What happens when the most frustrating part of customer service, waiting on hold, repeating yourself, and fighting your way through endless phone menus, finally starts to disappear?
In this episode, I sit down with Neil Hammerton, CEO and co-founder of Natterbox, to talk about how AI is reshaping customer experience in ways that feel practical rather than theatrical. We begin with a conversation about the gap between what customers have tolerated for years and what they expect now.
Whether it is a bank that still puts you through layers of outdated IVR menus or a service team that answers straight away and solves the issue, those experiences stay with us. Neil makes the case that voice is far from dead. In fact, he believes voice is becoming one of the most exciting places to apply AI, especially when businesses want faster, more human interactions at scale.
What I found especially interesting was Neil's view that AI should be treated like a new employee. That means training matters. Tone matters. Context matters. If businesses want AI assistants and agents to succeed, they have to teach them how the organization works, how conversations should sound, and when a human needs to step in. We talk about the difference between using AI for simple triage and using it to complete tasks end to end, from handling password resets to helping callers outside office hours or during spikes in demand. Neil also shares why the smartest path is rarely a giant leap. It is usually a series of smaller, lower-risk steps that build confidence and real results over time.
We also get into one of the biggest concerns hanging over every AI conversation right now, whether these tools are replacing people or helping them do better work. Neil's answer is refreshingly balanced. In many cases, AI is taking care of the repetitive jobs that frustrate staff and slow down service, while freeing human agents to handle the conversations where empathy, judgment, and experience still matter most. That shift can improve customer experience while also making work more rewarding for the people on the front line.
There is also a strong message here for business leaders who are still stuck in pilot mode, testing AI without ever quite moving forward. Neil explains why smart pilots need clear goals, good training data, and realistic expectations. He also shares how Natterbox is using AI internally, including producing board packs in a fraction of the time, while still keeping people involved to check, challenge, and refine the output.
This episode is a thoughtful look at where customer experience is heading next, and why the future probably belongs to businesses that know when to let AI lead, when to keep humans in the loop, and how to blend both into something customers actually value. What are your thoughts on the balance between AI efficiency and human connection in customer service, and where do you think businesses are still getting it wrong?
Useful Links
[00:00:04] Welcome back to the Tech Talks Daily podcast. Now, a quick question for you all. You know that moment when you call a company and before you've finished your first sentence, you're already wondering how long you're going to be stuck in a queue. Yep, we've all been there. You've heard that message. Don't you know you can resolve this issue online? Push this number for that. Answer this question.
[00:00:26] And all this occurs at the same time where we find ourselves living in a world where AI can handle a natural conversation. Anyone that's got the OpenAI app can learn about almost any subject while driving their car. It can help us prepare for meetings, help me prepare for this podcast today and even resolve problems in seconds.
[00:00:48] So, why does customer experience still feel so uneven and almost like you're going back in time when you pick up the telephone? Well, my guest today is Neil C. Hughes. He's the CEO and co-founder of Natterbox. And he's right at the center of this shift that we're going to be talking about today.
[00:01:07] And together, we'll explore what it really means to bring AI into customer conversations, but do so in a way that feels human, helpful and genuinely valuable for both the business and the customer. Everybody wins. And this isn't about ripping out contact centers or chasing the latest trend.
[00:01:28] I want to go a bit deeper than that. I want to help everyone rethink how work gets done, how employees spend their time and how organizations can meet the rising expectations that we all carry from our best digital experiences. So, today we'll talk about why the most successful AI projects start small, how to train AI in the same way you would train a new member of staff, and why the future of customer service isn't human versus machine.
[00:01:57] It's human with machine. And we'll also have a refreshing amount of honesty in this conversation around fear, adoption, and the reality of getting beyond pilot mode into real improved business outcomes, measurable business outcomes.
[00:02:13] So, if you've ever wondered what a great customer experience looks like in the age of agentic AI, or how are you going to introduce automation without losing the personality and the trust that your customers expect, this episode's for you. But enough from me. Let this Neil introduce you to another Neil right now. So, thank you for joining me on the podcast, Neil. Can you tell everyone listening a little about who you are and what you do?
