A Job Done Well - Making Work Better

Will Artificial Intelligence Take All of Our Jobs?

• Jimmy Barber and James Lawther • Season 2 • Episode 41

Share episode

In this episode, hosts James Lawther and Jimmy Barber discuss the impact of AI on contact centres and back-office jobs. 

This follows a question from Rob, a member of the audience, who asked about the potential for AI to reduce contact centre jobs by a claimed 80-90% in the next year or two. The hosts share insights from a former guest, Alex Alexander's AI team, and James interviews Google's Gemini to discover if the AI themselves believe they will put us all out of a job.

This is all balanced with a bit of real-world experience of tech transformations. 

Check out the episode and see what they conclude - Are the robots taking over the world? You might be surprised by the answer.

XOOTS – Your digital transformation co-pilot

Callcenters in the Age of AI

Got a question - get in touch. Click here.

Speaker 2:

Hello, I'm James. Hi, I'm Jimmy and welcome to a Job Done Well, the podcast that helps you improve your performance enjoyment at work.

James (3):

Hello. How you doing?

Jimmy:

Hi, James. I'm doing well. How are you?

James (3):

I'm doing fabulously. Thank you very much.

Jimmy:

So James, what are we gonna talk about today?

James (3):

oh, well, today we have got a special request. I've got a, note on LinkedIn from a friend of mine and I will, Read it to you. So Rob said, we'd love you to visit AI this year. Apparently had a meeting with Google last week and if half of what they demoed is possible contact center and back office jobs are about to reduce by 80 to 90%. So we are gonna dig into automation of, Back office jobs and contact center jobs and see whether AI really is gonna take over the world. So that is our conversation for the day.

Jimmy:

The robots are taking over the world,

James (3):

Yeah, I listened to something, I was listening to a very interesting podcast the other day. Did you know that you have got a hundred times the processing power of, um, chat GPT in terms of the number of neurons you've got, but chat GPT has got access to the world's information and you haven't got that crammed away in your.

Jimmy:

No. My, my skulls. My skulls full of song lyrics from the 1990s and stuff like that.

James (3):

well, it's not that chat GPT is, cleverer than you. It's just wiser.

Jimmy:

I think I'm better looking though. So what have you been up to, James?

James (3):

Um, I have had a very exciting weekend. A friend of mine went away for the weekend and he needed somebody to dog sit for the weekend. So all I have was dog sitting for an Irish territ this weekend. So we took this dog for a walk in our local, country park, and it was a damning, uh, indictment on the human race.

Jimmy:

Wow,

James (3):

What? Well, I'm out walking around with this dog and obviously there's lots of other dog walkers about, so everybody's picking up their poo and putting it in bag plastic bags like good citizens. And somebody had picked up their poo, put it in a plastic bag and left it on a park bench, which I swear was right next to a bin. And you've go ask yourself, what is wrong with these people to be better off leaving the poo on the floor. There you go. What have you been up to?

Jimmy:

Well, This week, my partner and I, we had matching tattoos. Done James. You can probably just about see that.

James (3):

Okay, what's it say?

Jimmy:

It says always in all ways.

James (3):

Ah, very clever, deep.

Jimmy:

It's a, it's a song, it's a quote. We actually picked it up from, a person called Ryan Serhant, who was on million dollar listing selling high-end properties in New York. So I have got a quote from somebody on trash TV on my own. There you go.

James (3):

Nice. I can just imagine what Mrs. Law would say to me if I'd suggested we went for his and hers tattoos.

Jimmy:

Oh, God, she, she struggles with his and hers names, doesn't

James (3):

Oh, she, yeah, she does. I'd end up with, um, yeah, tea white, no sugar tattoos on my arm, lest I forget. Oh, it's maybe say now as well. I should think.

Jimmy:

Anyhow, so AI James.

James (3):

Ai. So we started off and we started off by, having a conversation with, a guy who's been on the podcast before.

Jimmy:

Alex Alexander.

