Using data to create customer trust

The full interview with Jimmy Hosang from TMAC (The Modular Analytics Company).

Here Jimmy talks about the importance of voice communication, in building customer trust, together with some of the analytical approaches that can be overlayed to create even better outcomes. Voice firmed has an important place in the future for contact centres.

Find out more about TMAC-> Here.

Interview Transcript

0:00
So hi, everyone. I’m here with Jimmy Hosang saying he’s the CEO of modern analytics company today. Jimmy, thanks very much for joining me. It’s great to have you here.

0:09
Thanks for having me.

0:10
Maybe just take scribe a little bit about, I suppose, what you guys do and what you’ve been up to over the sort of the last the last 12 months or so?

0:17
Yes, so, we’d like to describe ourselves as a proudly different AI company that specialises in conversational intelligence. And what we really want to do is we want to maximise the value of every customer interaction. So specifically, contact centre environments. A little bit of a potted history of myself was Adam, from a credit risk background, I did marketing and pricing that was a bit like the littlest hobo. And then finally, I went contracting, and my first contracting gig was in a contact centre. And I’d never properly worked in contact centres, even analytics environment before, then all of a sudden, by just just lit a fire under under there because you know, the risk profiles, your marketing campaigns, like absence, sickness, holidays, it was such a diverse ecosystem that I just thought, this is where a panda wants to be. And this is, this is where I’ve kind of been working over the last guy, this fat six or seven years. And then we started the site, the module on six companies and outcome based consultancy, to drive value in contact centres. But over the last year, we pivoted to more of a SASS company. As we proved out some of the fundamental, fundamental products, which were our next best action, it was about speech analytics, it was about targeted coaching and understanding to productize those, and we started to be released in earnest over the past 12 months.

1:47
And what’s your feeling about it? So as from the contact centre environment, the amount of data that there is available in the contact centre is terrific mean? Is that been a big gap in terms of making most of it? And then how do we get access to more that data? Because it feels sometimes like there’s a lot of data that’s just left on the table, and we don’t use it?

2:05
Well, people, people don’t care about it in businesses, like it’s a huge, it’s a huge annoyance for them. Like, I feel like companies often see sweets, in particular, they get drawn towards this big shiny thing. And the big shiny thing over the last few years, you know, first of all, it was kind of big data. So it was like, say, like, Where’d you get all of these digital touch points? And then that kind of led into, you know, the digitization? And, but nobody thought about, well, what about the humans that you’ve got interacting with the human human customers? And how do you collect all of the information of them surely like that? That is the starting point. What people tried to do is they tried to cut into the saw the contact centre and the diverse ecosystem and all of the potential data and all of the stuff that they’ve done wrong over the last 20 years in collecting data, and they thought to themselves, Oh, well booger that we’re not going to fit that’s too big. That’s to let someone just leapfrog that will digitise it? Well, that’s not that’s not how human beings work. And they tried to like piggyback over the digital piggy over the problems in the contact centre, tried to go straight to digitization, automation, chatbots AI, that kind of technology. But if you actually look at life, you know, I think 30 years ago, the contact centres were going to shrink and actually grown and have grown because because you’ve not fixed the fundamental problems and how you create a great human to human interaction. Because it’s only when you do that you start to want to automate and digitise. And by the way, I

3:54
suppose that’s the sunlight, the speech analytics is becoming that’s become quite a big theme over the last, even the last six months, I’d say it’s sort of been accelerating. Obviously, you guys are, you know, the heart of that as well. I mean, is that is that around sort of optimising that human to human interaction, because again, in speech, there’s huge amounts of data that sort of untapped and almost like As humans, we process it. And as soon as it’s trying to get into some sort of machine readable format, you can actually use it to help out I suppose,

