The NiCE World London Takeover of the Punk CX Podcast Part Two

0:00So welcome to the Nice World London takeover of the Punk CX podcast. This is part two of a two-part series of chats I had with nice executives, one of their clients and an analyst while here at Nice World London. In the first part of part two, keep up, I am delighted to say I am speaking to Taz Chowdhury, consultant and former strategic delivery lead of Birmingham City Council. I think that's right. Is that right? That's correct. Yep. Hello and welcome to the podcast. How are you doing? I'm good. I'm good. I'm actually not just at Nice World but in a new world for myself personally. Okay. Setting up my own consultancy, trying to bridge the gap between the suppliers and the customers.

0:35Fantastic. And sort of sharing my experience and the story that I've had with Burmese Scouts. Perfect. Well, you could just literally preempted my nice question was asking if you're going to do a bit of a background sketch on yourself. But that's kind of brilliant. But let's get into, well, we're here at Nice World London. And there's part of their big tour around meeting their customers kind of closer to where they're at. And there's been a lot of sessions, a lot of talk, keynotes and things. What have been some of the big highlights for you at the event? What have you seen that you've got? Oh, that's cool.

1:02I love that. I think they've done a great job of actually simplifying what AI is doing. Okay. You know, there's been a lot of talks previously coming into this about all the technology that goes into it, a lot of strategy. but actually you know seeing face to face and up front some of the way that the technology is properly implemented the tasks that it can can action okay you know it's nice it brings people a bit closer to what is tangible i think for a lot of people at the moment ai it's new it's exciting yeah probably feels a bit far away for them yeah coming here actually lets them know it is within touching distance there are certain things that you can do and actually you don't have to do it all straight away yeah i think it's also that it struck me as well as like some of the stories i know that you spoke on the stage kind of behind us kind of earlier on it's it's kind of open and quite candid yeah about like oh we did this yeah and i saw it didn't work or it did work but up to a point and then we had to kind of roll back and then started doing it again and as there your approach kind of matured um but talking about that i don't want to kind of give the game away I wanted to ask you about the Birmingham City Council sort of a sort of example the story you were at the heart of that helping them to kind of do that but before getting into what you did give me a little bit of background about what was the challenge what was the problem that you were trying to solve you know before you stepped into this kind of journey with with with nice so prior to my being at Birmingham they had gone into section 114 administration so they got no money right yeah yeah so it's very hard to do something when you've got no money so we we did what any good council does and we spent lots more money on some consultancy to tell us where we should go and spend some more money okay right to create some savings to bring some money back in right so that's what happened before i got there we were guided um whether it was the right or wrong thing at the time.

3:07I think it was the right thing. Automation is always a good idea if you're trying to achieve savings and there are manual processes and real transactional tasks that can be automated. You save lots of time, lots of money, lots of resource. Now, voice automation was the suggested guide for us. Go to the contact center, call into one of those trick IVRs. You speak to a you know somebody that almost sounds human intelligent voice agent yeah intelligent voice agent and um try and reduce some of the some of the capacity that flows through to the person that's handling calls at the end of it now we went with a really basic sort of setup now like i said this was before i i got there not not saying that to try and abdicate responsibility no no not at all you know but what they did a lot of really good work or started some really good work before i got there they were on that journey and something that i just want people to try and understand i said it on stage which understanding the the different types of uh of ai that's available you've got your deterministic and flow model you know very much boxes and arrows you know very much like the old ivrs it's flowed it's controlled you know what you're going to get Then you've got the generative AI, it's based on the LLMs, but it's very much about a response.

4:30You ask it something, it tells you something. Really simple way of looking at it. Probably the best way is for you to understand it. And then we've got what we've now moved into, the agentic space, what Cognigy offers, the kind of tools that are at our fingertips. And it does something. That's the difference. Now we're in the space where we've got AI that has flow, responds to you, and then does things for you. So it's going beyond just generating a response for an email. It's actually able to execute tasks on your behalf. And where have you sort of focused that kind of on in terms of the, you know, because, you know, a council, Birmingham, is what, 1.5 million residents, maybe 2 million?

