conversational AI

DRUID Talks Ep #7 AI for Everyone - Making Technology Work for Both People and Businesses with Laetitia Cailleteau

DRUID Talks Podcast Ep. 7 explores the intricacies of conversational AI and touches on examples of how this technology is used for the good of the people.

Episode #7 of the DRUID Talks webcast features Laetitia Cailleteau, Managing Director for Accenture and Data&AI Lead in Europe, and Subject Matter Expert Kieran Gilmurray. See the full episode and transcript below.

Subscribe to the next episode!

Kieran Gilmurray:  Welcome, folks, to the latest DRUID Talks. I'd like to introduce you to Laetitia, Managing Director for Accenture. Laetitia leads Accenture’s Data & AI business in Europe and is the Global Lead for conversational AI. She drives innovation, sales and delivery across the globe, across multiple industries and clients. Laetitia has more than 24 years of experience in the consulting industry, has authored multiple academic publications and holds patents in the conversational AI domain.

Laetitia was the founder of Accenture's Innovation Studio in London, the space where her team co-created with clients to realize the potential of artificial intelligence, augmented reality and IoT. Laetitia is the reserve member of the Artificial High-Level Group at the European Commission and also a board member of OASIS, the Open Foundation in Europe. Welcome, Laetitia.

Laetitia Cailleteau: Thank you, Kieran. Absolutely delighted to be here with you today! 

Kieran Gilmurray: Well, let me jump in and ask lots of questions because I've been waiting for this moment for some time. There's been a lot said about chatbots and chatbot technology, intelligent virtual assistants, digital workers, and cobots. They're called a lot of different things that can lead to confusion among these technologies. But what do you think is the difference or what differentiates conversational AI from other chatbot technologies?

Laetitia Cailleteau: Well, thanks for the question. I think that's a very important question, especially for us in Europe. When I started in the conversational AI space, we didn't want to be associated with the word ‘chatbot’. And the reason why we didn't want that is that in Europe, we were trailblazing a lot of those technologies in early 2000, and they were not particularly successful. They were quite, you know, the technology cloud didn't exist, AI wasn't there. So, they were very scripted and very rigid, and most of them were also limited to text. And we didn't want to carry that baggage of an environment which is quite constrained, which is quite rigid. So we thought it was important to actually acknowledge the progress of cloud, AI, technology, and natural language processing and really kind of embrace something which is much broader.

So this is why actually we've been in Accenture, one of the first to call this area and this domain conversational AI, really to mark a new step change. 

Kieran Gilmurray: I love that you pioneered that, by the way, as well, because it's very rare that people see a technology right from the very beginning. And then bring it all the way through. Your experience must be exceptional on it! Do you believe conversational AI is a special technology amongst emerging tech? 

Laetitia Cailleteau: Well, I think what is very special is the possibilities it brings just to… it does really change the relationship between human and machine. So let me take you a few years back, you know, like we used to have to learn how to code and how to interact with a particular machine and that used to take a long time. I think now, with conversational AI, where we do is in natural language, we can start interacting with systems, and that changed the complete paradigm of the relationship we have with machines because we don't need to learn it any longer. It learns us. And I think the possibility around this is actually mind-blowing around democratization, access and all of that.

So yes, it's a very exciting technology because it brings so much more possibility for all of us, professionally but society, more generally. 

Kieran Gilmurray: You must have seen tremendous change since you first pioneered that technology. What're the biggest differences between now and ten years ago? Is it the lack of rigidity? Is it that technology is, as you mentioned, learning us as opposed to it or us learning it?

Laetitia Cailleteau: Well, I think it's a number of dimensions. I think, obviously, starting to have the cloud has been a bit of a game changer for this particular technology. And with the cloud, the ability to store data and utterances and then starting to use AI in some of those systems that were very scripted before. So, you know, I think the paradigm really kind of evolved with those kinds of technologies, kind of being available and being more, you know, accessible.

Now, obviously, with a large language model and some of the generative AI, you know, new technology coming up, you know, we are taking another step change into this particular technology and the world of possibility that's becoming even stronger. 