4:21
when I worked in marketing, I worked in marketing for loans and credit cards. And what I always what I always said was digital marketing ZZ, because I build a predictive model, and I deploy it onto a website and the website generally doesn’t have a bad day. It does exactly what it’s told. It’s much more interesting to build a predictive model, and then give a script to an agent and get them to get them to try and try and deliver it in the most optimal way. Because, you know, humans, humans are bad days. Humans have good days human, the way in which it’s delivered is very, very important. And so then So coming from that marketing kind of credit risk background and then go in both. I have deployed models, but my models aren’t really performing as much as they would. So, you know, I thought to pick an example, but I thought a model would deliver either 1.5 million pounds a year, like that deployed into a contact centre for for upselling and cross selling, and it’s delivered, you know, a few 100 grand why? So, it was the why, why is it not delivering? That kind of made me to using speech analytics to say, you know, what was the incoming sentiment of the customer? What was the reaction? Can the sentiment of the agent like did the actual they say the things that they wanted to say? No, not really, like culminated in how do you how do you adhere to the processes that I created when it was about my campaign so that so the use of speech analytics for me and I think a lot of companies over the last kind of decade I don’t even Missoni speech analytics. And it’s become it was expensive, like, transcribing audio historically, had been very, very expensive. And, and therefore, it became a speech analytics project became like a vacuum for all the all of the benefits like you had to do everything in cannibalise. Yeah, everything, I have to do everything. Whereas to me, speech analytics is just an input to better customer experiences, and better and better agent coaching. It’s another data set, it’s a bit, it’s another database, it’s kind of a big data set. And that’s what I really want us to, that’s what I really focus on when it comes to speech.

6:52
But what you’re saying is kind of interesting is if you always like take a customer journey design, right? So it’s almost like you have all these pieces and the logical processes that need to happen. But you’ve got to, you’ve got to include the the human interaction, execution elements of that, and you got to have measures to measure how it’s getting executed, or they find the process and receives otherwise you don’t get the benefits, right. So like the speech analytics piece sort of overlays around that quality measure and the execution measures if you miss that, and that’s where you lose benefits.

7:20
100%. And I always say like, so when it comes to customer experience, you’ve we’ve all been in those workshops with those brown paper exercises where we go, Well, this is the happy path, and this is this unhappy path, this is what happens. And then we go on the customer calls, and then it gets resolved or it doesn’t get resolved. Well actually, what you have to do is have the customer call, you have to extrapolate that and look at them six minutes, 12 minutes or 45 minutes that that customer go and look at all the journeys that happened within within that conversation. Because that conversation that human to human element, by a lot of got to go back again, like a lot of companies are trying to digitise their experience. But I think by digitising their experience, they’re actually removing some of the thing that makes them valuable, you know, the, the commoditizing experiences and and the commodities Well, if I’m if I’m just doing everything through a website, if another website, does it exactly the same, then why should I be loyal? Whereas actually, you know, just historically, and the, you know, the brands that really, really understand their customers, they understand the value of human to human, that’s why Apple has got these Apple servers, right? You go and have a human to human interaction. And so I think when we design experiences, it’s looking at all of the touch points I’m looking at, extrapolate in the conversational touch points, looking at what is happy path within that goal. And what is the unhappy path within that column, redesigning them so that the agents get better outcomes?

See also  Advancing Digital Payments

8:59
I suppose one of the good things about talking to human is an NF it’s done right is you can get to the point pretty quickly. So I can see taking in face to face, I can see the way you’re reacting, I can see here what you’re sort of intimating on the phone, or the way you’re talking to them. And I get to the point pretty quickly, and you saw like the shortcut that that getting the information quickly. When done, right, I mean, done wrong. And if you’re following a process, then you get a long process. And that becomes frustrating. But it’s almost like how do you how do you use the humans to get to the point quickly and cut down time for the right customers? I suppose the same is true of I’d have arguably of digital processes and use the data to get to the point quickly.

9:37
In almost any scenario. It’s easier to do virus than it is to type. Yeah. So that’s why that’s why I find like chatbot technology and the the endless spin that you get into when it comes to chatbots are like I don’t know how you call it like artificial intelligence because it’s nothing new. Telligent about most of these chatbots so artificial intelligence, like you get caught in a loop of trying to get something done now bots we use our volunteers like much more than we did before through the use of smart smart tools like and smart homes. We have Alexa set in a Google Home using more modern more because we understand it, you know, that is the quickest way to do something. It’s quicker than typing. And so, where I think we we became misguided around customer service a long time ago is we actually kind of real customer service made it look at humans human customer service, we made it look rubbish, and then gave ourselves like reasons to digitise you don’t like, one wait times in a contact centre, like not arming your, your agents with the correct scripts and the correct tools to do the job properly. And then when it’s rubbish, when all everybody wants to do to digitise nobody really wants to digitise me. Digital advice doesn’t mean doesn’t need anything, people just want to get the stuff done as quickly as possible.

