Podcast | Winning hearts and minds with process automation – why more consumers are opting for self-service

Jean
Shin

Director, Strategy & Content | Podcast Host of Mobile Interactions Now

30 min Podcast
José Horta, Co-founder & Chief Innovation Officer at SimpleTECH

In this episode, we check in with José Horta, Co-founder & Chief Innovation Officer at SimpleTECH. José’s fluency in machine learning and process automation will help us understand how businesses are starting to use self-service as a customer-facing differentiator.

Podcast transcript

Jean:

Jose, welcome to the show. I am excited to have finally deep dive into really what is making process automation possible. It happens to be a topic that I think is becoming very much of a customer facing differentiator. But before I get totally distracted by that, I would love you to just quickly tell us what you do and who you are and start us from there.

 

José:

Hi Jean, it's a pleasure be here, my name is Jose, I'm the Innovation Director and Co-founder of Simply Tech and our [inaudible 00:01:13] two companies that we run with some partners and they're focused on unified communications and applying value to companies in their communication to the customers and other stakeholders.

I have a PhD in computer science and did my PhD in France about IOT blockchain applied to the energy sector, and has nothing to do with I do with my companies, but with the companies around with my colleagues, I'm a telecommunications engineer, and we created the companies before I started my PhD. Then at some point I started my PhD, left the company came back a couple of years ago and I'm focused on the innovation part in making partnerships and creating new products, new experiences with customers so that's my goal in the company.

 

Jean:

Awesome I simply could not have picked a better person to have tackled this topic. Let's start from very basic understanding of how we can approach this meaning when it comes to automation of most anythings, I think part of the decision is like what to automate, what not to automate, is there any general approach you take with your clients or when you're developing, how to think about this?

 

José:

I think that the best use case that you can have for automation is not only when you look at us a way to risk through the use costs for a company, but also simultaneously to provide let's say, an improvement in the user experience. So I think in those cases, when the user goes through a process that is cumbersome, that's takes time that it's annoying or it's complicated or whatever. That's a huge opportunity for automation and if simultaneously, while you automate the process for the user, you can reduce to reduce costs to the company that's providing the services, that's the honeypot.

 

Jean:

I have a long list of things that I hate doing, but I have to do when I need to get something done, so I'm assuming from outside of looking in the list can be quite different from, inside-out kind of analysis, but I know you guys worked on many different use cases, but there are some exciting things I saw in the press as well, some of the things that you're working with automating some claim processes and things like that, using that as a more specific example, can you tell me what you just talked about in terms of a use cost versus the whole pain that users have to go through, how that come about?

 

José:

In that case, I know if you ever had an accident of something, but it's a very, very stressful situation and as a user, you are in a very stressful situation you have to call your insurance broker and then your insurance broker would tell you "I'd have nothing to do, you have to call the insurance company." Then you call the insurance company and then insurance company, at least here in the region, they would say, "Okay, you have to wait there, we'll send you a technician to fill in the claim form take pictures and so on." So you have to wait there and usually people wait between 30 minutes and one hour and once they arrive, they take pictures, they ask you questions and so on, and then they leave and you get nothing, I mean, you don't get the claim report or whatever they send it to you 24 hours afterwards.

So for the user it's kind of the worst situation ever. What we did is focused on the user if we focus only on the user experience, what we tried to do is to ease this process and avoid them to have to wait, so what we provide them is a way of declaring their insurance requesting for making a claim for the insurance company directly through WhatsApp.

There's an automated part that would allow them to go through a process of more or less five steps, nice identification, so who you are, what's your plate number, these kind of things. Then the second part of the process is asking for the location, or when it happened and where it happened. The third part is sharing documentation, the fourth is sharing pictures of what happened, I mean, of the car damages, these kind of things, and declaring what happened, what was the damage? What did happen effectively in the process?

So instead of like 30 minutes, one hour waiting and so on, it takes eight minutes and you can leave because you are able to leave the place in only eight minutes, and usually when you have an accident with another person, if you are able to leave because your insurance allows you because they provide you with the service and the other person has to wait, it's a kind of, I can leave because I have this insurance that allows me to do this, so I can leave and you have to wait, okay, just wait and I'm leaving. It's like in eight minutes, you're done and you leave.