Kieran Gilmurray: I love that phrase, the world of possibility. That's so cool! From your experience, then we… the business world has a great number of pains, and this technology can potentially help resolve those pains. What have you seen the best businesses doing? How are they using conversational AI to make their businesses much better than before? 

Laetitia Cailleteau: Yeah, so I think I think a lot of people - that is too particular angle people have been looking at. Ultimately conversational AI is a new system of engagement. So we used to have websites in early 2000, then we had apps, and then I think conversational AI is the next kind of level of how we engage with a brand, you know.

So we had a rush for brands to kind of have a website to have an app, to have X, Y, Z, and now they want a system of engagement. I think people wanted a system of engagement to improve customer experience. And what has been really difficult in the market is that a new system of engagement is typically, you know, more expensive to run if you just do it for the customer experience. And I think that that's where the industry wrestled a little bit. So there were some, you know, great amount of keenness to become, you know, multichannel and all of that kind of things. But at the same time, people were conscious, you know, this was an additional investment. So the key for us was to find the right use cases or business cases to actually, you know, help transform and help found that new way of engaging.

I think one of the use cases that's most deployed out there, and I think you know – there’re not many companies that haven't considered it - is the calls in the customer care element because, ultimately, when you start digitalizing your customer care using conversational AI technology, you can start, you know, have besides the cost saving and solving some of the challenges of having to ramp up agent all the time, people that don't really like the job and having something quite seasonal as well in your business, you know, like a lot of brands have like a big rush at Christmas and need to hire a lot, and it's quite difficult to ramp up all of those things. So the call centre in the customer care… the technology really solves some immediate challenges and really helps, you know, running operations in a more streamlined way.

But what people also realize is by going… using that technology, they can also digitalize the voice of the customer and things that were usually in tapes buried under, you know, in the background, they could start analyzing what it meant and what customer was really after, what were the challenges and all of that, and really use that to better up their service and, you know, potentially review their product and their services as a company.

That's where I talk about, you know, the back office. You know, customer care used to be a back office kind of function. It's really kind of putting it back in the front as the voice of the customer that actually really has a strong touch, which is in strong touch with what's needed and what's not working and what needs to be done. And can really start impacting the backlog of particular brands on their product.

Kieran Gilmurray: What do you think businesses, Laetitia, want from automation? What do you think they want from conversational AI? Like do they actually understand the art of all of the possible that you mentioned a moment ago?

Laetitia Cailleteau: I don't think... it's quite difficult to make a generalization there, but I think, obviously, brands are quite complicated. There’s the marketing department, there’s the customer care department, there’s the technology department, and I think it depends on who is driving this particular agenda.

I think in the past, it was more the technology department or the IT department trying to size and understand the technology and the possibilities, and therefore it was maybe a little tactical, you know, in its implementation. Obviously, the power of conversational AI is much more than automating conversation or contact. It's much more; it's much more about reinvention of your brand and its customer relationship.

Just to give you a few examples, we had a number of customers that were very keen on automation and, but you know, when you start looking at automation of existing script, you know, those scripts, they often are very system-focused. They were created a long time ago for a company that wants to go through a particular process. But now we are much more savvy than that as end-users; we want to speak in our own terms, our own language; we don't necessarily want to be, you know, speaking a system language. Yeah? So things like, you know, when you call a particular brand, and they ask you for your order idea, which is that long are they asking you for your customer rather than, you know, things that you have at hand or things that can be deduced because you know, then they know about you and they may need a snippet of information, and they should be able to pull it up quite a lot.

So, all of those things are not in the old-fashioned way. And if you automate them with a conversational layout, you kind of end up with a very clunky system that nobody wants to use because it's the same as calling somebody. It's probably even less flexible because it’s still embodied in the old ways of thinking or establishing that relationship.

The opportunity is huge, though, because, you know, you can take the opportunity to reinvent this whole experience and streamline the experience and then obviously, you know, do some automation and use the technology to even make it smoother. I think if you miss it if brands miss the opportunity, they typically just create something that's not used, that won't get any return. And then it's not helping the industry because, you know, it's another chatbot that's not working and that's not nice, or that's another system... So the transformation element is super important, and it needs to be an opportunity for reinvention and really putting the focus on the human and the customer rather than on the system. Obviously, systems are still there. You still need to collect information, but maybe some of it you can infer things like that. You can make it much smoother for the end user. 