11:11
And it’s a bit like um, so I we had some telephone issues this morning. And we’re saying about me the story, right? So I can telephone issues. And I phoned up and I tried to get it resolved over the phone. And I was going round and round and round in circles, right? Eventually, I got an engineer to come out and the engineer came out. And within 20 minutes, the whole thing had been fixed. He knew exactly what to do. And it’s just around having specialist knowledge deployed at the right time, sort of there and they know what to do. And they can just fix it straight away. Because they’re experts, right? Yeah, well,

11:40
I was on I just made a comment earlier, like I’m I’m moving home and I’m moving on with my services and things like that to my new new house. And I was on, I was only there with a chatbot with chat started, I was moving my broadband, I started at half past eight in the morning, and by three o’clock in the afternoon, that’s still not resolved it. So that kept moving on to sort of the people, obviously, I’m working, so then I have to jump on to something else, then like I get passed on to somebody else, then it goes on, like, all I was doing was checking that the mass movement data from abroad button was correct, because I got an incorrect text message. And it took until hours and I was unable to get that done. Whereas I was moving out my TV services, I put in some details and they said, Can you please read up, I rang up and within 20 minutes it had been resolved by by a contact centre agent. So we don’t we don’t monitor the amount of time and benefits you don’t want to do all the time that’s wasted by the consumer. Like having to go through this process. You just see the operational efficiencies on the on the on the company side. Yeah,

12:54
so it’s interesting, I tend to do a bit the deep dive around that type of speech piece, and then how to use the data to speech analytics to then make it more efficient, or make you get the outcomes even quicker. That’s, that’s interesting.

13:04
But what so what we actually do is we so we can we can maybe touch on this later. So I believe in scripts, so and you know, six or seven years ago, everybody kind of completely got turned off against scripts, and they said, No, you need to be an active listener, you need to be empathetic, you need to do this, you need to do that. But I consider performing arts background, like all of the best performances that you’ve ever seen in any any film as natural as be look, come from a script and eat so your best actors in the world, the Americans to see Robert De Niro’s, like you don’t say to them, I’ll just make it up as you go along. Just just actively listen, like the scripted and the old learn how to perform it and work from the craft. And we’ve taken and those are the very, very best, what we do is we take out 1.2 million agents in in in the UK at the moment, and that centre agents, and we kind of just only said at that just react, I’m just not sure what that look like. So we I hated him throughout. But I think I was relatively relatively good at it. But just yeah, just just not just read just freewheel. And like, can you believe that we did we actually say that I have had the arguments with people I was like, no, what we think there is we should design the best experiences from what we understand. We should use the best terminology. Because, you know, this certain terminology that you know, in and puts people off we should remove friction. And then we should we should use each analytics to look at when which parts of the journey sentiment drops off. Was that the right thing to say? Or just generally like is it agent like sticking to the script or are they friendly and if it’s really really been getting good outcomes tech, maybe we could learn something from it, if the free really getting bad outcomes, but start a coaching conversation and this is why you should be sticking to the sticking to these scripts and sticking to these routes, because it gets you to better outcomes, both for yourself and for the customer. Yeah,

15:08
let’s just go back to the activities. I’m not performing art background, but it’s almost like, you test it, you learn. And then you do it again, and you test it and you learn you do it again. And it might be slightly, every performance might be slightly different, but you’re sort of constantly reinforcing and getting better, right? So you have to be you have to be able to control it minds of science back and you have to control it, to be able to learn and then get better.

15:26
100%. And it’s also like your motivation. So yeah, so how I would put it is, like, every performance is, every performance, every phone call you have is slightly different. But you understand like the parameters work, too. But before, before we even get to that stage, you need the training wheels underneath, and you need to go through there. And you know, we talk about how long it takes to train. I think Malcolm Gladwell, when was your 10,000 10,000 hours. But but I’m not strange choice, necessarily. That’s the case. But you should be should be putting some training wheels on people and going but this is what good looks like until you get and then just get it move. Deliberate practice, like in the groove of going through this. You know, you go well, we don’t want them to sound like robots. You don’t want them to sound like robots. But you don’t want to get into problems with the FCA. You don’t want complaints. You don’t want them to sell cyber stuff. So we should be able to get training wheels. And then when you know when they start to get more confident that will just naturally they’ll be able to be able to free reel a little bit.

