Automate problems, not just symptoms - Interview with Nick Clark of BCG

0:00Welcome to the next edition of the PunkCX podcast. With me today I have Nick Clark. Now Nick is a partner and director in the service and support operations practice at Boston Consulting. Group, otherwise known as BCG. He's also the author of the Service Matters newsletter, which I'm a subscriber to, which I really like. There's a weekly newsletter about customer service that he writes, he notes importantly, in a personal capacity. Nick, how are you doing? Welcome to the podcast. I know that I've been reading your newsletter for like a long time now. I really like what you do. I mean, it feels just very, I don't know.

0:34It's very insightful. It's very, it feels quite light touch, but it's got like heavy content in places which i really like so dan you do a really great job of sort of framing some of that stuff so thank you for coming to the pod on the podcast i think it's over way overdue so thankfully we got here in the end how are you doing and anything else that i missed from my intro that you'd like to add well thanks for that fantastic intro adrian and thank you for inviting me on long time listener first time caller so yeah it's a real uh privilege and a pleasure to be here with you. And yeah, that was a really nice intro.

1:08Thank you. Yeah. The Substack newsletter that I do, yeah, servicematters.substack.com. I try and get it out every week. But I find, yeah, I mean, it's kind of a Zen thing for me to do as well. It's a way that I can kind of settle my thoughts into what's the latest things I'm hearing from industry conversations, those sorts of things. And it's starting to grow a bit of an audience. So yeah, it's good to hear that it's something you like reading i was trying to think on in terms of other things i actually had two things that i wanted to add for my intro one is because this is the punk cx podcast i used to be a singer in a punk band when i was at university and that's something i'm very proud of i would love that uh there's no pictures exist anymore the other is the other on my uh on my bio is my something i'm extremely proud of is my first job coming out of college once i sort of given up the punk band idea was I worked on the phone.

1:58So my first job out of college was working in a contact center for a water company. It opened my eyes to a lot of experiences that I didn't really get at university, to be honest. And I still draw on a lot of those experiences today. So yeah, that's something I always feel very proud about. Awesome. Now, I mean, this is a bit of a sidebar, but like punk band, what style of punk and who were your influences? well so this was the late 90s so i would say the style of punk was the sort of u.s skate punk ska kind of thing so think like rancid or no fx or those kind of bands if those but actually for me personally that's not really my scene like i was much more into back then i thought i was very cool because i listened to a lot of stuff from the late 70s and i really like the sort of post-punk bands like wire and magazine and joy division and things like that that was really my scene and i guess even now that's quite cool i quite like that stuff but yeah i don't really listen to music as much as i'd like to anymore yeah oh you know you know that sort of whole no effects and sort of stuff coming out of the epitaph label which is the home of bad religion one of my favorite bands who are just cerebral sounds like a fancy word but they are and kind of like and also I'm a real massive fan of the whole hardcore discord scene that was from the other side of the American continent in Washington in Washington yeah particularly kind of like minor threat and the old Ian McKay sort of like involvement through to Fugazi who I think are just incredible I was just listening to actually have it on my phone.

3:39I have these things. And I was just listening to this track called Burning Two. Now, if you've never heard that track, that was written more than 30 years ago. And it's about environmental concerns and people's potential apathy towards some of those concerns. And the song is like so prophetic. It's quite incredible. And so any punk fans are listening in, you should definitely check out bands like NoFX, Rancid, the whole epitaph kind of catalog Bad Religion, and also check out kind of things like Fagazi and things as well. Not that you should all be listening to just solely US punk kind of like bands, but there are some pretty cool kind of examples.

4:22But yeah, dig into some old Joy Division, and the old New Wave Scar stuff as well, because some of that stuff is still incredible even today. Anyway, oh yeah, we're here to talk about customer experience of customer service, I think, rather than punk music, although we could probably talk about punk music all day. It could be a bit of a side project for you. Yeah, no, indeed. Now, I know that, as I say, we were trying to get you on the podcast for a wee while, and I'm glad that we finally got there. And there was a number of things we were kind of just spitballing in the run-up to the podcast that I wanted to explore, because I thought some of the things that you said I thought were really interesting, particularly because the position that you occupy in BCG and working with clients and seeing some of the challenges that they're facing.