5:16Yeah, something like that, yeah. in the kind of the area so it's like you're covering a lot of people um so what area of the council were you focusing kind of on to start with the contact center affects everybody so voice automation every any every call that came through was going through that nlu based voice automation okay when we ended up moving to nice we pulled it with the cognitive platform there's a real battle between what we'd achieved with aws which we'd achieved some good things and what we wanted to do going forward. So we were almost at risk of putting a Ferrari engine in a Fiat Panda, right?

5:55That'd be fun to drive. Yeah, it would be fun to drive. But not the ideal fit. No, no, you know, we've taken on this really incredible tool. Sure. And we were trying to replicate what we'd done before. Now, I get it as a starting point. People were really, they were used to building out those voice automation journeys in a certain way. They were used to having certain conversations, understanding how they were supposed to build those concepts and iterate those journeys within AWS. But it took a while to accept that that wasn't going to be the same with Cognigy. We had an agentic soul, back by an LLM.

6:34We could choose to be a lot more agile. We didn't have to spend as much time developing. um and i suppose for for a time period is the lack of control that people felt and the lack of control is based on the fact that it was different yes not necessarily that we had no control but when you've moved to a more of an agentic but also proactive type of uh approach that's gonna have like data and infrastructure sort of implications and so i guess to coin the the phrase like what sort of ducks did you have to get in in a row before you could to facilitate all that sort of connectivity? Well, if you're going to be really strict, you want to get your processes, your data, your line of business systems, your integrations, if there's any integrations, understanding what kind of integration that you need that's going to achieve through MTP or APIs.

7:27Lots of different things that you want to try and get in a row. There are actually some things that you want to continuously work on in parallel. data is never going to be perfect. No, no. Okay. The other thing is like completely spick and span data. No, there isn't. Because it's not a fixed thing, right? It's not a fixed thing. It's not a fixed thing. And it's even less fixed when you've got an organization that has the same customer in 20 different systems. And actually that customer is input into 20 different systems. It's not that they just exist as an entity because that data is flowed into it.

8:04It's Because we physically input that customer there. Sure, sure. So, right, how are we going to go through that type of data cleansing and data matching at that scale? It's not that it's impossible, but to do that before we do anything with AI would almost create a blocker. Right. So you choose one. You choose the most robust one that you've got the most confidence in. You match it up with the relevant journeys. You figure out what can be achieved with that data that's available. And then you almost run it as a pilot. Yeah. And then that's you proving your assumptions and putting yourself in a position to say, we need to continue, you know, cleaning our data so that we can do this everywhere.

8:44And I guess it's a process of building confidence as well. Absolutely. And that's the thing that people don't necessarily pay as much attention to is the idea that you talked about there's change and we've done it this way before, but we move to something else which hopefully has a bigger, more future-proof potential. And then it's still different. and people have gone like oh but it feels new and we're doing the different and do i need to build confidence in this and that's some of the soft side of things in an organization also but is a really important thing that doesn't necessarily get the attention that that it often deserves and that's why i was really purposeful about trying to break down and simplify how ai can differentiate in a different modern thing yeah that message rarely gets across yeah you know and it doesn't have to be this big scary beast that some people worry about that hallucinates and then goes off and does its own thing and then suddenly you know it's terminated to judgment day right so yeah i know indeed it can be achieved again going back to what's what's available or what's been available here it is tangible yes and it it can be something that people do understand and they can understand it intimately without being able to build it themselves right and so what the results been what sort of like what's been the response from your from your citizens or the residents of Birmingham, what sort of what's been the outcomes that you've been able to drive?

10:08I'll be honest, people want to talk to people people want to talk to people and I can be completely honest about that, there's nothing wrong with people wanting to talk to a person, right? We're humans, that's the way that we're built we're built to talk to other people but what we want to try and do is if they're not going to talk to a person is how good are we at making sure that there's an outcome for them sure we're getting better at that we're not perfect yeah right we started off at like a 25 capture rate we've improved that just that oh i say we i'm no longer there but it that's improved that will continue to improve as we build out the remaining journeys in Cognigy yeah and we're able to use the investment in the technology to review what's available so the AII giving us the data the analytics to almost re-automate what's there.