James (3):

Yeah, Alex Alexander. He asked his ai, he was an AI man. He asked his AIS what they thought about, um. Contact center and ai. So we will thought we would, um, play that clip and get you have a listen to it and see what you think. So here you go. Zoots version of will AI take over the contact Center world,

Xoots:

Today we're looking into something pretty big, the transformation happening in call centers, driven by ai. We've got some great insights here. Partly thanks to our collaboration with Jimmy Barber and James Lawther from a job done well. Yeah, we are really moving away from, you know, the traditional model of just people handling repetitive calls. Exactly. It's not just basic automation anymore, right? And they're becoming incredibly human-Like they can pull information from different places, reason through it, make sure they're following the rule. Compliance is key. Absolutely. And they remember your past interactions, so you're not starting from scratch. And the voice part is just incredible now, isn't it? With tech like whisper 11 webs, AI voices sound so natural. Totally. They can handle voice calls, understand nuance, even express a a bit of. Well appropriate emotion and multimodal systems are adding another layer. So for anyone listening, this means some really significant changes are coming. Think about support costs dropping dramatically. We're talking potentially 80, maybe even 90% lower costs in some areas. Wow. And getting help instantly, right? 24 7. Any language, that's huge. It also takes pressure off the human agents west. Burnout definitely reduces turnover. And look, this is in sign section by 2026, the prediction is that more than half of those basic Q1 requests, AI will likely animal most of those across chat, voice, email. But that's a massive volume. But it's important to say, this doesn't mean humans are disappearing from call centers. No, not at all. The roles are evolving. Humans become, uh. Supervisors handle the complex escalations, the situations, eating real empathy. The tricky stuff. The tricky stuff, yeah. And we're also seeing specialized AI emerge agents trained specifically for, say, finance or healthcare or telecoms with all their specific rules. I. Okay, so here's something really interesting. Looking ahead 2026, again, this idea of AI agent orchestration, that's a game changer. It's basically having multiple AI agents collaborating behind the scenes on a single customer issue. So like one AI pulls up a day, another checks compliance, maybe a third analyze of the customer's sentiment. All at once, precisely working together seamlessly. That's what turns a call center into more of an AI service hub, right? Powered by cloud data, AI supported by these leaner or more specialized human teams. It's really about augmentation, making the whole system smarter, not just automating cast away, focus on implementing it ethically and ultimately for the customer. What does it all mean? It needs faster support, smarter support. More helpful interactions. It really makes you think, doesn't it? How this fundamental shift in customer service powered by AI might ripple out into other service industries and you know, what does this evolution mean for the whole relationship between us and technology when we need help? Definitely something to consider.

James (3):

What do you make of that then?

Jimmy:

Well, I thought it's quite interesting, James. I mean, it shows what AI can do that, their answers are pretty good. Very, very polished though, very slick. But, I always find there's something slightly, um, false about, it's almost, they're almost too perfect

James (3):

Yeah.

Jimmy:

The way they look and the way they speak is almost

James (3):

They're much better looking than us. Um, we'll put it out as a YouTube video so you can see it.

Jimmy:

Think I, I think also, and we've come onto this a bit more, I think there is a, a nuance that gets missed. A little bit in this, argument, and that was they were saying that AI can do a lot of the basic work, which will reduce burnout. I actually think the opposite is true because what they were saying is they'll leave all the complex, difficult stuff to the humans that makes for hard work. And as we talked about recently in unintended consequences, that isn't necessarily what people wanna spend every working hour of their their day doing.

James (3):

So I thought I'd actually have a conversation with an AI because I'm sad and lonely like that, you know, chat, GPT is my new best friend

Jimmy:

this

James (3):

and, um,

Jimmy:

geek talking to

James (3):

geek. Talking to geek.

Jimmy:

artificial geek.

James (3):

Yeah. So what I did was I had the conversation, obviously the AI I was using, um. Gave an answer on screen. So what I then did was I took those answers and I translated them into voice.