5:10Now, you said this first thing, which was, you said that success with AI and customer service is much less about technical capability and much more about strategy. And I was like, ooh, let's not talk about technology. Let's talk about strategy. So tell me a bit more about what you mean by that okay let me give it a go i mean first of all i should say obviously it is quite a lot about the technology as well and you know that's not to be discounted there's some extremely clever people work very hard to you know solve problems like you know low latency for ai voice and stuff like that so i don't i wouldn't want to diminish that but i think often you know what a lot of companies are learning now you know having you know we're sort of three years in to the you know the latest ai wave and people are learning that it really isn't you know success really isn't about the technology you know that's that's the starting point and to sort of explain what i mean i thought it might be worth stepping back a bit and thinking about you know why did this sort of get too philosophical but like why does customer service exist in the first place um and obviously there's lots of different reasons for that but i think the more the kind of the closest to a sort of universal definition that i've got to in my career is it's really there to kind of fill in the gaps right for of of businesses and enterprises and other other organizations so whenever there's a process that doesn't quite work or whenever there's a a new product that's been launched but it doesn't have all the features yet because the company wanted to get the product out quickly to go to market quickly or where we have customers that have exceptional situations that nobody's planned for in the past.

6:51That's the role of customer service. It's that kind of glue that makes sure customers can get what they need from the services that they are paying for or using in any other way. And it smooths over those experiences. Now, the reason that exists is like there is some kind of trade-off going on there, right? Because it's for most organizations, particularly commercial businesses, it's usually better for their own profitability and for their customers, they would judge, to have a customer service team rather than trying to strive for perfection, like trying to go, there's never going to be any problems.

7:29There's never going to be any need for a customer to get help. You could, in theory, have that kind of perfect company, I guess. But the result of that would be, it would kind of, by definition, have to be a company that never innovates or never grows, like never develops new products. And that would be why you don't need a customer service department. And only have one type of customer. Yeah, right. Yeah, exactly. And only serve a very homogenous type of customer. And I don't think, I mean, who would want to invest in a company like that? Like that's not a company that's going to be going to be growing.

7:59So to me, like an essential ingredient of growth and innovation is the role of customer service to fill in those gaps and to enable the company to be flexible and to move. Now, the problem for a lot of businesses and any customer service organizations is usually as an enterprise as a whole, they don't have a firm grasp on this dynamic. Right. Real understanding of why is it that I spend so much on my customer service organization? Why does it cost me so much? Why are calls queuing? Why are customers complaining? and that link between everything that's going on in an organization that's creating that growth and that innovation and what that's doing in terms of the downstream impact on customers, that link isn't always that strong.

8:47And so I think when it comes to AI, the first thing, me as well, the first thing we all thought of when AI really suddenly became really popular again is chatbots or voicebots. But that's really about automating the symptoms. And if we automate the symptoms without really understanding the problems, then it doesn't work for lots of reasons. Like often the automation doesn't really solve the real issue the customer needs. Customers get frustrated. People don't use it, get a bad reputation, those sorts of things. And so I feel like at the moment, the most powerful way to use AI, especially this generation of AI, is to use it to understand that dynamic.

9:28So to get much better insights in what's really causing our customers to contact. When they do contact, are we helping them effectively? Like, do our processes actually do what they're supposed to do? Are we making the right trade-offs? That kind of thing. I mean, do you think, though, that here's the thing, is that a lot of the time, I mean, that is not a problem exclusive to the development of or the evolution of AI in its current kind of like form. I mean, that's a problem that's existed kind of like for millennium or not millennium. That would be a bit of an exaggeration. But for decades where, you know, almost access to the insight was always available to people.

10:11They just had to go and find it and spend the time understanding the problem and thinking about what's the best way to respond to it. I remember having a chat with somebody who was doing a project for a financial services company, and they were looking at trying to improve their mortgage application process. and they were this was before this was a few years ago so it was before the advent of generative AI and they said well okay what we're going to do is we're going to employ a bunch of data analysts or like bring a team in from outside and gather all the data and crunch all the numbers all the data and the multimodal kind of data and we're going to identify all the main kind of problems we're going to come back and it'll take us a few weeks to do that and one bright spark Mark kind of said, well, I'll tell you what they did off their own bat.

