. I think this is really good food for thought for a lot of people because it's so funny how we used to always talk about outsourcing and now it's. No one talks about anymore because everyone's so busy talking about the AI hype. I think it's going to come back for people that are really interested in what you do. Tom and thinking, oh, this is something. I do want to build an AI enhanced team. I do want to understand this a little more. Where's the best place to find you online and the things you're right about, the things you're working on.
- [27:32]
C
So you can find me on LinkedIn, I'm on Thomas Daugherty Jr. You can find. That's the main place where I post and do most of my stuff. Most of my articles and contact information is there. And I've done a lot of stuff with a lot of different verticals. And it's amazing to me the big toolbox that you get from working with all these different organizations, whether it's legal, higher ed, technology, finance, it's just amazing to be able to apply the tools. They all have the same issues and problems. You just show up in a different space.
- [28:08]
A
Amazing. Thank you so much for being here and everyone. I'm going to put the links to Tom's stuff below the video and in the show notes. Thank you being here for another amazing episode of the Artificial Intelligence Podcast thank.
- [28:21]
B
You for listening to this week's episode of the Artificial Intelligence Podcast. Make sure to subscribe so you never miss another episode. We'll be back next Monday with more tips and strategies on how to leverage AI to grow your business and achieve better results. In the meantime, if you're curious about how AI can boost your business's revenue, head over to artificial intelligencepod.com Calculator Calculator Use our AI Revenue Calculator to discover the potential impact AI can have on your bottom line. It's quick, easy, and might just change the way you think about your business.
- [28:51]
A
While you're there, catch up on past.
- [28:53]
B
Episodes, leave a review, and check out our socials.