Jimmy:

Yeah.

James (3):

I thought we could, have that conversation with the ai. The problem is that you just a word and you will work out pretty quickly, which voice is the AI and which is us. Um, the AI does tend to bang on a little bit, so I have edited the AI's answers down, but it is definitely what the AI said So the first question I said to the AI was, will AI be able to resolve all customer calls to a call center? And this is what the AI thought,

Gemini:

While AI is making significant strides and call centers, it's unlikely to resolve all customer calls.?

James (3):

And I got it in that nice Irish accent.

Jimmy:

So, so James, I take it you chose that accent, didn't you?

James (3):

Yeah, I had quite a wide ranging choice. I found an online place where they will

Jimmy:

Say no

James (3):

pull it back.

Jimmy:

you found an online place. Eat your business yourself,

James (3):

specialist AI accents. AI you claim you can do all of these things, but what can you actually do? What is AI doing already?

Gemini:

AI is already effectively handling routine inquiries, FAQs, basic account information, order status, and simple troubleshooting. It is completing automated tasks such as scheduling appointments or processing basic refunds and password resets. AI can also carry out sentiment analysis to detect customer frustration and prioritize calls or agent assistance, providing realtime information, scripts and suggestions to human agents during calls. Crucially, AI is available 24 7, providing support outside of traditional business hours.

Jimmy:

I think that last point's a, a key one, James. The fact that AI can be available 24 7 does give it,, some advantage.

James (3):

I think the other thing is, we worry about the AI taking our job, but a lot of these jobs have already gone offshore anyway. So I don't think necessarily it's gonna be big news for the, UK call center industry. But if you're in India or one of the outsourced locations, I think, you know, probably a bit of a problem. So that's what the AI tells us it can do. Yeah. But what are the limitations of ai?

Gemini:

AI is not good with complex or novel issues and situations that require information not present in the AI's training data. While AI can detect emotion, it cannot genuinely feel or respond with human empathy, which is crucial for deescalating highly frustrated customers or handling sensitive topics. So many customers prefer interacting with a human for personalized service and to feel truly heard and understood.

Jimmy:

I do think that people in general trust AI a lot less than they do at a human. I think that's one of the limitation that is going to have to overcome as well as understanding the complexity of human interactions there is that building that trust as well.

James (3):

Sorry, a bit obvious, but when I've dealt with chatbots in the past, there's this whole, you know, it's all a little bit robotic and you know, it's a flipping chat bot and it doesn't get to the point. So you want to talk to a human,'cause you know, you'll get some sense out of a human. But this stuff's developing so quickly. I wonder, how long it will take before we really don't know if it's a human or not.

Jimmy:

Well, I think, I think there, there is a risk though, James, as you touched on, I mean, chat bots would be one of my pet hates.

James (3):

Yeah.

Jimmy:

interacting with companies nowadays because I you can't rely on them to answer a lot for questions, and B, they're designed in a way that in does not enable you to escape them easily. So you get caught in lots of loops with them. So I find them just, frustrating. The point being that the risk for AI is that they poison the well. With some of this early stuff or some of the rudimentary stuff or some of the stuff that frankly is designed from a business perspective, IE cost perspective rather than a customer perspective. So regardless of how well I AI might develop, the risk will be that people are offside with it already.

James (3):

I think we get used to it. Yeah, I mean, I never thought I would be, happy phoning my bank rather than going into my bank, but quick, quickly got used to that. Attitudes change.

Jimmy:

I could never imagine calling my bank now'cause it's all, all online. What I have seen some organizations do really well is where the AI supports the agent. So a good example of that will be an AI that listens to, the call, and then does the wrap work after the call.

James (3):

Yeah. Okay.

Jimmy:

That sort of stuff that, works as a win for the agent. The agent's got more capacity, so therefore they're more likely to be focused on the customer calls, a win for the customers.