11:01It'll tell us where the next best opportunity is, what should we do to improve the current journey. We're moving into that direction. There is so much to do in terms of the build and the layout from moving from one system to another that we couldn't do that straight away, but we know that we're on that journey. But actually customers have been receptive they've been receptive because they're not having to wait and also if it gives them access to answers and potentially resolution like 24 7 and this is the thing is that you don't turn the lights off that's the automation kind of at the end of the day and because many many residents will be working kind of people they'll have jobs and they don't necessarily have time to go and do this other stuff so being able to give them that opportunity to be able to self-serve or increase the capacity to self-serve and not have to work to office hours is probably going to be welcome.

11:55And people know that people are smart, right? They kind of understand that things get better over time. People are smart, but the phones are smarter. And what they're realizing is that the phone can do a lot of the work for them. So that self-serve thing that you're talking about goes beyond just the voice automation. It's the opportunity of, okay, we'll send you a link. we'll send you a text with a link to go on and pay your counsel tax right or we'll send you a link with the relevant information that's gonna give you the answers that you're looking for and you're no longer having to wait in a queue to get that answer from a human being that you can get off the website sure so that the opportunity to self-serve is only gonna be greater the more we're able to access the data that we talked about yeah of course so So whilst it's limited right now, the opportunities will continue to increase to do more to clean our data.

12:50And we do more in terms of embedding the different line of business systems, giving the access to the different systems to the Cognitive Platform. The one thing I want to ask about, I saw that the council's up for an award. Congrats on that. And finally, big shout out to Sheraz Yacoub, who is the head of customer experience at Birmingham City Council, who's named in the award. that I wanted to ask specifically about this approach that the two were told about innovation has been anchored in a human-centered design and an ethical AI kind of adoption is that something that you that tell me explain to me about that so they what what does that mean in practice so it's I mean it sounds you're very inclusive we wanted to create journeys that were simple and easy for people to yeah to access them and if what we didn't want to do was over complicate it by making it too agentic if you like you know yeah right we don't want the customer sat there having a full-blown conversation and end up talking about you know their lunch and their dinner and telling us how they do so we had to put the right measures in place make sure that the outcome is clear and it's achievable.

14:05But we also did a bit of work in terms of the language modeling and what different languages the book could accept. Yeah. And the translations were available. So there was a lot of work that went in, especially from our user research side and UX designers. from the original piece of work we were able to carry over as learnings and it did make a big difference. We were able to build off really solid foundations and move things forward. I think you got in the part of the design process, did you get residents involved in some of that as well? Because it's to help enable sort of or to try and drive.

14:44Well, we got them to pick the voice. There you go. We put that out to a vote. We had a number of different voices. They picked a voice and that was the person they wanted to speak to when they called in. Perfect. So there was lots of different opportunities for us to get the Razor Spon. Obviously, we got them to test it as well. Yes. We were doing the first wave of it. We go in, call, you happy with that? Does that sound good? Yeah, we're happy with that. Can you tweak this and tweak that? But then you have become in danger of becoming the Tinker Man and constantly tweaking, getting tweaked and getting tweaked.

15:16We've got to start somewhere. Sure. So we had to accept the baseline and then move on from it. Absolutely. And then another thing on people in general, I mean, you've involved the residents in the process. And what's been the impact on the service agents themselves? Because a lot of people, they might see some of the automation and go, oh, that's different. That's kind of things. How has it impacted their roles? Has it created space for them to do more and different sort of things? How has it kind of changed? Or is it a little bit early to tell? I mean I can give you my opinion on it and I think that for the customer services they've got to start moving away if they're going to bring in agentic AI they've got to move away from this the old metrics to measure success so average handle time, classic measure how long on average is an agent handling the call well if we're putting agentic AI in place to deal with all of the transactional stuff.

16:19We're only going to be left with the complex queries. So trying to measure that metric in the future won't reflect the same way that it used to. No. So now you've got to think of a different metric that you want to review. So customer experience, so how do you measure that? Well, we're already qualifying how good a call was. Yes. So we do a lot of quality frameworks and we can automate that. But really, how are we able to pinpoint the customer journey and how good was the customer's experience on this call? And I think that's the direction that we need to go if we're going to really measure what success looks like.