11:07They gathered up sort of around about 10 different agents on their lunch break, arranged for some teas and coffees and a tray of donuts, told their supervisors to bugger off so there was no influence in there. And they said, right, collectively, let's see if we can identify the top 10 kind of problems that customers can have with the kind of the mortgage application kind of process. They went through that and they identify all these different kind of problems. Now, they had a highlight. Now, six weeks later, the data kind of team came back and went. And what they found was that in half an hour, the agents were able to cover 80 % of the problems in half an hour, rather than actually resorting to spending thousands of pounds and wasting six weeks to get to be like 100 % right and identify all the kind of the problems.

11:57And so do you not think there's also this problem where we default to technology to look to try and understand the problems rather than actually realizing that the intelligence is there? It's just distributed in a way that we sometimes don't appreciate and we don't necessarily spend the time and the effort to go and try and access it. and I think that's more to do with it's a problem with us trying to use the tools that are available to us and our own wherewithal to try and understand the problem rather than rather than just relying on the tools as it were to be those kind of like these surfaces, these are the nuggets as it were they're there, you just need to they've always been there, you just need to go and find them or want to find them yeah, that's a really interesting observation I think it's, and there's, I've always seen, there's always a difference between, you know, you can really feel it in the culture of companies where, you know, the leaders will make a point of spending time in contact centers, sitting next to agents listening to calls or, yeah, running those kinds of workshops that you talked about, or even, you know, even talking to customers themselves, God forbid, you know, compared to what, you know, and that really makes a difference.

13:16because I think you get that innate understanding of, you know, the trouble is if you just look at the data, particularly in the old days when the data was just manually coded intent graphs, you could look at the data, say, if you're a bank, and it would say the number one reason people call is to do a balance inquiry. Or if it's a telco, the number one reason people call is to check their bill. And then you say, well, why are people calling for that? It's easy. You've been able to do that online for 15 years. like why why we go and do this and it's only when you listen in that you realize actually the ones that are calling up not always but by and large it's people because there's something else it's not just they're asking about their balance they're asking about their balance and they don't recognize something about it they think something's wrong and they need some help and so getting that innate understanding of that complexity i think is really helpful yeah but to sorry just to finish that off i think where we're seeing the real power is the combination now of the human side of listening to frontline agents who can talk about their experiences and combining that with at-scale generative AI analysis of large core volumes.

14:27And tools that we have now, we can have those workshops for the agents and they can say, oh yeah, there's this problem because customers always call up about this thing. And then they always call up again because it hasn't worked. and what we wanted where that may have fallen down in the past is everyone kind of knows that's an issue but they've never really been able to quantify it and so it's never got on to the work it never got to the top of the backlog yes but now because we can analyze all of those calls at scale we can have a conversation with that data and say tell me how many customers do mention this when they call up and we can get a natural number yeah and we can quantify it and that creates you know a proper business case and i think that that's quite powerful when we use those things together it.

15:08Absolutely. I mean, it's something I've been musing on. I'm used on kind of like, I think it was last year or something, or maybe this year. I can't remember exactly when, but I said that I was talking about success in terms of implementing AI into customer service. And I said, there's two types of AI. There's agent intelligence and there's artificial intelligence. And when you combine those two, that's when you start to get the results. And it's a bit of a play on words, I understand that, but it has a deeper kind of point. but I really like that and the idea about understanding the organizational realities around you can speak to a whole bunch of people and they can identify some things but if you don't understand the scale of the problem it doesn't necessarily go up the priority and that's just I guess real politic in an organizational sort of like sense but I think also the whole thing around understanding the problem then that helps if you spend the time understanding the problem then that's really going to help you formulate what is the strategy that you need to employ in order to propel your service forward rather than just kind of like, well, you just want to fix, fix, fix, fix, fix, fix.

16:14Actually, the plan should always be, how do I elevate? How do I understand? And then how do I kind of try and just rather than filling in the gaps or fixing the leaks, how do I elevate it onto a next level where then hopefully that problem, the recurrence of that problem starts to go away. Right. you know. So the other thing that you also said was that you also said that something along the lines of that you said that brands need to go beyond the obvious with AI and that companies are differentiating themselves by doing more than automating existing kind of processes. And I thought that was really kind of interesting because it feels like there's like, here's the herd and then there's the people over here that are doing something kind of very kind of different.