Kieran Gilmurray: Yeah, I wonder if anybody ever did a comedy show and they acted as some of the old chatbots and then realized what they were putting in place and then put that in front of an executive team or whatever. I do sometimes worry that the team where, you know, throwing a tool and said, get on with digital transformation. It's never truly worked. And I love what you said a moment ago around the… look, it's actually the back office that’s coming to the front—the customer insights and bubbling the away to surface of the organization. And then you're redesigning what you've got to create great experiences. That's probably the power of the technology and the design or the use of the tech.

You mentioned a couple of things there, Laetitia… you mentioned, you know, back office and front office and brand and customer service and customer experience and bringing together lots of applications. What are the crucial or critical capabilities that a conversational AI platform should have to allow it to help the brand and the business to actually meet and possibly exceed their business expectations?

Laetitia Cailleteau: Well, that's a very good question. I think one of the key ingredients is user experience design because even with the worst technology out there, you can probably do quite a lot if you have an amazing design team. Yeah, because ultimately, it's not about the quality of your natural language processor as much; it's about the quality of that interaction.

And very often, you know, you need to second guess the customer, and he would just nod, “Yes, you got me or not”, you know, like so… actually crafting that user experience is the skill, you know, like we used to have a designer for the app or for the website. You know, crafting conversations is actually more in the realm of, you know, movie scripting or things like that… You know, how do you craft something that's compelling, interesting, that's fast for the end user or other situations, and things like that? And then ultimately, if you have done that well, well, the end user shouldn't have to tell you their whole life story. So, you know, you should be very much in context and, you know, drive the conversation as a conversational system rather than just opening up the world to, you know, the world on his brother.

So I think the design is key. And I think that's what a lot of companies don't get right. And it's not easy because, you know, the designer… there is still a left brain and a right brain meeting up together and they just don't necessarily understand each other because we've been segregated as little kids into your right-brain pocket and your left-brain pocket. But yeah, but it's really that magic of being able to trade the experience with the understanding of what the tech can do, I think, really important.

Obviously, you know, there're a lot of channels up there if you want, WhatsApp, Facebook, you know, Instagram, you name it. I think it's also important to have something which is cross-channel, you know, so you don't necessarily have a bombshell to maintain all those channels separately with different things. So, it's important to have a technology that can, you know, enable value channel and potentially swapping in between each channel. Because, you know, when we're on the move in the train, we will do something different than if we’re at home, if we have a problem with a bill, we probably use a different channel than if we are in between meetings trying to solve something.

So I think that fluidity behind the channels is important, and I think the technology needs to have that. Obviously, the natural language needs to be working well. The context, you know, remembering the context is essential and the fluidity of the conversation. So, establishing trust as you kind of go through the experience is important. So, people feel like the system got them, and they don't feel like they're going in a rabbit hole and say, “Oh my God, what's going to come out?” you know, and start being suspicious about some of the outcomes.

Obviously, integrating… all brands have a lot of systems in the back. So, the integration with the system is also key because, you know, you can have the best conversational AI system, but if it can't do the job, then, I don't know, cancelling your credit card or doing X, Y, Z, then it's just, you know, it's just a little pointless for the footprint.

So, it's a mix. But I think for me it is really… I would like to really kind of point to the experience element. I think it is really essential. 

Kieran Gilmurray: It's interesting; there's a lot in that particular cake mix. I wonder if we ever listen to our human workers and have the same expectations as we do of technology. I sometimes think when we introduce conversational AI or tech, our expectations are here, but when we're dealing with a person, we’re sometimes a little bit more forgiving or understanding… I’ve never quite understood that…

Have you seen a really great example within an industry or within a business that has actually implemented conversational AI or conversational AI experiences that excited you, delighted you, inspired you, challenged the way the business was thinking, came out the other end much better? You don't have to name a company unless you want to, but have you seen that where we're going, “That's the gold standard everyone should follow”?