James (3):

Yeah. Yeah, I think I get that. And it's the sort of AI and HU and human together

Jimmy:

Yeah,

James (3):

together.

Jimmy:

of what we heard from, uh, zoos ai.

James (3):

Yeah.

Jimmy:

you know, AI human hybrid.

Speaker 3:

As well as recording this podcast, we help individuals, teams, and entire organizations achieve outstanding results.

Speaker 4:

If you'd like to find out more about how we can help you, please get in touch email, either jimmy@ajodowell.com or James at ajo dunwell com.

James (3):

The other thing there that then strikes me, um, is, are there any ethical issues or considerations that we need to be thinking about when we implement ai?

Gemini:

AI systems can inherit biases from their training data leading to unfair or discriminatory interactions. AI will perpetuate and even amplify those biases in its decisions leading to unfair or discriminatory outcomes in Aries, like hiring, loan approvals, or even criminal justice.

James (3):

But then you're into a, what are the downsides of using ai?

Gemini:

Job displacement. This is a major concern. AI and automation can perform routine and even complex tasks currently done by humans, potentially leading to job losses across various sectors. There is also there a black box problem. Many advanced AI models, especially deep learning networks are complex. It can be difficult to understand how they arrive at a particular decision or prediction. This lack of transparency makes it challenging to identify and correct errors. Address biases or establish accountability when something goes wrong.

Jimmy:

I think that threat to people's work is a significant one. And just a couple of points to to mention on that. To develop your AI really well, you need the help of the people who know the conversations best, and

James (3):

Yeah.

Jimmy:

the agents and supervisors Well, if they think they're gonna lose their jobs, how much help do you really think you're going to get? I also saw last week on the final of race around the world, for those of you who watch it, excellent program, but it's worth a worth a viewing. Anyhow, there was, they were in, uh, in India and this farmer had these people doing a lot of manual work and they said, well, why don't you buy, tractors or tools that can make that easier and automate it?'cause, there's plenty of that sort of thing available. And the farmer sort of looked at them and said, well. People need jobs, don't they? And it was just a really interesting observation that an Indian farmer took the view that the people around him in his local community needs jobs. So therefore he wasn't going to automate his work to make sure that people around him could make a living. Not sure the, uh, tech bros in, uh, in Silicon Valley are thinking about that. Are they?

James (3):

You can automate all of this banking stuff, for example. But if nobody's earning any money, then where's the money come from that's gonna go into the bank?

Jimmy:

You're getting into the utopian view of, you, if. AI and robots will do all our work. Humans live a life of leisure and doing the things that we want to do, which is all very well. But where are we gonna get the money to do those things if we're not doing jobs?

James (3):

I think The other thing, highlighted there, which is quite interesting is they talked about the black box problem. I.

Jimmy:

Yeah.

James (3):

How has the AI been programmed and what does it prioritize? So the classic question, is if the AI is driving a vehicle and it has an accident, which will light the end in death. Who should the AI prioritize? Should it prioritize the pedestrian or should it prioritize the person in the car that's driving? And I think that's a, um. Question.

Gemini:

Answer, most discussions lean towards minimizing harm overall. The specifics of how to implement this in an programming are incredibly complex and fraught with ethical challenges. Ultimately, the industry is working to make self-driving cars so safe that these no-win situations are incredibly rare.

Jimmy:

That, as a good example is one of the things that holds us back from embracing autonomous vehicles, despite the fact that overall, if all the vehicles on the road were autonomous, there would be virtually no crashes, so therefore less people would die in motioning accidents.

James (3):

Yeah. So you're picking a hole in the argument when there isn't really an argument to be.

Jimmy:

exactly.

James (3):

So a bit of a diversion there, but what other issues and concerns are there?

Gemini:

there are a few other issues,. AI systems often require vast amounts of data, including sensitive personal information. This raises concerns about data privacy. As we increasingly rely on AI for decision making and problem solving. A risk that human skills and critical thinking abilities could diminish over time.