17:02So Amirite, can you explain to me what you mean by that? Yeah, this is something I personally feel very passionate about. I think that the industry, we need to move away from these, more radically away from these historic customer service models that have been designed around basically constrained levels of human resource. so customers have to navigate you know complex IVR routing they have to queue before they can speak to somebody or chat with somebody when they speak to someone that person's in a rush because their handling time has been restricted or the call gets transferred because actually that person isn't an expert there's a very limited, even more limited number of experts to actually speak to to get the real answer or it goes to a back office and you're having to wait And all of this is about, historically, all these tools have been put in place to throttle that demand because the people that are available to handle the demand are constrained, particularly at certain times a day.

18:07And the really exciting thing about AI is we don't have to do that anymore. AI isn't completely unconstrained. There is a cost to it. And I might talk about that later. But we can definitely think differently. so this year for example i've been working with a bank in the middle east where they're using data and ai to make their service proactive right so for example if you're on the app and you're trying to make a payment and it hasn't worked they're not waiting for you to call up they have a chat bot will start to have a conversation with you there and then and start to help you resolve it but if you do call up the voice bot that you speak to will also know that you've just tried to do that and we'll straight away be getting into kind of helping you right so not forcing you to go through all these like hoops because we don't have to anymore we can just start solving the problem straight away another example another bank uh this one's in brazil they this bank's been built basically so their entire servicing model is through whatsapp right it's a continuous conversation it's mostly delivered through ai sometimes there might be a human answering but It's just a continuous stream of conversation.

19:16So it's, I mean, talk about low friction. Like there's nothing to stop you from just getting the help that you need. And because, again, it's mostly being delivered by AI, that human constraint isn't coming into play so much. And so when, you know, the idea that you, to me there must be hundreds of models that haven't even been thought of yet. Once we move away from this idea of we've got a throttle demand, then, yeah, that's when it gets exciting. So what's interesting to me about what you kind of said, one is that the first example, I think, is interesting because it speaks to that omnichannel promise that was never quite delivered.

19:55where it was around how you can retain context across different channels, and therefore it becomes one conversation and their ability to do that. But a lot of it's to do with how you can, I guess, deploy in channels, but then retain all the context in a conversational layer. And that's fine. But then I also really, really like the Brazilian example because it also speaks to one of my pet sort of, is it a peeve? I don't know. Maybe it's a drum I beat. And it's this sort of like this idea that many brands talk about, oh, we have to be everywhere on all the channels that our customers are. And I'm like going, hmm, do you really though?

20:40And because I really like the idea, and I think many of the leading brands, some of you standout brands kind of do this really well. They don't necessarily have to explain themselves. they just kind of go, we're going to concentrate all our resources on these channels. And if we do that, we can dedicate more resources to it, and then it can increase its efficacy and so on and so forth. And in many ways, that sort of signposts itself. It's almost like excellence over here, by the way. I don't know why you're going over here, because we're not over there. We're over here, so we'll travel. We know that we're going to do this.

21:10We do this in real life, right? I always use the example of if there's a restaurant that's meh, kind of like it's one street over, but there's an even better one that you've always wanted to go to, two streets over, of course you're going to travel two streets to go to it because it's just a better experience. And I think the same applies in the digital kind of realm, but we don't really kind of like bust that assumption in many brands. Go back to the point around strategy. It's like, it's not technically, it's about strategy and about thinking about understanding kind of the whole problem. So I think what I'm taking from the kind of examples you're sharing is that actually you say, rather than go beyond the obvious with the AI, people are actually starting to be, and sometimes it's not the mass, but it's almost the outliers are starting to be really thoughtful about what they do and where they deploy it and how it kind of promotes better outcomes and how it elevates the experience and everything else.

