Type to search

Intelligent conversations? Smart technology, CX and the difference between digital butlers and stalkers

Technology & Data

Intelligent conversations? Smart technology, CX and the difference between digital butlers and stalkers


Marketing speaks with Dr Nicola Millard, head of customer insight and futures at BT Global Services about human-digital customer service, ‘intelligent’ tech, chatbots and satisfaction.

This article originally appeared in The Experience Issue, our February/March print edition of Marketing magazine.

With almost three decades’ experience at BT (British Telecommunications) Global Services, Millard’s roles have included stints as a human factors researcher, a customer contact strategy consultant and a customer contact futures project manager.

During a recent Melbourne visit, she spoke with Marketing about the insights found in BT’s ‘Digital Customer 2017’ research. The global study conducted over 10 countries sought to find trends in consumer preferences, expectations and fears, and looked at their satisfaction – or lack thereof – with a number of platforms including telephone, online chat and smart capabilities, including bots.

The conversation surrounded the use of personal data in customer experience and business strategy, and touched on why AI isn’t that intelligent, and the importance of building context into what it says.


Marketing: What were some of the main trends and preferences that emerged from your study?

Dr Nicola Millard: There are a few things that have improved, but there are also things that are tangibly not improving. Customers saying they’re exhausted from dealing with customer service issues, that’s going up. There’s more exhaustion going on. We’ve been looking at easy journeys and frictionless journeys for a while, so you’d think that should be changing, but it isn’t.

One of the big issues, I suspect, is that digital customers – when they go into digital – expect it to be easy, and therein lies the problem. Customers like self-service because it puts them in control. If it isn’t easy and they can’t find what they need and then they can’t find a phone number or any contact because you buried it 63 clicks down into your site? That’s problematic.

That’s been coming out as a problematic trend for a while. Also, there’s a kickback trend off the back of that where customers are serving themselves more. They’re only actually involving organisations for the complex and the emotive stuff. We’re seeing core channels emerge.

The phone: it’s not new or sexy, but it’s still coming out pretty high in terms of people’s preferences, and has done for a long time. It’s an accessible channel, and it deals with complexity and emotiveness well. But the real big growth channel is chats. We’re seeing massive growth in chat and, if you’re looking at digital customers, that makes sense. You’re online, you don’t have to find a phone number and a phone. You just click a button and have a chat.


How satisfied are customers with chat experiences?

It’s all to do with the design, isn’t it? Where are the chat boxes popping up? There are the ones that I call ‘impatient’ chat boxes. As soon as you get to the site, the chat box comes out, and it’s, ‘No! I’m just looking at the site!’ There’s that, and there is that perception that if they know the agents are on multiple chats, and there’s a delay, digital customers get impatient. They want instant access to things.

Chat is perceived as a much easier and faster channel than the phone. Realistically, I’m not sure that’s true. But it’s a perception – the customer feels in control.

They can go and do something else online, rather than sit and listen to hold music. So we’re seeing growth in that channel and, obviously, a link to that is perceptions around things like chatbots, a very hot topic at the moment. Two years ago they weren’t sure. Now, they’re going, ‘Well, if it’s quick, easy and gets me to my goal, brilliant. I’ll have a chatbot please.’

Obviously, don’t make it that they can’t talk to a human if things don’t work out. There’s that ‘IVR’ (interactive voice response) for digital. Like the ‘press one, press two’ that you would go through on a phone. They’re triaging you – seeing if they can get you to the goal – if not, dropping in a human agent. You can then see what they’re interested in. You can triage much more intelligently to the right agent with the right skillset. There’s lot of really interesting stuff around bots, they’re deploying them everywhere.

I think they’ve been slightly oversold, because I’ve yet to have a particularly profound conversation with one.

But it’s around appropriateness; have they got the data? They don’t magic data out of nothing. Is there an interactive FAQ? Can we build a process in the decision tree? Can we learn from agents as well? There are a couple of start-ups that we’re looking at – true machine learning ones – where the agent actually teaches the bot. It’s not cutting the human out. It’s actually enhancing them.