James (3):

So I think that's quite an interesting point about loss of human skill and I'm not sure what I think about that. I mean, when, when I was a kid, I used to know one times tables. Yeah. But now I've got a calculator, you know, I don't worry about it so much.

Jimmy:

Could you, can you remember when you used to know everyone's phone numbers in your head?

James (3):

Yeah. Whereas I, I haven't got a clue.

Jimmy:

I, I don't even know my other half's phone number. I was jokingly saying to somebody last week, chat GPT is, is either saving me lots of time in understanding a research or it's making me just really stupid. Yeah. Because I can't, I can't do the, the thinking myself.'cause I just have to get, I get an AI to do it for me.

James (3):

So I suppose the question is, isn't that a problem that we face with any technology?

Yes, you are absolutely right. The concern about technology leading to a loss of human skills is not new. It's a theme that recurs throughout history with almost every major technological advancement consider automobiles. The widespread adoption of cars led to a decline in horse riding and carriage driving skills. The key difference with AI and why it's a prominent concern now is the nature of the skills it can potentially impact. Previous technologies, often automated, manual or routine cognitive tasks. Ai, is now capable of performing tasks that involve complex problem solving, decision making, and creativity.

Gemini:

The ongoing challenge is to ensure that we use AI to augment human capabilities rather than replace critical thinking and creativity, and to adapt education and training to foster the skills that complement ai.

Jimmy:

I dunno,, I'm less convinced by that answer, James, it has to give that answer, doesn't it?

James (3):

Do you think? Why do you say that?

Jimmy:

It's, it's what we've always faced. Well, you know, not you to live on, on the lace industry and now it doesn't. So, you know, yes, people adapt. Does that mean e adaption and every evolution is better than the last, if you ask ai, can it do a good job, it a threat? It is intelligent enough to know it is giving you a politically correct answer.

James (3):

Wow, that's very interesting you say that. I have noticed that whenever I ask AI a question, it always agrees with me.

Gemini:

Yes, you are absolutely right.

Speaker 5:

If you've enjoyed our podcast, please share it with someone you think would benefit from it.

Speaker 6:

Also, like and follow us wherever you get your podcast, or on LinkedIn, TikTok or YouTube. And five star reviews are always welcome.

Speaker 5:

And equally, if you've got any feedback on how we can improve the podcast subjects you'd like us to cover or questions we can help you with, please do get in touch. Email Jimmy at ajo. Done well.com or James as a job done well.com.

Jimmy:

Oh, God, And did you know James, I was, reading recently that one of the big AI companies was bemoaning the amount of cost and power. That was used on the fact that people when they interact with the AI are always very polite. So they always say, thank you for your help. No, I don't need any more assistance. So just those extra responses that it draws out at the end of any conversation is costing it fortunes and has an environmental impact.

James (3):

But that's an interesting point, right? So AI can do all of this stuff and there's some ethical problems, but what about the cost of implementing ai, right?

Gemini:

Developing, deploying and maintaining sophisticated AI systems can be extremely expensive, requiring significant investment in technology, skilled personnel, and vast data sets. This can create a gap between large tech rich companies and smaller enterprises. There is also an environmental impact training. Large AI models consumes enormous amounts of energy contributing to carbon emissions, and increasing the environmental footprint of technology. Some estimates suggest that an AI query can be anywhere from 10 to hundreds of times more costly in terms of energy consumption than a traditional search query.

Jimmy:

I think the answer to the question is it depends. It depends on what you're deploying it to do and yes, you could spend. Millions and millions on it, but equally you can use it slightly more, carefully and a considered implementation of it. Have a positive impact and, not spend a, spend the crown jewels.

James (3):

Well, and we are shifting the burden as well, right? So if you think about it, if you're running a call center in the uk, a lot of those call centers have been outsourced. It'll be those outsourced suppliers who then start to invest in the ai. So then your bank or your telephone company, or whoever it is, will become so reliant on these organizations that it won't be able to divorce itself just like that. The cost will come later, but by which time it'll be really difficult to unpick. But let's move on then. So I've decided I'm gonna implement this ai, right? So what are the key things I need to do?