22:09And that feels to me that is things are shifting, but it's taking a while for people to sort of re-examine some of the assumptions and the ways of doing kind of things in order to drive these kind of outcomes. Is that fair, do you think? oh yeah i think so i like the i love the restaurant analogy and i think yeah like it's funny that you know omni channel is a word that's been around i don't know at least 10 years probably longer and it's it i think of it now it's a bit of a flawed it's an ideology and that's the problem it's the kind of flawed ideology and i think one of the big flaws in it from a customer service perspective is it's so based around this when people think omni channel they think yeah channel switching and you can switch channels and we'll retain the context but i think in servicing the minute a customer has to switch a channel you've already introduced unnecessary friction for them and so you know and that's not to say you can never do omni-channel i think there are some industries like an example like airlines i think actually omni-channel works a lot because if someone is actually on a flight or they are at an airport they are likely to need different channels to get help maybe they need to go to the desk they call they chat that makes sense but in a lot of cases yeah like i think where technology is allowing us now is you can create excellent service that does everything in a in a single channel or in a smaller group of channels like another great example of that would be i'm sure people mention this on your podcast all the time but octopus energy of course have a you know great reputation yeah their preferred channel for service is email yes just like phenomenal because everyone else hates email right but they've really focused on how do we make email work really well and the way they've done it is you know the way that they organize their teams so that it's a small group of people look after a group of customers and so when you send an email it's not just going to a random queue it's going to someone that's served you before and they put a lot of ai into that as well to make that work but i you know i love that.

24:12I think it's just going beyond ideology to actually being smart about what the customers really want. So, I want to circle back to maybe the first point you talked about before, about it's not really about the technical capability, more about the strategy. I just want to ask you, because there's another thing I've been thinking about, and it maybe speaks to that sort of Octopus Energy example, also the bank in the Middle East and also maybe the bank in Brazil. And it feels like one of the things that, and I always think that this might be a bit of an Achilles heel about some people's approach.

24:47And it feels like oftentimes there's a lack of a vision of what they want their service experience to be. And I think that then feeds into that lack of the vision because it's almost a bit like now obviously technology is moving so fast so you have to educate yourself in terms of what's the art of the possible and that informs your kind of like kind of vision but if you have a limited sort of like vision then that then leads into your strategy and so and so forth i mean do you think that's accurate and people aren't necessarily spending enough time because i think i look at the thing about the octopus energy example i'm thinking that's brilliant that feels like it's a story that they've told themselves who says, we want to get to a point where we are operating on a channel that pretty much all of our customers are comfortable with, but also that they get to speak to people that they've been in touch with before.

25:45So there's a continuity of it because it's all about relationship building. We need to think about how we can do that in order to drive better outcomes and how we can use the technology in order to deliver that. And that just feels like a beautifully crafted sort of like vision and story and everything. Okay, now we need to think about how do we deliver that? Now, we might not be able to deliver that now, but we can build towards that over the course of years. I mean, do you think that's kind of fair that there's this lack of a differentiated vision kind of in many cases? Yeah, and I think, you know, a lot of customer service organizations, you know, they have incredible leaders.

26:23Like people really care about customers. They care about their teams. They really know how to do the job. It's very impressive. And however, sometimes or often operating in a context where the vision can feel constrained before you even start. We've got to hit a budget. We've got to, you know, here's a whole load of KPIs that we need to hit. And I think the task of any customer service leader that wants to really make a difference is to persuade your C-suite or inspire your C-suite that customer service is something to be interested in as a broader value driver, ideally even the CEO. And actually, this is one thing where, although AI, it's inescapable at the moment, there's definitely a bit of hype on the AI.

27:18But this is an example of where hype is good. Yeah. Because actually, in the last few years, customer service has become a bit more of a C-suite agenda topic than it had in the past. And so that's the opportunity to come with that bigger vision. And that bigger vision is, yeah, it needs to be things like, how should our customer service reinforce our brand? We have millions, maybe tens of millions in some organizations, hundreds of millions of conversations with customers every year. How can we make those the most valuable brand building conversations that we have, whether that's with a human or an AI or a combination?

27:54So, yeah, that's how when things get exciting. But, you know, I appreciate when you're stuck in the day to day and delivery, it's not always obvious to start doing that kind of stuff straight away. No, exactly. When you're kind of budget and kind of resource kind of like constrained, I mean, the bandwidth that you have for creative kind of thinking and pushing the kind of boundaries kind of like converges to zero very, very quickly. So I understand it's quite an idealistic position, but I think it's, to your point, is that sometimes you have to advocate for yourself. And that would be a long-winded advocate for customer service leaders advocating for their own importance and value add within a broader organization.