17:03Because once it's all baselined and we've, to an extent, either removed or redeployed those agents elsewhere, and we're now having more meaningful and complex conversations, having a half an hour conversation isn't something you should be worried about how good is that half an hour conversation how good is that customer's experience how likely are they did we answer all their queries in that one fall were we able to address more because with the pressure wasn't of us to end that call and move on to the next one yeah because we've now got the bandwidth to do it i think that's a really important point is like you can't use an old frame of reference to measure what you're going to go to do in the future and i think this is something that people need to get their heads around is like going you're talking about aht and FCR and et cetera, et cetera, deflection and containment and all these different things.

17:46Some of those are still relevant, but actually they're not 100 % relevant. They're 100 % map into this kind of future kind of like in a model where it's about the outcome. It's about did you help somebody achieve what they needed to achieve? You know, how did the agent kind of perform? And can you benchmark that, I guess, your best performers and can look at do this automated kind of coaching to try and help them, everybody improve and so on and so forth. So it's like a completely different sort of like kind of frame. yeah absolutely and that's that it's but it's fuzzy right now and it's becoming clearer as we get closer to it i mean you can take certain steps where you time when you deploy a certain thing so you could you could deploy all of the automations on the operational side before you do anything with ai so that'll be like auto summary auto scoring for the quality framework you can get the system to score how good the calls were without also evaluate.

18:48There are a number of other things that we can save in terms of operational time before we actually look at the impact that AI can have on what calls are coming through. So by the time you've done all that and you've reduced your AHT as much as possible and you're basing on the classic measure, At that point, you can say, well, actually, we can ignore this now. We know we've done as much as we can to save the operational time. We're now going to introduce the agentic AI and we're going to remove all of the transactional calls. So we know that we're only going to be left with all the complex queries.

19:21So now Aht is no longer a good measure. So now we've got to consider a different way of measuring what good looks like for our customers. Yeah, perfect. So one final question, as you may. May. It's been a bit of a journey with Birmingham. But we know that there are people that are keen, there are different parts of their journey. Some people are not starting, some people are just starting, some people are a bit more advanced. But I think in the majority, there's a lot of people that are just getting going, as it were. And we're still early into this new kind of era. People that are just getting going, based on your experience at Birmingham, what would be your best advice say here's some of the big things that we learned at Birmingham and this is kind of what I would I would advise you to think about or keep in mind going forward I think the approach a lot of people spend a lot of time a lot of money invest a lot in getting to a proof of concept and then running a proof of concept but if you are really clear on what your goal is and what you want to achieve and you're happy to be a little more focused run a pilot and set yourself up to be able to scale that pilot out in a live environment yeah so that you aren't you aren't waiting to then have the approval after you've done a run approval concept to then go back and say oh okay well we're proving the benefits here we go so you run it as a pilot the money that you spend building up to approval concept and you're supposed to running it as a pilot yeah probably it's about net neutral at that point and you're in a better position at the end of it to scale up because you've created an alive environment and and you're probably going to be able to see some return of investment in the pilot if you're really clever and you've done the work in figuring out which journeys you want to automate first so you're almost using the pilot to create the investment opportunity for the next wave of uh of automation it's almost as if like you're saying stars stars you mean to go on yeah which means give yourself an escape route but stars you mean to go and just get going absolutely so like don't guess at the proof of concept so i think it's like going pilot we we are committed to going kind of down this blah blah we're in talent because you'll do all the work anyway and they'll just get you going and so you end up saving time and money and then the president's ultimate goal i mean it would be great if you know an organization comes they see an opportunity to automate they invest in that single journey that single journey then creates savings those savings are then reinvested in further automation scaling out what that platform's able to do right that's that's where they want to get to so if you do your homework and you're really confident you're able to prove your assumptions and then set up the pilot and run with it you know you're going to save time and money perfect um taz anything else you'd like to add yeah give it a go take take the risk take the risk um don't overcomplicate things.

22:18I always go by three really simple questions, right? It's like, what is the problem? Do you know what your problem is? And this applies not just to AI, just in broad spectrum. As a consultant, I try and answer three questions really well. What's the problem? Who's the audience? And what's your metrics of success? You can answer those three things really well. You're in a good position to be able to take the next steps. Awesome. We'll leave it at that. Tashyadri, thank you so much for sharing your time, your insight and your expertise with us today and congratulations on the success that you've helped engender up at Birmingham.