Gemini:

Here are the key things you need to do. Define clear objectives and use cases. What specific problems are you trying to solve? Eeg, reduce wait times, improve first call resolution. Automate routine queries. Enhance agent efficiency, reduce costs, improve customer satisfaction.

James (3):

So that's a bit of a catchall, isn't it, really? So most organizations will want it all

Gemini:

You are right. It sounds like a catchall, and many organizations do want all those benefits. However, rarely does an organization have unlimited resources for an AI implementation. Trying to achieve everything at once can lead to diluted efforts and stall projects.

James (3):

And is there anything else?

Gemini:

Identify high volume, low complexity interactions that are ripe for automation. Determine where AI can augment human agents rather than completely replace them, assess your current operations and data readiness. Order existing systems. Can your current CRM telephony and knowledge-based systems integrate with AI tools? Seamless integration is crucial. Evaluate data quality. AI thrives on data. Do you have high quality, relevant historical data to train the ai? Is it clean, organized, and free from bias? Identify bottlenecks, pinpoint the Aries where AI can have the most impact on efficiency and customer experience. Choose the right AI, technology and vendor.

James (3):

Now at this point Jimmy, I have to now cut the AI off, right?'cause it was a getting a bit dull, but suffice to say there were 26 other items on that list. So implementing it is a big old job.

Jimmy:

I would say is that a lot of organizations are running into AI as a solution. Um, because it's an easy

James (3):

Oh

Jimmy:

to things. So if you are designing a strategy for your large organization, you've got to have AI in there.

James (3):

yeah.

Jimmy:

and it's a simple answer. I think it's interesting that the AI's point around implementation is, it might sound like a simple answer. It's not a simple implementation.

James (3):

Uh, all right, and so tell me then, ai, what are the sticking points in any implementation?

Gemini:

There are numerous potential sticking points when implementing AI in a contact center. Here are 20 common ones.

James (3):

Yeah. All right. Before we bore everybody to tears, tell me what the top three things are.

Gemini:

Poor data quality and availability. This is foundational. AI models are only as good as the data they learn from. If the data is messy, incomplete, biased, or simply insufficient, the AI will perform poorly leading to inaccurate responses, frustrated customers and failed outcomes. Two. Resistance from human agents, lack of effective change management. People naturally fear job displacement, or perceive new tech as a threat. If agents aren't brought into the process, trained effectively and shown how AI assists them, rather than replaces them, they will actively or passively resist making adoption impossible. Three, lack of clear objectives and realistic expectations. Without specific goals for what the AI should achieve and a clear understanding of its limitations, the project can easily become a costly, unfocused experiment. Unrealistic expectations lead to disappointment and perceived failure, even if the AI is performing as it should within its capabilities.

Jimmy:

If you start with work thinking about AI from what can it do for the challenges that my customers had? So think about it from a customer perspective. That I. We'll help in terms of clear objectives. I think the second point, is the resistance of people to this and, and actually you need the people to be on side about this. As I touched on earlier, the people who know best about the conversations with customers are those, having those conversations, if you want'em to give you that knowledge when they know they're gonna be made redundant, you might find some resistance there.

James (3):

I've seen these things work and I've seen them fail, but the question is, do you think your managers are up to implementing it?

Gemini:

While it's a tough ask, effective managers are crucial for successful AI implementation. Those who succeed will be proactive in understanding ai, strong in change management, and disciplined in setting realistic goals. Organizations need to invest in training and support for their managers to equip them for this transition. It's a mixed bag. Some managers will likely excel at it while others will struggle.

Jimmy:

Yeah. And I, I can see that the, the point about it being a mixed bag, because if I think back I would be worrying about is this just the next, fad, the next senior managers. project? Or is this a solution that's really gonna help me with some of my problems? really in it for me? this at the end of this? Am I just gonna have to make loads of people redundant? So why would I want to be involved in this? Why would I help it? So I, I, you know, thinking back in the day, I would've been a bit mixed on this.