16:30
Yeah, so it’s the data, isn’t it and the data with the human interact with the human into the life, how do you then get better using data to learn and then get in gradually get better, I mean, for all of the all of the all of the stuff we’re talking about?

See also  The data explosion... changing how we see and measure the world

16:41
Exactly, exactly. And as well, what I find very, very Museum is remove that remove the contact centre agents, but to build the chat bots, then you have to have a script up to go. If they say they say this, if they say they say this, if they say they say this, and the best chatbots are the most complicated because it’s got all of the different decision trees in it. And then it makes decisions as you click through and actually responded to. So we’ve given chatbox a better, better guidelines and a better start up than we do our contact centre agents.

17:15
So I was looking through some of some of the stuff that you guys do and Naja come up quite a bit in terms of like, who’s going to chat bots. But I suppose that’s also in terms of guidance or scripts for an agent as well as like, how do you nudge the customer conversation longer to get to better outcomes? How do you think about that? How do you think about implementing nudge, because it can be quite a, you have to use quite a lot of data to really understand what are the dynamics of how customers might react in that situation, those kind of things,

17:41
you’ve got an interesting collection, I, I’ve got a real interest in behavioural science when it comes to certain collections of the collections industry. And I’d kind of be interested in your thoughts as well, because I believe there’s, I worked in collections and credit risk, sure, in the 2008 2009 financial crisis. So really, really interesting times, and but what I found was we, we actually, we actually did a lot of disrepair to like, our, our customers, because we started to treat them, we started to treat them like pretty, pretty horribly. And like my belief, and I’ve always had this opinion that, you know, there are certain there are certain types of customers who, who can’t pay, or won’t peg, the certain types of people who psychologically are like very responsible, you kind of got like a kind of got to make a matrix. In fact, many, many times in credit issues, start do these matrices. And if these people in the middle that are the big, the most valuable real, and especially to any kind of you know, credit risk, our our, our credit, kind of company, because they they’re always hovering between a hovered between, you know, you don’t want to so good that they just pay us off, straight away. Even the people at the top end here. Even if we are like, These people don’t really need credit, it’s probably like, it’s probably they just got they just got it, it’s always these people. So you’re always like walking a bit of a fine line to get these people to behave in a certain way get them to pay on time and stuff like that. And so that to that end, like is about behavioural principles is like getting them to not be as impulsive, making sure that they’re the understand when the directive is putting out and making sure like you understand the impacts if you don’t pay things like that. And therefore, like, you know, we can build the best kind of dialogue models, collections, models and collections principles, but there must be some some some if not all of it, which is around training people to be better at managing credit.

20:00
Yeah, yeah, I think it’s I think it’s also like, makes me think of almost like politics, right, which is elections are decided on the margin flight for people who are the swing voters, right. And it’s almost like that in terms of like collections in terms you got the people up the air, if you can pay the people down who really can’t pay, but it’s the the difference from a company point of view is trying to persuade people on the margin, right? So it’s like, how do you get the people who, who maybe they need a bit of a nudge to be able to pay and correct their credit? And it’s like, how do you get further down that that line, rather than as much as the people are either end, right? You making me think is that really, can you use nudge to to move those people up and make quicker payments? Because it’s in their interest as well? What do you feel about transparency within that, within that sort of nudge cycle? I mean, a part of it, I think, is, is also how you transparent with what are the consequences of doing something? What are the consequences of your actions in terms of like joint responsibility? I mean, do you think transparency is a thing that there’s this that’s gonna come across more?

20:59
Transparency is interesting. I’ve not, I’ve not thought so much about transparency, just in and of itself, what we talked about is trust. So we have to build trust through our processes. And I think transparency is probably one of the things amongst others, that ladders up into like a into trust. And I see a lot of the lot of people talk about that Trustpilot scores and things like that now, not 100% sure whether that is true trust. It’s how, how, through your conversations through your messaging, through the way in which you behave, like how do we build? How do we build trust with with the customers to the point where, like, especially in financial services, and that when they’re in a crisis, or when they need to support a competition, rather than thinking that you’re an institution that always rips them off, are discharged, necessarily on just our practices in ways that isn’t ethical. And a lot of the stuff that we’re doing at the moment isn’t just around, is it just around like, you know, the data scientists aren’t just legit or necessarily, it’s found, how we identify customers who are distrustful of your business? And how through communications that we start to build more trust. More, obviously, that choice these two way like, you have to be like, yeah. And so yeah, I would say that we’re not I’m not doing a lot of thinking around transparency. But we’ve been doing a lot of things about trust trust.