[00:02:43] Of course. Yeah, nice to meet you, Neil. Thank you for inviting me on. So, Neil Hamilton, I am CEO and co-founder of a company called Natterbox, which is a UK-based, well, I suppose it started as a telecommunication company, but we now brand ourselves as a customer experience platform. And you're quite forthright in challenging the idea that IVRs are unfit for purpose. When I was doing a little research on you, that's one of the things that stood out about you to me.
[00:03:10] So, what has changed technically that make voice AI far more capable today than in previous generations? Because I think we've had so many bad examples. Well, a lot of previous examples feel a little bit sluggish now and in need of an upgrade. I think the Amazon device is one example. I don't want to say the name because everything will start going off in here, but what makes it more capable today, do you think? Well, technology has moved forward.
[00:03:37] But just to answer the first part of that, I mean, IVR, interactive voice response, I mean, it's kind of not, though, is it? It's you phone somebody up and it's press one for this, two for that. You get through, you go to the next layer, you go to the next layer. That's frustrating enough, but it feels as though we've almost accepted that's how things work. And the frustrating thing for me years ago was actually, I can't say names, can I? I had to ring a media company in the UK and every year their devices went wrong.
[00:04:06] Almost like clockwork every year and every year I'd phone back and there was about seven different layers. And I'd listen to all the thing. And I, at that point, thought, why is it they can't recognize it's me calling again and ask me, are you calling for the same issue you had last year? And if the answer is yes, put me straight through to the department, save me about an hour of going through all these different things.
[00:04:29] And that's, to be honest with you, that's what gave us the idea for Natterbox is there must be a better way to treat customers when they're calling to your business. Everyone's a winner. So to answer the second part of your question, technology with conversational and gentic AI now is incredible. So you can phone up and have a right old chat. So you don't even have to say, I need to speak to your sales team or technical support team. You can phone up and say, well, I'm struggling today.
[00:04:55] You know, I've got a bit of a headache when I got out of bed, maybe had a drink too many last night. And by the way, oh, the reason I'm calling is my widget seems to be fluttering and not working so well. And the AI will make a decision. Oh, I can see that you bought that, you know, less than a month ago. Let me put you through to this team that will replace it. Or is that a warranty? Let me put you through to a technical support person that can help you with that. And that, for me, is the exciting part of customer service here.
[00:05:22] That AI can actually do that job for us better than old-fashioned IVRs. Yeah, completely agree with you. And I was speaking to someone on here a few weeks ago, and he was talking about that when he goes on a three-hour road trip or something, he will have OpenAI's assistant on there, and he'll have that conversation and learning about things as he's driving. And you compare that experience to, I don't know, Siri or Amazon's digital assistant, and that slight delay, which just makes it almost unusable. It shows you how much things have changed.
[00:05:51] But what are you seeing actually driving adoption in customer experience right now? Is it cost pressure, customer expectations, talent shortages, or is it something else entirely? What are you seeing driving this? Well, it's twofold as far as we can see. First of all is the desire to improve the experience people have when they contact businesses. And the second one, which is always, is to try and reduce cost. And they kind of go opposite normally, don't they?
[00:06:19] If you reduce cost, generally the equation is that you decrease customer experience. You improve customer experience by investing more money normally. But with AI, the latest generation of AI, we can do both. We can reduce cost and improve customer experience. So by way of an example, if you call into a business rather than have to sit on a whole queue or go for an IVR tree, just have somebody instantly answer.
[00:06:44] But I've always said that one of the best companies I know for customer experience is actually a company we use for our travel, our corporate travel. They're not a customer in Atabox. They don't even have a phone system. But if you call them 24-7, they pick the phone up within about two rings every single time. And the human answers who knows what he's doing as well. And that is just tremendous. And that's what AI allows us to do now for every business.