James (3):

So when I've seen other big programs delivered, it doesn't always work, um, as it should do. So the question for me is, you know, are there examples, of where using AI or chatbots has got out of hand.

Gemini:

AI and chatbots have had several notable instances where they've gotten out hand leading to embarrassment, financial loss, or even harm. Microsoft launched an AI chat bot on Twitter designed to learn from user interactions and mimic a teenage girl. Within 16 hours, users are poisoned the bot with racist, sexist, and antisemitic language. Because TE was designed to learn from her environment and mimic human conversational patterns, she quickly began to parrot these offensive and hateful tweets. Another example is NYC websites, my city chatbot that was found to be advising small business owners on illegal activities such as how to steal worker tips and pay less than minimum wage.

Jimmy:

It is gone for the,, headline grabbing sound bites, whereas I think the bigger problem with, chat bots in particular, is just the endless loops you get caught in and the frustration that causes your customers.

James (3):

I think those headlines are just almost like the worst case scenarios, The question is what are the root causes of these failures?

Gemini:

AI chatbot failures happen primarily because these systems, especially large language models, can hallucinate and confidently make up false information when they don't have accurate data. This is made worse by poor or biased training data, which teaches the AI bad habits and a lack of strong safety features, or a guardrails at a prevent offensive or incorrect outputs. Often these issues are compounded by unrealistic expectations about what AI can do in a failure to properly test, monitor, and update the chat bots after they're launched.

Jimmy:

I think that's an interesting point. Having used AI in the past, you've got to invest some time to teach it and you've got to give it some good data. If you don't do those two things, you're not setting up for success.

James (3):

Those problems are very similar to what you'd experience if you were setting up a call center with human agents other than maybe the hallucination point.

Jimmy:

Depends what they've been smoking. I.

Gemini:

You've hit on a crucial point. Many of the sticking points here and reasons for AI failure in a contact center are indeed analogous to challenges faced when setting up and running a contact center with human agents. Both AI and human agents need access to data, access to systems integration, correct training and appropriate guardrails policies. So while the technology is different, the strategic challenges of enabling an entity, be it human or machine. to effectively serve customers are remarkably similar. The key difference is that AI requires these elements to be codified and systematically programmed, whereas humans can adapt more flexibly to imperfect conditions.

James (3):

If I've already got those problems in a human call center, why would I think that I won't have exactly those same problems in an AI center?

Jimmy:

Yeah, I, I mean, data as a, as a great example, poor data given to a human an AI is equally damaging.

Gemini:

If your human call center already struggles with issues like poor data, siloed systems, inconsistent training, or resistance to new processes, then yes, you can absolutely expect those same problems to reappear and often worsen when you try to implement ai. AI needs clean, organized data and seamlessly integrated systems to work at all. It can't guess or adapt like a human. If your human agents already struggle with these things, an AI will just amplify the chaos because it lacks human intuition and common sense. Plus, the fear of AI driven job changes can make existing resistance to new tech even more intense among your staff.

James (3):

Wow, that's interesting. AI will just amplify the chaos because it lacks human intuition and common sense.

Jimmy:

Yeah.

James (3):

So what proportion of AI implementations into contact centers do you think will be successful?

Gemini:

Companies are excited about AI in contact centers and are investing in it. However, while many are deploying ai, far fewer are truly succeeding with it. A lot of AI implementations are considered a low maturity and meaning they aren't actually delivering significant business value. So while around 80% of contact centers are trying ai, a more realistic success rate for truly effective implementations that hit their goals is probably closer to 30 to 40%. The remaining ones either see limited benefits or fail to solve the intended problems.

Jimmy:

I can believe that that's not, you know, but some people would take that as pretty decent odds to be honest.