Laetitia Cailleteau: Well, I think what I like in a conversational AI system is also the learning side. You know, when we used to deliver an IT project, you know, you do a lot of hard work, and then you go live, and then that's it. You’re obviously in maintenance mode, but when you're in a conversational AI system, there's all that learning, learning of different ways of expressing an issue, different subtility on how to deal with a process. So the learning element, I think, is really, really key.

And the reason why I'm bringing this up is I'm thinking a few years back, a lot of years back in the early days - and this is one of the patents we created with my team at the Liquid Studio in London - we wanted to put the telephony system with the Irish Revenue Commissioner of the time. And the rationale for them wanting to kind of put a conversational AI system was that it was a new tax that they had created in Ireland, and suddenly this particular tax was impacting the older generation, which was not in working life any longer. And therefore, you know, much more old-fashioned ways of dealing with that particular tax, yeah? Because, you know, the people were just in retirement, basically, and that was created… That particular tax introduction created quite a bit of a volume of queries coming in on phone calls.

And we started to create this conversational AI system. And at the time, we had some challenges around the pace of it because he was on the phone, and you know, at the time, you had to wait for a pause in the sentence to be able to understand what the person was saying and then, you know, decide the next step. And that was quite tricky because we thought that that particular pause was too long for the conversation to be interactive. So that's where we created that pattern—some continuous listening of the conversation so we could understand on the fly. So we didn't have as much latency in the conversation. And one of the… so we were going to go live for one week to start, you know, capturing all those data and those experiences and learn, you know, how, you know, what we needed to adjust and things like that.

And then we never switched off the system. But we had daily reviews, daily learning sessions and things like that. And you know, one of our biggest successes is we had an 85-year-old lady, which was one of the first users of the systems, and she went through, and everything got sorted. She got the tax code on whatever she needed, and she was in a special scenario of being disabled. So, you know, in that particular interaction, we solved the problems put on because we, you know, we ask her some questions, “Are you responding to a letter we sent to you?” Blah blah... so we were trying to really get under the bonnet of who was calling us early so we could really make it as friendly as possible. We saved her a lot of hassle because that would have probably been forcing some of the family to take it to the tax office and all of that kind of thing. And it could have been a, you know, big thing when it's just been less than a three-minute call, basically.

Kieran Gilmurray: Oh, I love that! It's kind of like… I don't think we talk often enough about technology for good. We hear all the bad and the horrendous stories. I think sometimes they make the news. But that's an amazing story. There's something about… there was a phrase I used before, “There's nothing more natural than a conversation”, and that is the power of this technology. It allows someone who maybe has disabilities, maybe someone who was housebound or, for lots of reasons, can't actually get to a physical location to allow them to actually interact.

Should we fear Laetitia, AI, and its creative power? Should we worry that it might become more creative than humans or present a threat? Or, as I said a moment ago, is that exaggerated or where is the actual answer? Is it somewhere in between the two? 

Laetitia Cailleteau: Yeah. And I guess the answer is probably somewhere in between the two. I am very optimistic about our future and the future of AI. I personally think there’re a lot of opportunities coming up, and they would be good for society. Obviously, it's really important to have the right guardrail, right? Yeah. And I think some of those guardrails are being established, but they're not quite there yet, yeah?

One of the examples I always use when people are asking me about, you know, the kind of technology killing jobs or things like that… I use the car example. You know, it took, it took, or I have another example of that… I can tell you afterwards when… I use the car example. So before we had, before having roads and traffic control and signs and garage to fix the car and all of that, we had a few cars out there, you know, in the middle of the oasis to kind of trying to pass their way. It’s just as you, you know, obviously it did challenge, you know, kind of going from one place... it goes back up to the bicycle and all that, but you created a whole environment on the back of that, it created a lot of jobs.