[00:07:11] You can have somebody that answers your phone within two rings actually knows what they're talking about and can help you. It's so interesting you say that because I don't know if I'm just a grumpy old man, but I'm getting to the stage where if I have a business that makes me jump through all those hoops and push this for that and say this and confirm that, I've started removing them from my life to save my frustrations. And I have two bank accounts. We won't name any names here, but I do have one which is like a bills account.
[00:07:38] And if ever I do need to call anyone, I have to jump through those hoops. So they're going to be going soon. But I have another bank account. And as you said, that example, they answer within two rings. It's a human. I get an answer straight away. So they're in my good pile. But it is that old phrase, isn't there? The last best expectation we have anywhere. That becomes the standard expectation for what you want to see everywhere. And I think we all feel that as individuals, don't we? Yeah. And that's a good thing, right?
[00:08:04] So I think the expectation now is that things should be improving. The old press one, sales two for support, we are now not so happy accepting that. Certainly in the circle I work in, obviously, I'm obsessed with this now because it's my business. But I think the general public are starting to catch up that it doesn't have to be as bad as it used to be. It can improve now. And as we record this, it's 2026, year of agentic AI and agents.
[00:08:31] And Natterbox have also launched AI assistants and AI agents that can resolve up to 70% of routine queries. But what does that actually look like in a live enterprise environment for any business leader listening? And how do you ensure that those interactions maintain that personal touch that we've just talked about that is so essential in customer service too? The first thing I would say is in my head and what I always say to people is just imagine AI is another employee that you're hiring.
[00:09:00] So if you hired a new employee into your business, what's the first thing you need to do? You need to explain to that person how your business operates, what's expected, and you need to give them training material and make sure they understand what you're training them so they can actually do the calls. In an ideal world, you would put them next to somebody who's experienced and they would listen and learn that not just the technical answer, but the manner in which you deliver that and the way you engage with your callers into your business.
[00:09:29] And AI is very similar. If you expect AI to do a really good job, you need to actually do the training with it. You need to allow it to listen to how humans interact. And it's not just the technical answer, it's the style as well. So having AI answers that speaks in a certain way, whether it's formal or casual, whether it's fast or slow.
[00:09:51] So if you've got a demographic of people that are perhaps hard of hearing, you want the AI to speak a bit slower and clearer than perhaps if you're, you know, to use a sort of a stereotypical on a trading floor where time's of the essence. You might want the AI to speak really fast and cut corners in the way in which it speaks. And that's the ability that we have now with the AI that's out in the wild.
[00:10:14] And just to come back to you on your timestamp, Neil, you said this year, this month, with AI, you need to bring it down to what day of the month. It's moving that fast. It really is. And you're in somewhat of a privileged position here because you're speaking with businesses all around the world and they're going to be coming to you and asking for your help and implementing this. And I'm curious, from all those conversations, if you put all those in a big melting pot, what are enterprises you're working with? What are they getting right about AI adoption in customer experience?
[00:10:43] And also, where do you see the most hesitation or internal friction? Because there are a certain amount of cautiousness involved with the certain industries around this too, right? Yeah. So it's like many things. It's early adopters. You've got individuals out there that work in corporates or even smaller companies, and they are passionate about moving the needle and they will be early adopters on technology. You've got other people that are a bit more cautious that will go much slower.
[00:11:11] So what we find is the baby steps first, doing steps of low-hanging fruit. So, for example, if somebody calls into a business at, say, 2 a.m. in the morning, there's probably a reasonable chance that businesses not got any staff there. So the person calling in gets told to call back or they can leave a voicemail, perhaps. Why not have AI answer in those situations? So it's a low risk because you're definitely going to improve the experience because the AI can take a message as well if it's not able to help resolve the issue.
[00:11:41] Then the next level would be, what about if someone calls into an organization, say an enterprise, and everyone's on the phone? So rather than go into a queue, why not have the AI take the slack? So you put it as a lower priority to the human. And if you do that, then that covers everything. If, for example, there's a tube strike and no one can get into the office, then AI can answer all calls because no one's in. Or if somebody's sick or there's a peak in traffic.