28:41because actually the reality is, and this is kind of what's been showing up with AI, is that Contact Center or any customer service or customer support team is that you're sitting on top of the biggest real-time kind of data set in any company anywhere in the world. It's like if you want to know what's going on in a business, go and speak to people in service, and they will tell you exactly what's going on kind of any time of the day about anything, billing, products, kind of like outages, all these different sort of things. And it's like, it's, and for people not to be, and I think that's the opportunity is if you are not leaning into that and trying to organize that and support it and invest in it and then type all of those insights into other parts of the organizations and get them all connected into that data set, then you are totally missing an opportunity.

29:30And if you want to, if you want to, if you want to grow, all businesses want to grow, especially at the moment. Yeah. The contact center is a fantastic source of unmet customer needs. Customers every day tell you what they don't have, what they want. And some of those things might be things they're willing to pay for, but you've got to go and listen to find that out. And I think the thing is, going back to the complaint you talked about before, is that about the cost of AI. This is a thing that I've been mulling on a little bit, is that we think about AI, sometimes we think it's cost neutral. Well, it's sort of not.

30:06and whilst it might can help you make some decisions about headcount and stuff and I hope people would lean into this idea that it will free up resources and you can redeploy them to other kind of places there's also this other angle where that if you're just deflecting and helping people self-serve it doesn't eliminate those kind of problems and actually there's little cogs that's whirring in the background, helping people solve those kind of things. And as you add more products and more services and complexity kind of goes in, then those cogs get bigger and kind of bigger. So the cost of AI, there's like this hidden cost that could just grow over time.

30:53But I wanted to kind of just, I don't want to prejudice kind of what you're going to say, but I was like, I just wanted to circle back and ask you a bit more about what you meant by the cost of AI sort of thing, because I think it's a really fascinating kind of thing. The way we're talking about it, we're talking about it in terms of investment, but thinking operationally that it's possibly cost neutral, but I don't think it necessarily kind of is. I mean, yeah, there are levels of this, aren't there, I think. On the fundamental operational level, the cost of an AI that can perform the same task as a frontline agent, you know, blended, it's roughly about 20 % of the cost at the moment.

31:38So, you know, at that level, if we have people that are doing work, which is, you know, transactional, simple, and we can offload that, then that, and, you know, often then give a better service to customers as a result. um the what what else is going on there is that usually in most organizations we work with there's a whole latent set of demand that isn't being met at the moment so you know you free up the capacity that has been working on that not you know relatively low value add simple things actually there's plenty of spare need for additional additional resource to to to do things which are value adding.

32:22And, you know, we've worked with companies that have, you know, that's in a number of ways, you know, companies might reinvest that in just having shorter queues, or they've reinvested it in retraining those. So some of those frontline reps to become salespeople and, you know, to really to grow customer loyalty and grow value. And that's all happening there. Yeah. And I think the other thing that we at the moment are underestimating is the cost of the new organization that is needed to continually build, shape, govern AI. Because that isn't a thing that stops. So both the AI doesn't stop because the tech gets better and the models evolve and change, but also the needs change, regulation changes, products change, those sorts of things.

33:12And we also need to keep differentiating. So the good news is some of the organizations we're working with now, they're retraining, upskilling and redeploying frontline representatives or their managers to be working on tasks which are more about that. So the new skills to build, improve AI. So that might be doing, creating knowledge-based content, improving the knowledge-based content. But it might be even more technical skills, like actually managing the quality of the AI that's coming out, the prompt engineering, even supporting around data security, those sorts of things. So there's a lot of new capabilities that were needed to make it work.

33:57Now, I still think it's on a like-for-like basis, there's a cost saving there. But often it's more like when we reinvent for something which is radically better and is creating a lot of value and is helping us grow, that's a different story around cost because it's all productive cost. No, I mean, I agree with that. I think my point really was it's more complex than it seems on the surface. Yeah, definitely. I think the fundamental kind of thing is that, and I've been saying, again, I've been saying this for, exploring this sort of idea for a little while, is that the fundamental nature of enterprise software has changed.