22:52That's been great, so thank you so much. Thank you very much for your time. Been a pleasure. Take care. Cheers. Cheers. So welcome back to the Nice World London takeover of the Punk CX Podcast. This is the second part of part two. Keep up, this is maths here. And I'm really looking forward to chatting with Oru Mohyuddin, who is Research Director at IDC. Hello. Hello. How are you doing? I'm good, thank you. Fantastic. So lovely to have you on the podcast. Can I ask you just to give me a bit of a background sketch, an introduction to kind of Oru and the work that you do? Right. So I am with IDC.

23:27And as you've mentioned, I'm research director there. And my focus area is customer service and CX. Fantastic. Now, we are here at Nice World here in London, part of the world tour, bringing the Nice World message to customers around the place. there's been a lot that's been talked about, right? Different kind of technologies, different sort of customer case studies, all these different sort of things. Sometimes it feels like, wow, it's overpowering. What have been the highlights for you? What are the things that ever stood out? Right, so one of them is intelligence orchestration, hybrid workforce, including AI and humans, and A2A interactions.

24:03And when I say A2A interactions, I don't mean where one AI communicates with another AI to delegate the workforce. the workflow. What I mean is that AI talking to another AI on behalf of a consumer or a customer. So in the future, let's say it will be my AI talking to a brand AI to do a transaction. And I think that is really interesting. But these three things have stood out to me because it's just doing things so differently. It just fundamentally changes everything. and that's why these three things stood out. Wow. I mean, so I think you're talking about machine customers. Yes, machine customers.

24:46And actually, when this goes live, I think it will be two weeks since I just released a podcast talking about machine customers with a lady called Katia Forbes who we're talking about all that. It's fascinating sort of stuff. I mean, it's going to change all sorts of different sort of things. So yeah, no, it's very future sort of focus. I mean, I love that. But I wanted to ask you also, because you've come from a research background you're an analyst in this sort of space now we know that the you know this cx ecosystem this space is it means really crowded right and it kind of it can feel really confusing and you talk to a lot of different vendors you kind of you map the marketplace i'm from a strategic positioning perspective i wanted to understand you know what's working best for vendors in terms of how they strategically position themselves so that they stand out in customers' minds, but in what is a crowded marketplace?

25:43Right. So a vendor that stands out is a vendor who's not selling technology. Ah, love it. It's a vendor who is a transformation partner. Okay. So by that, I mean, you're not helping to scope out and design the IT, but understand the customers need to transform and everything that is involved in the process. So obviously technology is a part of that, but there is a cultural side to that. And it's also helping them transform the culture. So the main thing is to think out of the box, be open-minded, be receptive to changes, to think things differently. And it's to educate brands about those requirements.

26:33So, for example, you don't think of contact centers as cost centers. Think of contact centers as revenue centers. So, how do you measure the success of AI? Are you looking at the traditional KPIs or do you look at new set of KPIs? So, it's like thinking things radically differently. And it's the role of the vendor to help them change their mindset. and that is being a true transformation partner. So if you want to stand out in the market, you have to be the transformation partner rather than just sell technology. I think it's important, I would ask, I think it's important to add to that is like when you talk about partner, it's not about positioning yourself as a partner.

27:13It's not just about telling different stories and better stories, that's part of it. Of course. But actually it's about, as you said, doing different things. The thing I saw, which I thought was really interesting, was how they're making an investment in this the nice AI labs and they're sort of creating some new kind of applied research, like benchmarks and things, but also using some of the insights that are coming from that to inform their forward deployed engineers that they're going to be pushing closer to the customer so that they understand that some of this stuff is complex, right? Absolutely.

27:47And they're taking some of their own expertise and pushing it closer to the clients to help them be successful, to be that transformation. Exactly. Rather than going, oh, you buy it, off you go, good luck. Yeah, and that's it. We don't know you anymore. And on to the next customer. But you're absolutely lying. Right. You know, that does, you know, that is a standard feature for NICE where, you know, you are establishing yourself as a thought leader. Right. You're discovering new things and you are sharing it with not just your customers, but also the academia. Yes. Right. So you're spreading the word.