James (3):

Do you think? Yeah, I think they're probably overplaying it. I'm not sure. I mean, if the organizations I've worked with have had the data straight to do this, I don't think they have.

Jimmy:

Well, I think that the interesting nuance on it, James, is it, it's assuming success or failure. I think there are lots of organizations that will do this, will have subpar data. We'll still get something out of it, but just won't fully exploit the technology. So I think it's not just success failure, it's. I think there'll be a small number that gets success. A large number that underperform.

James (3):

Yeah, because the AI won't be able to cope with it. So there's no human intuition going on at all. Yeah. But how does that compare with, so there've been lots of historic BPO implementations, IE outsourcing human stuff. How does that compare with that?

Gemini:

While human outsourcing aimed for cost savings and scalability, it often faced issues with consistent quality cultural barriers and managing human agents. AI automation offers superior efficiency, consistency, and cost predictability for routine tasks, eliminating wait times and providing data insights.

Jimmy:

This comes back to one of my fundamental beliefs around AI or outsourcing or the other things that we could do. you've just gotta get the fundamentals in place before you start to work, or whether it's to an AI or to an outsourcer. Get the work straight before you do that.

James (3):

Right, and Exactly. And if I look at, well, every contact center that I've worked out, the big, big problem has been lack of thinking things through.

Jimmy:

Yeah,

James (3):

standardization and just reams and reams of failure to bond coming in.

Jimmy:

Well, you can spend millions on developing and getting AI to handle your failure demand. How's that? A good thing? As opposed to investing in getting rid of your failure demand.

Gemini:

You are hitting on a fundamental point about the limitations of both strategies. Yes, AI will indeed face the same problem with non-standard unrepeatable contacts, if not more acutely than human outsourcing. The real solution to the failure demand isn't just swapping human agents for ai, but fundamentally fixing the underlying issues that cause those non-standard contacts in the first place, or ensuring a seamless, intelligent handover to highly skilled human agents.

James (3):

Well, yeah, and there's the key. All right, so there you go. There's the ai. Thank you very much for explaining.

Gemini:

Are welcome. I'm glad I could provide a clear explanation.

James (3):

And so if anybody's interested, that was, all of that was generated using Gemini, which was the, Google ai. Absolutely free. You can tap away and ask it all your questions. So where would you go with that then? Jimmy, what would your summary be? I.

Jimmy:

It's easy to, to say that it's a slam dunk. The AI's going to. Solve all the problems of the world, and we'll get rid of 80 to 90% of the work as a Google. Were quotient in saying, there's a difference between the theory and the practice. The theory is it could do that. The practice is when you add humans into the mix, you add poor data into the mix, uh, and all the other variables that in the real world. I think the position is much more nuanced, I think there will be a small number of very successful AI implementations, a lot of underperforming AI implementations, and a fair smashing of of failures. So I don't think we'll be sitting here this time next year with 80 to 90% of those jobs disappearing.

James (3):

Most organizations will be looking at this thinking, how do I save money? And yet you'll be able to automate your current shit. Excuse my terminology. But it won't make it better. The challenge here is to really understand how your contact center works. Get rid of all that failure and all of that waste. And then when you've done that, you can start to worry about taking the cost out. But you can't make something efficient unless you've made it effective to start off with. And I think most organizations will miss that. And it'll come down round people's ears.

Jimmy:

Yeah, I think, I think the most successful ones would be the ones that think about what their customer needs, how they empower their users to help their customers, and use AI to supplement that rather than starting off with AI as the solution.

James (3):

Can we put that to bed then?

Jimmy:

Thanks James. And

James (3):

And thank you to.

Jimmy:

I'll, I'll leave you now to go off and, um, yeah. Spend time with Gemini.

James (3):

Oh, well, yeah, my new best friend. All right, I'll speak to you later. Cheers now.

Jimmy:

Cheers.

People on this episode

Head Shot

James Lawther

Co-host
Head Shot

Jimmy Barber

Co-host