It’s almost like a staffing structure or training process?

I regard bots almost as children. You need to be a responsible parent, and you need to teach them well.

There are subtleties. For example, bots don’t handle complaints terribly well, because complaints are often very long. They’re complex. How do you pass that into something a machine can understand?

Sarcasm gets used quite a lot. I’d say Australia was just as bad as the UK on this one! An example I usually give is when one bot picked up something on Twitter about a UK train company, which was: ‘Thank you to this particular train company – which will remain nameless – for my free sauna this morning.’ Now, of course, you put that through a typical sentiment analysis engine, that’s going to come out as a positive comment, but we know, as human beings, having a sauna on a train is a bad thing.


It’s not what you want…

But why would a machine know that? So a lot is around context. Then, of course, we’re looking at things like emotion recognition.

There are lots of really scary start-ups that have started to use the cameras on smartphones to assess micro expressions on the face. In marketing there are people going, ‘Ooh, that could be interesting!’ But it’s also a little bit scary.

What’s acceptable, what’s creepy? Am I, as a customer, happy that you’re reading my facial expression or taking my personal data and doing something with it?

On the personal data side, we have this thing called the ‘me-conomy’. So, I’m actually more willing than I used to be to share social media or location data with somebody but, as a large corporate, I’m not going to do that just for free. What am I going get back? What’s in it for me? So I will share it, but I want more personalised service or better offers. If I don’t get any of that, why would I share it? And then if you use it too much, it starts to become creepy.

There’s a fine line between a digital butler and a stalker. The digital butler is helping me to do things I probably don’t want to do. A stalker is following me, tapping me on the shoulder and I don’t want them.


It will be interesting to see whether actually not using a consumer’s data for anything beyond fulfilling their purchase will become a successful model and attract loyalty from customers.

There’s an interesting test in Europe going on at the moment, because we’ve got the GDPR, the general data protection regulation. It is actually mandating companies. First: obtain consent from customers, but also explain to them what their data is going to be used for. I think it’s a good thing to be honest, because it’s starting to crystallise conversations around ‘what data do I need from customers?’ Consent is needed to use it for a bot, or for personalisation engine.

It’s going to be an interesting testbed if it works; I suspect a lot of other countries around the world are going to look at implementing something similar.


The human element in digital CX – how do you see that changing in the next five years? What will an effective human-digital CX solution look like?

There’s a lot of talk about AI wiping out jobs, and I don’t like the words ‘artificial intelligence’, because, frankly, there’s very little that’s intelligent about it.

It is dependent on the data.

Plus, machines are good at doing certain things, and we’re good at doing certain things. I prefer ‘augmented intelligence’ to artificial intelligence, because I think our human brain is amazing. There are certain things a six-year-old can do that a machine is going to find quite hard for a number of years. 

In the core CX space, a lot of those skills are the ones that add to the brand: empathy, caring, innovation, creativity. Negotiation is an interesting one and even just basic conversation.

Bots are not particularly good conversationalists, because human language is anarchic. It wasn’t designed to interact with a machine. The really interesting thing is we’re getting all these voice interfaces coming along: the Google Homes, the Alexas. You can’t have a profound conversation with them, but they’re an interesting interface, as the technology starts to disappear into the ether. The smartphone is now absolutely the portal for the customer. I don’t think that’s going to go away for a while. Smartphones will evolve. People are using PCs less, people are using tablets less. The capabilities of the smartphone we’re only really just tapping into.

For example, geolocation – can I figure out where you are, and can I contextualise things from that? You’ve got your digital butler sitting there. They’re a bit dumb at the moment, but they might evolve. I always say that, at some point, maybe my digital butler will be handling all my service issues for me. So it could be a case of my bots talking to the company’s bots.


That’s an interesting point. If we reach a stage where your bot is talking to the company bot, all the work that’s been put into accuracy, voice recognition and replicating human communications may be rendered obsolete – because bots don’t need to talk to each other like humans.