28:19You're going beyond just selling technology. And that's what I mean. Like, you know, if you want to stand out, you have to do things differently and really drive the value of AI in its truest sense. Yes. And so I think that's what's interesting. But there are no, there are no sort of, there's no, we don't know where we're going. Because we're laying the road as we go. So I think, so it's like there is no clear map of what's going to do. So it's about almost you have to highlight and iterate and sort of experiment in this kind of partnership. Some stuff might work, some stuff might not work. It's a new world that we're actually.

28:58I think that's really important. But then also one is just like flip it because it's not as if brands, they're not passengers in this. They are active participants in this because they have to partner and they have to be. What are you seeing from a brand side of things in terms of some of the challenges that they need to overcome? in order to be successful in this? And also, what are they doing to get things right and to accelerate their sort of development? Right. So let me first talk about challenges. Yes. And then I'll talk about what they need to do differently. So in terms of challenges, there are many.

29:33So let me simplify that by putting it under three categories. Okay. So first is data. Yes. Second is organizational alignment. And thirdly is compliance. Okay. So data, which we all know about it's scattered, it's incomplete, it may not be good quality data. And your AI feeds on these data. So rubbish in, rubbish out. So what's the point in having AI if it is churning out rubbish stuff? So you need to make sure that the data is right. So that's one challenge. The second challenge is about organizational alignment. And that's more kind of diverse. So it means not having the right technology stack, like being stuck with legacy technology.

30:19Sure, sure. Then it's also the siloed operational structure. You know, one team not talking to the other team, one stack not on the other stack. Then it's also about resistance. Yeah. Right? A lot of the times employees think, well, you're deploying AI. That means that you're getting rid of me. So I don't want to accept it. So there is pushback. And then you have unclear ROI. How do you measure the benefits or how this AI is delivering results? And it goes back to my earlier point where you become the transformation partner and help them understand the true value of AI. And it's not looking at AI or understanding or evaluating the value in the traditional sense, but looking at new set of standards, matrix, in order to value AI.

31:15So in terms of organizational alignment, what we're seeing is that brands are still operating in the old ways. and in order to transform or in order to deploy AI, you also have to work in a new way and that is a huge leap. So that's another challenge. And the third challenge is about regulatory compliance. So there are some regulatory matters that you have to comply with. For example, if your AI goes rogue or it leaks confidential data, it can be very damaging in terms of financial consequences too. So these are the main challenges. I think also it's very heightened if you operate in a regulated and compliance-heavy sort of environment.

32:03And so there's over and above the brand risk and the reputational risk that comes from that. If you're in a regulated environment, it's an even bigger fanatic risk. Absolutely, yeah. So it's a new thing. And so therefore, being able to govern it, being able to audit it, being able to kind of like control it is kind of really important. And especially in the context of Europe, where you have stringent regulations when it comes to data privacy. So security and safety of AI is prevalent across the world. It's relevant for anywhere. But in Europe, you have GDPR, you have data privacy. And if you violate that, the consequences could be severe.

32:44not just in terms of you know denting your brand reputation but also having financial consequences from the government yes one final thing because i know we're getting to the end of the day everybody's tired it's like it just wants to get out and kind of relax take a breath fresh air one final question before we kind of wrap up or is best advice we know that people are are considering getting embarking on this kind of journey some people that are thinking that people are further ahead than they are, but actually not. There's a lot of people that are still getting started. What would be your best advice to a brand that is going to embark on this journey, that wants to embrace some of this technology to drive better outcomes to both the customer, their employee, and their customers, employees, and their businesses?

33:32What would be your best advice to them? What would you say, do this, you'll be set. Don't just jump on the bandwagon. don't do AI for the sake of AI do AI for the right reasons it's for transforming your organization in order to be relevant in the future and once you have that realization then you will do it in the right way it would be about scoping out where you want to solve problems then finding the solutions but also when you're doing that you will associate it with cultural changes And then it will be part of your strategic vision. So AI is more than just a tool that you're implementing.