22:38
And how does it lead to trust in the in the day? And how do you measure trust them at the end of the day, so? And how do you measure these getting better? So I like having data that you can actually show, we were here, and now we’re here. And these are the drivers of how you get from A to B? And it’s like, how do you measure trust? And how do you actually measure that it’s got better from a customer point of

22:56
view? Yeah, so we use proxies for trust, the Trustpilot score and stuff, it’s not important, but I wouldn’t say like whether or not it’s a, it’s a trust measure, or whether or not it’s a or it’s, it’s something psychological something to do with your personality. But if you if you dropped into if we if we built a model, which was likelihood likelihood to go one payment in arrears, you would say that the people who were like somewhat high likelihood to go but one payment in arrears and then and then a model, which was upon going on payment areas like the hunter contact. If you looked at the likelihood to move into contact model, what you’ll find is the people who are low, right, limited contact and highlight that to go and those are the those are the populations that are most distrustful of you because we think just trying to you now going to try and like some of them are, rip them off. You’re gonna go out and aggressively by the by the glass. And then what you’re trying to do is try and look at those look at the communications that have gone on with that customer over time. And see like why would be if you don’t want to speak to you but listen to some of that you’re never going to be able to change some people just put their head in the sand when weather financial difficulties. But you as a as a institution, as a company, like your responsibility is to try and look at those people and try and get them try and get them to contact and I will say that just treating people overtime, treating people like grown ups. And I think it’s a problem with you know, it’s a problem with a lot of communications from Twitter, Facebook Pillay, everything that’s happened in COVID. I think all the time like treating people like grown ups. He said that Trucking is an attractive quality.

24:53
Trust is such an interesting concept. I love it because it is something you can relate to New York’s like as a human. You can later and it takes time to build up trust, doesn’t it? You’ve got to do the stuff. Trust is about what you did six months ago is what about what you do now. And it’s very hard to change, change it now you’ve got to think looked back over six months and say, what did it do over there because that goes into, you know, the way people process around whether they trust you or not.

See also  Change Regulatory outlook: What is the outlook for 2021

25:16
Yeah, and I think it goes back to it goes, you can look at it from a customer experience perspective, you can also look at it from an employee experience perspective as well. So like, there’s a tremendous amount of distrust around smart analytics from employees, because the liable interest always monitoring me. When we do work, hopefully, by we put people in the ground, I sat in contact centres, we’d be agents around me watching what I do, and going, you know, this isn’t make your life harder. This is, number one, it’s to make your life easier. Number two, not make you feel like trying to get the best trying to get the best out of you can put that on increase sales. And I think the third thing to touch upon is human to human. And augmenting service is the best, I don’t know what anybody says. But it’s the best that’s knowing that needs to a cube and helping the agent be armed with all the tools with the best scripts with the recommendation engine like and Coach correctly, like that. That is the best experience. And at what I kind of say to some of these guys, if I’m being like super, super brutal, we want to give you all of the tools because like, most people just want to alternate. And that’s, that’s a much worse place to be because then like, you know, people’s people’s roles and stuff disappear. Whereas if you maximise the value of the human to human interaction, like that’s gonna take years and years,

26:49
humans human interaction is like, incredibly powerful. And he talks a bit about it from a customer, state customer to company human to human interaction, I suppose that’s also true in terms of like, employee to employee interaction, or, you know, employee to manager interaction human to human is kind of like is, is kind of one of the best ways of sort of getting to know people. How much do you think we’ve lost with, with I suppose, the pandemic and remote working? And where do you think will go as a result of that in terms of like, coaching tools as well? I mean, what how does that kind of link together in terms of some of the customer observations,