[00:12:07] So by doing that, you keep humans, what we call humans in the loop, and maybe human first, but then AI to back up. And the next layer after that, once we achieve that with clients, is to look at what's the most repetitive items going on in your call center or in your business that your humans actually don't like doing. And it's actually a detractor from their job to the point that it increases staff churn rates in businesses. So we're all the same, right?
[00:12:34] If we have to do something we don't enjoy and it's repetitive, we're less likely to enjoy what we do. So questions like, I forgot my password. Can you send me a password link? Or what time do you open tomorrow? These repetitive things that are quite common. That's the sort of thing that AI should be doing, in my opinion. So when you call into a business, AI can do that triage or what used to be IVR and say, oh, hello, Neil. Thanks for calling back to Acme. What can I help you with today?
[00:13:02] And if you turn around and say, well, look, yeah, I'm really sorry about this, but I've lost my password again. Can't remember it. The AI can solve that. It's no problem at all. Would you like me to send you a link that you can click on? Now, even if you turn around and say, well, actually, I haven't got access to that. I'm driving down the road at the moment. I really need to change my password. So I'll give it to my partner or something. And then the AI could go, no problem at all. I need to identify you. It can do that whole identification, verification, ID and V.
[00:13:29] Once it's successfully done that, then it can say to you, OK, can you give me a password that you'd like me to make? And you can have it as a one time only. You then say to the password, it's a one time use. So somebody can log in and you can log in and change that. As an example, there's no need for humans to get involved. You get a much better experience because you've got no hold where you're driving. The AI can sort that out for you. But the moment you say, OK, I need to do something a bit more complicated or I'd like to speak to a human, please, then the AI should pass that call across.
[00:13:59] No question. Just like a human would. You spoke to a junior member of a team and you wanted to speak to somebody more experienced. The junior team member would pass the call across. And those password resets is such a huge problem. I come from an IT background and I know firsthand the frustrations of having a service test that I've got. I don't know, they want to add value, they want to get involved in projects, but they're handling 200 password resets a week or even more sometimes.
[00:14:27] And to further bring it to life, when I was doing a little research on you guys, there's a great example of We Buy Any Home. And they're handling 1,000 plus calls every month without an overflow of service. A compelling use case. But what were the practical steps allowed them to reach that outcome, doing an extra 1,000 calls a month? And what can other businesses maybe learn from that journey? Anybody listening, what could they learn from it?
[00:14:53] Well, that's an example of a forward-thinking company, to be honest with you, where the company wants to embrace technology to further improve customer experience. So We Buy Any Home have two things. As it says on the tin, I suppose, they buy any house or any home. But they have a lot of people phoning up, asking questions like, would you buy a house that's got tenants? Would you buy a house that says council house?
[00:15:22] Would you buy a house in this situation or in this location? And quite often, I'm told, the answers would be given to the caller. And the caller would go, great, I just wanted to know, actually, yeah, I'm not going to sell my house, but thanks, bye-bye. So it's not a qualified lead, in effect. And that type of call, as you can appreciate, takes time from a human. And somebody who genuinely wants to phone up to try and sell their house can't get through because they're in a queue. So now the AI can do that qualification. So the AI can answer all the questions correctly.
[00:15:50] And then if the person wants to sell their house, say, one second, will I get a colleague for you? Puts the call straight through to the colleague. But to avoid repetitiveness, the AI summarizes what's been discussed and the outcome and sends that on screen to the human. So the human now gets the call. He's reading on screen. And when the human's ready, presses the button to say, I'm ready to take the call. Meanwhile, the AI is still engaging with the caller. And it's a very seamless experience for all parties then.
[00:16:17] And what came out from this was that, as we said earlier, the interest level from the staff went up. At first, when they heard AI was coming into their business, they were concerned for their jobs, like everybody seems to be. But once it was in, they actually got more excited and embraced it because they said, now every time the phone rings, we know it's a qualified lead for us, in effect, as opposed to potentially a tie kicker. So that's quite interesting.