When you look now, the automobile industry is huge. And, you know, if you look at the beginning of the… hundred years ago, it was nothing! So you know like it will open doors and a realm of things that we can't even imagine yet. In the same way, I also use a little picture. I don't have it now, but it's that little boy who has a big pole and is knocking at a window, and I'm asking very often… it's in black and white, and the little boy is probably 13 or 14… I'm asking people, “Do you know what that little boy does? Is he just knocking at a window?” And people are like, “No”. And he's actually the alarm clock for the miners to wake them up before we invented the alarm clock in the 1900s. So, you know, like, who would imagine having somebody knock at the window to wake them up? So, this is evolution; this is progress!

I think, you know, there's a lot of opportunity for all of us into reinventing a better world and a new society, which I think, by the way, is heavily biased at the moment. So there's a lot to do better. And this is going to be a tool in the toolbox to actually create that. Obviously, it's important that we have guardrails because we know there's a number of, you know, bad actors, consciously or unconsciously sometimes, that can create challenges.

Kieran Gilmurray: Yeah… I wonder if we look back in 20 years time at some of the things we're doing now are our stick-at-the-window moments, so it’s quite interesting... I suppose that leads me to the next question. We can’t avoid it, Laetitia, anymore! ChatGPT is everywhere at the moment. How do you read that or see that in the context of conversational AI for business?

Laetitia Cailleteau: Well, I think it is quite revolutionary, the type of technology and architecture that's coming through. I think this is the beginning of the rupture point. There's obviously a lot of startups now who are looking at, you know, specific large language models and all of that kind of things. So, this is the beginning of a rupture point in AI technology, which is going to blow minds more.

So when we think that cloud and data are something big, and we could use some more of the traditional AI, I think this is just the next wave. So it's going to accelerate even more this collaboration between humans and machines definitely. Obviously, at the moment, it’s still, you know, it's like the car moment. You have a few accidents before you kind of invent the traffic control. Well, it's going to be a little bit the same, you know, as some of those things are opening up. Obviously, people are a little worried about the regulatory framework and the possibilities. But as soon as we get used to and control some of those topics, well, it'll be another one coming up! I think one thing is for sure is that the pace of change has never been that fast.

And that's exciting because I consider myself an innovator, and I'm like, “Oh, my God!” This is… normally, we used to wait like for years for things to come up or things to kind of make a change, you know, practically at scale. But the pace of change and the pace of technology is really kind of going much faster than it's ever been. And I think there's no way back. I think we need to want to understand for all of us, you know, how are we going to keep up with all of that? Because that's mind-blowing!

Kieran Gilmurray: It's kind of scary. There's an old English phrase, “You wait for a bus, and then two turn up. With the pace of technology, it's: you wait for a bus, and then 19 turn up!” It is kind of exciting and scary, the art of the possible. 

We talked about 20 years’ time into the future, and we talk now about the pace of change… is a little bit crazy in some regards. Laetitia, if you were to go back 20 years and you were to talk to a young Laetitia or any other young woman going into business, stepping in on the very footstep of the very first AI in the world, what would you tell yourself? What advice would you give to a younger you with a whole career ahead of you? 

Laetitia Cailleteau: Hmm. I think, looking back… so when I started in the late nineties, it was fairly, you know, I remember working for Andersen Consulting at the time, so that was my first job, and I’m still there. But, you know, we had, like... we had to have blue suits with white shirts and heels, and we used to give OCTEL to people. So, you know, we had a computer that was dialling for mail and blah, blah, blah. And we did the dot.com, right, so it kind of created the first websites for the, well, basically, brands. Yeah? And then, obviously, this kind of moved on. So this is what happened, meaning the Internet, we're just coming out 20… 20 plus years ago. So I guess, and personally, I jumped into the technology element, although it was not really my deep calling. I come from a quite humble background, and I thought, well, actually, this is the future, so I should double down on this. I wouldn't change that. I think I would give myself more compassion and optimism, you know, like more compassion and optimism, you know, embrace change all the time. I think that's what I would do.

Kieran Gilmurray: Were you very hard on yourself often in your early career? Because it's funny, looking back to the nineties… I started in ‘94 or ‘95 with DOS, Windows 95, NT4; you name it. Novell was around at the time, and I don't regret a moment of all those different experiences. I don't want to relive them… I can't imagine going back to some of the tech, but it was an exciting time at the same time. Were you hard on yourself for a period of time? 