27:25
I always go like, 18 to 18 months ago, can’t remember, all the ones went into one. But you know, 1616 months ago, we gave it guys a laptop and a headset, and like sent them off to set them off. Oh, yeah, off you go. Do do that job that you did in the office, but go and do it like your house. And we did so like, did silent necessity, but we didn’t do a lot to we didn’t do a lot to to help them to give them the tools necessary to to be able to perform in those types of environments. Now, some of us are very well cared by, you know, some of us, you know, got house and garden and can go and take a break and stuff. A lot of people like work in that situation and didn’t realise that we’re locked, locked in a room and locked in a room, or sometimes, you know, some of my friends who worked on the phones. There’s three of them that live in the house. Yeah. Yeah, exactly. Just listed on the screen. So it was the service in lots of different ways. It’s linked, select contact centres in particular, for an element of robustness there, but just I’m just digress a second, but you know, you can see customer service in restaurants and things like that just isn’t as good as it was, like two years ago. And that’s because like the guy is like, guys in those types of sectors are like, Well, why would I put myself in a position to be locked down and not have a job again, like I need to go and find something else. So general, this is kind of this kind of got difficult. I come from a data visualisation background. And now with visualisation, I love making things like spin on the page fritter. But I’ve started to over the last kind of 20 months think, well, is that really the best way of displaying information? Because you when you take data and display it in like a visualisation bar chart or pie chart or line graph or whatever, you then add a barrier, which is an a person’s ability to understand that and it’s that it’s that kind of, like kind of semantic layer. We were doing a lot of coaching stuff and say look this is these are top performers, easy button performers, but in visuals and it’s nobody was adopting it or a very small percentage we’re adopting it through our tools now we’ve kind of changing that can he knows this person needs coaching on apps because Why in plain English? Yeah, go and publish this book and go and schedule a coaching conversation, this person read well across the board, give them a pat on the back and send them an email. Yeah, that type of direct information like specifically prescriptively saying what to manage, you need to be worried ops manager needs to do. I think that will become more and more prevalent over the next kind of now, forever, because why? Why put visualisation in when he can just directly tell people what to do. It removes it removes all nuance.

30:36
So one thing I wanted to ask you before we finish was a little bit about so I’ve been following you on LinkedIn and social media and those kinds of things. And you say, you certainly have a lot of fun in your posts. And so the question I want to ask you, do you think we’ve all become a little bit too, too dull and corporate in some of our insular interactions? What’s your approach? Or what’s your approach on on that?

31:01
So I think I’ll go up to kind of like who I am as a person. Yeah. So I just like I like to have a joke. And I like to have a laugh. And I like to promote conversation beings just didn’t like my, my non digital life. And that would have been like bad for, you know, sees a dot. Yeah. However, it’s like not, it’s not that, and so on, on LinkedIn. That’s I haven’t been shooting day by eight 910. Like, for me probably like 1213 hours a day at work. So you always just talking about work and talking about work stuff. And data science, it can be like completely, like boring talking about, like, specific algorithms or specific tech of this. Like, that’s not, that’s not really what my life has ever been. Most of it’s about, like, they can take you friends with people and joke with people and a lot of people, even the most passionate people, people are most passionate and competent at the job. And the truth of the day spend, like, what’s the word, like, trying to avoid doing their job. And take it a little bit of a little bit of a break from there, and smoothing things over. So I. So I kind of took that into my LinkedIn, and just started posting, you know, things that I thought would make people laugh, things I thought be funny. I honestly believe that LinkedIn is the best place for that at the mall. And Twitter is terrible. Like, Twitter is terrible for conversations and terrible abuse today. And Facebook isn’t too far too far behind the kind of habit and a closed community on Facebook anyway, what I’ve found, because of the professionalisation of LinkedIn where most people would their professional capacity, you can have great conversations around work. We could also have decent conversations about just like your mic. And then you can also you can also have relatively civilised and relatively civilised and passionate debates about about other subjects that don’t often spill over into like, just name calling. And that and therefore like that’s why that’s why I think like, baby hasn’t become too serious. Lots of people post, posted it and going to the post that it’s like, come on, you can just ignore it, or you can just walk.

33:46
Yeah, Timmy. Thanks very much for for making the time and chatting with me. It’s fascinating. We got a lot of topics there and very, very interesting. So I really appreciate it. And thanks very much. Thanks for having me. Yeah, let’s do it again. You’re quite welcome.


RO-AR newsletter

Receive notifications of new content notifications: Subscribe here - unsubscribe anytime