No. But again, I suspect that there are certain things that they can talk about that are easy, and then certain things that are difficult. If you’re looking at things like brand, one of the interesting things around digital is it’s actually quite difficult to really differentiate your brand from another in the digital space.

It then often comes down to how you handle that as you go from digital to human, whether that’s contact centre, branch or retail store. How do you connect that together? How do you decide where the employee bit comes in? How do you enable your employees to have the right information in order to actually deliver that brand promise?

To be honest, it’s still the people bit that builds the brand. So where do you invest in training and the enhancement of jobs? Well, it’s the customer-facing people that make that difference. We’re starting to see things like contact centres changing. They’re going from a cost centre to a key strategic hub, because they’re where all the information about customers comes in, where you can use analytics to start to really understand the customer. But, second, they make the difference to the customer. Their work is shifting from very transactional, very predictable to quite emotive, quite complex.

For banks, for example – where they’ve got a high penetration of apps – people are doing a lot online. People don’t call their bank unless there’s a real problem and they need advice or reassurance. Or they’ve been mugged and their card’s gone.

When people are in a crisis it’s an interesting one.

You can’t put complex technology in front of somebody who’s just been mugged. The chemistry of the brain has changed. We lose half our short-term memory capacity if we get angry, frustrated or anxious. So if you’re going to put someone through a complex IVR, for example: ‘press one for this, press two for this or press three for this…’ Our short-term memory is usually between seven and nine bits, so when we’re calm, ‘press three’ is about the limit. If anxious, ‘press one for this, press two for…’ we can’t remember what one is for.

Customers in crisis need to be accelerated to something that solves their problem simply. There’s actually an online bank, a start-up that just has this very big button on the app that says ‘cancel my card’.


That’s smart design.

It’s just simple. ‘I’m anxious. Let’s just do this. I’ve just had my cards stolen. Let’s stop it, so I don’t have to go through authentication, because I won’t remember my password.’

Unless we use biometrics, which of course is coming in quite a bit around ‘let’s make it secure and easy’. It’s on the personal device again, it’s smartphone, it’s got the fingerprint, it’s got the camera I can take a selfie on. That’s really becoming a reality. Biometrics – that’s quite an old innovation.

Even voice biometrics has been around for a very long time. I think it’s now coming of age simply because the cost has gone down because we’ve got the devices.


Talking about this fluent integration of tech on the human, customer-facing side, do brands recognise the importance, or are they too obsessed with that next big technological step?

Some are coming around. People are often under-investing in their phone channel, because the phone isn’t sexy. A lot of the investment has got into chat. Understandable, because it’s a huge growth channel. Social media’s got a lot of investment as well. The phone still works pretty well for quite a lot of things.

I think a number of companies now are getting completely bewitched by the promise of AI and chatbots, and things like that. They do work. I’m not saying they don’t. Customers are actually saying, ‘If you can get me to my answer, brilliant.’ But don’t then underinvest in your frontlines, because actually they are going to be adding quite a lot of difference.

The skill sets are starting to change in them as well. So it’s not a skill around processing lots of calls and the really fast reading of a script, it’s actually having to be creative. I always describe the role as ‘sticking the fork into the spaghetti of backend process on behalf of the customer’. How do we make sure they have the skills? Communication skills, problem-solving skills? Caring, empathy, all of those things we undervalue often now.

They’re skills we need in the future. Some companies are realising that.

In the industry, there’s a growing awareness that the contact centre is becoming a little bit more strategic than it used to be. Because of that, it’s getting more investment than it used to get. I think that’s a good thing, but there’s still quite a lot of people who may be lured by the attraction of the technology without thinking of the consequences.

* * * * *

To purchase a copy of our latest issue or a subscription to the mag, visit the online shop »

* * * * *



Image copyright: sdecoret / 123RF Stock Photo


Ben Ice

Ben Ice was MarketingMag editor from August 2017 - February 2020

  • 1

You Might also Like

Leave a Comment