[00:16:43] I'm glad you mentioned the concern around jobs there, because there is a big debate about whether AI will replace contact center agents and entry-level roles. But from your perspective, how do you see AI and human agents coexisting? And how should organizations think about designing that balance and getting that balance right? Because if you are angry or you've had a frustrated experience, sometimes you do just want to speak to a person, don't you?
[00:17:08] Yeah. So as we said earlier on, people want to improve customer experience and also potentially save costs. And that's the holy grail. If you took a, let's take a more typical, a typical business in the past that had just humans, and sometimes you would call them, you would have to sit in a whole queue. We've already talked about that. If you found a business that had no whole queue, that would mean they'd have a lot of humans in their contact center. Now, in that scenario, AI comes in and takes some of the jobs.
[00:17:35] But most businesses just have IVRs and queues. So you're not replacing jobs. You're just making the customer experience better. So when people call in, they get answered straight away. And you're also allowing humans that are in the contact center to do the calls they want to do. So you're giving the repetitive calls to AI, and you're giving the better calls to the humans. And there's no necessity to eliminate jobs in that situation.
[00:18:05] But that's down to the business, right? Any business with or without AI can turn around and go, right, I'm going to save 20% of my costs. I'm just going to make my customer experience worse. That's their decision in their business. It might have a short-term benefit, but not necessarily long-term. We will have some people listening who might have been burned by delivering ROI or solving a real problem with their AI projects and struggling to get it out of pilot purgatory.
[00:18:30] So when businesses look at AI for customer experience, I think many are tempted just to go all in and attempt a large-scale operation. But I know you're a big fan of smart pilots. So what does a smart pilot look like, and how do you avoid getting stuck in endless experimentation, which many have seen over the last couple of years? Yeah, good question. Small steps, but step by step. I would say low-hanging fruit first.
[00:18:57] So the first step is to plan what does good look like for our organization? What would we like to achieve here? Then consult with a company like us that has experience so we can describe the art of the possible. And then we work out steps to get there. It's like many things in life. If you're trying to do too much too soon, it gets frustrating. It goes wrong. You get dark days. It's much, much better to take it in small steps. But I have an example, Neil. We as an organization use AI internally quite heavily.
[00:19:27] And we are pushing really hard to increase the use of AI as a productivity gain tool. And I've noticed if we try to do too much too soon, it goes wrong. People get frustrated. They actually burn and become very inefficient playing, in quote marks, with AI tools. And the same could be true for our service that we offer the world. If people try and do too much too soon, then it can go horribly wrong. For example, we said at the start that AI needs to be trained so it knows what to do.
[00:19:56] If your training material isn't good or is incomplete, then you're going to get a poor result from the AI. It needs to be good background training material. So why not train it on how to do ID&V first and stick to that before you teach it how to do something else? And I'm curious, outside of what you're delivering, what you're offering to the world and your customers, what big successes have you had internally with using AI? Anything you can shout about there or just genuinely positive experiences with AI?
[00:20:25] I think doing exactly what I've just said, small steps. So rather than say to somebody, here's an AI tool, use it to improve your efficiency. And everybody's interpretation is different. What we do is go, here's an AI tool. When you're doing this job or this job or this job, this is how you use it. Other jobs continue as you have been before. And it might be the next week we can tick another one off that you can use AI to benefit. So classic, this week I've got a board meeting. AI produced a board pack.
[00:20:54] And it did it in a fraction of the time for me. And it's able to do that because we've connected it to many of our backend systems so it can actually get all the data. But I don't just let it do the board pack and send it out. We get a board pack. And then I go around all of the different divisional leaders that have a section in the board pack and ask them to check to make sure that the information is correct in their opinion. And then once it's checked, it's Q8 in effect, send it off. And that's what you have to do. Such a great example there. So much food for thought.