Laetitia Cailleteau: Well, I think I'm always a bit hard on myself. I always want things to be happening now. I'm not very patient. Yeah, I’d give myself a little bit of patience, albite you know, I think this is over that time, I don't need that any longer! Things are moving much faster than they used to. Yeah, meaning, I guess… you know. Yeah, hard on yourself, you know, to kind of really going to get to the bottom of stuff and make things happen for real, you know, not talk your way through, but really kind of go on the depth of some of the things.

It's quite complicated! The reward is quite big, but, you know, it is quite intense. You know, it's not like, you know, all of this is coming from hard work, I suppose. Yeah, and I don't think it will stop. 

Kieran Gilmurray: I suppose there's, there's being hard on yourself, and then there's been driven. That very much sounds like driven. I think it's interesting today when we watch people coming into the IT industry. They won't have the experiences that you and I have seen. A lot of this space will be taken as a given and taken for granted. The complexity will be so abstracted. And I suppose there's nothing wrong with that because nobody wants to get out an abacus to do long division anymore.

And I think the very first cars, we don't want to pedal as hard as we do these days, but I sometimes wonder if there's a halfway house. But if we do see underneath the surface, maybe then the data protection rules, the ethics and whatever else might be that little bit more understood. But it's an interesting journey, and we will see what's happening in the next 10, 15, 20 years or for about and maybe I'll be like that 85-year-old lady that we spoke about earlier on talking to some robot or having some co-pilot do all the work for me so that I can sit back and relax. But let’s see, that's exciting!

Laetitia, thank you so much indeed for today's interview. I appreciate that. I've learned a tremendous amount. I think we have… and I don't think you're describing yourself as this, but I will. I think we have a pioneer in the conversational AI industry. And it's just interesting to have heard your story over the last five, ten, 15, 20 years to see how technology has changed and, how it's improved and where it might be going as well. So, thank you very much for sharing all that insight and sharing a little bit more about Laetitia from 20 years ago to today as well. I appreciate that a lot. Thank you so much!


Laetitia Cailleteau: Well, thanks a lot for having me! And to bounce back on you at 85 years old, I actually wanted to take the opportunity to talk about a project that I'm running in the UK with UKRI (UK Research and Innovation), which is called HomeCare, which is about conversational AI and older generation to really keep them connected with society and activities and things like that. And, you know, I think if some of the audience wants to go and check it out, it's a pretty amazing piece of work which I think is going to change the way we grow old in the UK at least.

Kieran Gilmurray: And what is… is there a web address?

Laetitia Cailleteau: Yes, there is a web address. I can share it with you: https://homecare.accenture.com/home. In the COVID time, we did things like food delivery and things like that. So, it's pretty amazing!

Kieran Gilmurray: It is! Technology for good! I love hearing those stories!

Laetitia Cailleteau: Yes, social isolation, you know… we’re really trying to help this particular element because, you know, we're in a quite complicated world with mobility and a generation being quite far away from each other. So growing old is actually a real… is a real topic now. And I think a lot of people don't appreciate, you know, what society did to our generations.

Kieran Gilmurray: I worry about digital exclusion. I just worry that technology is put in place and we actually create a technology-divided society. So it's fantastic to actually hear about technology bridging any gap and actually facilitating enabling. I think we need more of those stories rather than the fear stories or maybe a mixture of both. So we move at a pace, but we guard what's there as well to protect us, to protect everyone who's involved in using the technology, to protect who's vulnerable in society as well. I like that balance! 

Laetitia Cailleteau: Yeah, I mean, I've done a lot of work in that space, so I can talk to you forever on this particular topic. Really passionate about it!

Kieran Gilmurray: Laetitia, thank you so much indeed for today's interview!

Laetitia Cailleteau: Well, thank you for having me. It's been an amazing delight to kind of have all the souvenirs of my younger self all the way up to right now. It's always refreshing to think about the journey and how things are moving.

Kieran Gilmurray: It's nice to take a moment.