[00:21:23] I'd love to hear what people listening are taking away from today's episode and maybe share some of their experiences. But before we go there, I'm going to have a bit of fun with you before I let you go now. I always ask my guests if they'd like to leave a book they'd recommend to our Amazon wishlist or a song for our Spotify playlist. Now, guilty pleasures are allowed. Even maybe your favourite hold music. I don't care what it is. But what would you like to leave everyone listening as well? So I'm not very academic with books.
[00:21:51] I love Lee Child, Jack Reacher, to be honest with you. I've read all of those. That's just unwind. Although one of the best books I've ever read by an author, I think he's only written two books, is a guy called Alan Falsham. And I can't remember what the book's called, but really interesting. So anybody should look up Alan Falsham. Outside of fiction, I suppose the most inspirational book I've ever read is Richard Branson's Losing My Virginity. And that, again, was another pointer that encouraged us to start Natterbox.
[00:22:18] Because in that book, Richard's talking about how he started Virgin Atlantic. He saw a need and he decided to start an airline. He knew nothing about it. And he made one statement. And he said, I believe that anybody can learn a business well enough to be an expert within three months or 12 weeks. And when we had the experience I told you earlier about the media company I was calling, I decided to challenge that and say, well, can I start a telecoms company knowing nothing about telecoms? And we did.
[00:22:48] And there was four founders here. None of us knew anything about telecoms. And yeah, here we are. So it's inspirational, that book. Song-wise, you've probably published this near now I say this, but I don't, I'm not really a music person. So I don't know. Ellie Goulding always makes me, I don't know, emotional. It's a thought-provoking song with the words from that Vincent and Van Gogh. So, yeah. Awesome.
[00:23:15] Well, I'll get all those added to the book wishlist and the Spotify playlist. And I don't want to get too emotional, so we'll change the subject very quickly. But seriously, for anyone listening that just wants to find out more information about anything we talked about today, all things Natterbox and what you're doing and announcements coming later this year, where should they go? Definitely the best place is natterbox.com, our website. Excellent. Well, I'll add links to that.
[00:23:42] And I'll also add a link to your personal LinkedIn as well in case anyone would like to connect with you. But we've covered so much today about what's driving AI adoption and customer experience, why both AI and human agents are required for optimal CX, and why starting small scaling with AI works best for businesses. As I said, I'd love people listening to contact me and you. Let me know what you thought. But thanks for joining me today and bringing this conversation to life. Great. Thank you, Neil. Pleasure.
[00:24:12] So many things I loved about this conversation, especially about how practical it was. I mean, strip away the hype and you're left with a simple idea. Start with a clear outcome, take small steps, and use AI to remove the tasks that people never wanted to do in the first place. Sounds simple, doesn't it? But when you tick those boxes, customers get faster answers, employees get more meaningful work, and the business moves forward without the disruption that so many fear.
[00:24:43] And I think Neil reminded us that technology on its own never fixes anything. And the real change comes from how you introduce it, how you train it, and how you bring your people with you on the journey. Ultimately, it's about treating AI like a new hire. Give it the right context. Let it learn your tone, your values. And suddenly, that experience could become seamless rather than mechanical.
[00:25:10] So if today's conversation sparked a few ideas about you and your own customer experience strategy, your internal productivity, or even how you are personally using AI in your working day, I want to hear about it. You can find more information about everything we discussed at natabox.com. And as always, you'll find links to everything else in the show notes. And connect with me on techtalksnetwork.com.
[00:25:35] And finally, as you probably noticed, there were two Neils on the podcast, both guest and host. And one amusing little peek behind the curtain here, one of the first things Neil said to me before we started recording was, do people ever call you Ian? And we both had a big laugh about this because I don't know why it is, but many people confuse the name Neil with Ian. So I want to do a straw poll on this.
[00:26:01] Any Neils that listen to this podcast, have you ever been called Ian? That's one for me to leave you on. So let me know your thoughts on that too. Any experiences? And I'll be back again tomorrow.