 

Jean:

Yeah I mean, it's funny though that the context is such unfortunate situation, but there is that whole show-offish, look what I could do that you cannot, I don't have to put up with this kind of moment and I think that is what I was hoping in a way, the whole user facing differentiator where I think we went through the maturity curve enough to know that some of these digital experiences are really not just a little task automation or anything like that, it's becoming real differentiator.

Can you give us a perspective in terms of why it is happening even more so now and what makes it more of a now moment to do this kind of things, despite some of the legacies still happening and issues happening with insurance industries where those factors might not have changed that much.

 

José:

I think it's mainly due to the channels that the user had to make a claim, for example, or to do these kinds of things. Usually was the phones, so you have to call somebody and be in a phone call with somebody and then solve the things, or even physically you have to get a technician to go there wherever, because they had to take pictures and things like this.

So I think with social networks and the opening of social networks with APIs for allowing, automating this kind of communications, so not only being able to chat with an agent, but also complement the human interaction with automated interaction, that is the key that allows for plenty of opportunities, it's not that were impossible to automate, it's just that, it wasn't that easy as it is now, so everybody has WhatsApp or has Facebook messenger or Instagram, whatever.

So I think now it's easier because, you don't have to go through all this process of developing an app or a website and the experience of the user would have been anyway cumbersome and making it evolve, it was also very hard to do.

So maintaining an app and updating the user experience in the app now with the social networks and the platforms that we build that are no code, and you can modify something in the experience and it's instantaneously applied to the experience, it's not that you have to update the app, for example, download the last version of the app, no, doesn't happen, you don't need to wait.

 

Jean:

Yeah, I remember the whole big promise of a digital transformation was very much hinged upon people having to download the app and install the app and I remember those times where people are very optimistic, brands are very optimistic, people would download their apps and do this, all this and being part of it.

And last episode, I was talking with a solution architect here in Germany that works with a lot of airline companies, as well as automobile companies and he was mentioning two things about why the adoption, all those things are not happening as fast as some might have expected, but something like WhatsApp is kind of solving that, but they solve the device flavor, varieties and all those things issues so you kind of having different skin and that's kind of a failure of industry until some of this whole major platforms came in and provide this. Do you agree in the context of what you were just talking about?

 

José:

Yes, I mean, I think that was one of the keys that there is some like companies as WhatsApp or Facebook that are maintaining their apps and they're putting their apps everywhere, so everybody has their apps, so now they have this tool to get to information. For example, this one use case of automation that we may talk maybe in another session, but I mean more deeply, but it's voice biometrics authentication, usually for applying biometric authentication, you have to have an app because you have to know that the picture was taken at that time.

You have to be sure that there is a concept that's called likeness, that you have to be sure that the user is there taking the picture and for example, in WhatsApp, we are able to implement voice biometrics authentication because if you push the button of recording the audio, we know that you are recording the audio at that point, so it's not that you are attaching the audio.

If you attach an audio, it's a different endpoint on the API, so we know that the audio that's coming for processing was recorded at that point in time, so that gives you likeness and enables you to implement plenty of things that otherwise you would have to develop an app and put it on the hands of the user, which has a huge cost.

 

Jean:

This is such a pattern, QR code, having seen the first way of QR code adoption, this new wave that is coming and we're seeing it everywhere was really surprising, but the thing is, it became that after the device providers start embedding it into their camera functions, that you don't need a scanner once that little extra step, by having to have a separate app to solve that little problem that consumers are not willing to bother with, once that is solved by some other means I think the adoption is getting broader. Sometimes it sounds obvious, but is there specific indicator or is there something you're seeing really experiencing the broader adoption customers are really using it, and what the difference is?

 

José:

For example, claiming insurance automation, at first you have to call anyway, you have to call the call center and when you call the call center, they will tell you, "You want to continue through the call center, or you want to go through WhatsApp? You also can do it in WhatsApp." At first, the first time we have been running this service for a year and a half or something like that.

First, it was 15% or 10% of people was like choosing to go through WhatsApp and now it's 40%, 40% of all the claims that are done for San Cristobal that's a company that is implementing this use case, 40% of these claim insurance are declared through WhatsApp and they plan to, by the end of this year to get it to 50-50, 50% of people being declaring their claims through WhatsApp.

 

Jean:

I'll be the first one if it's actually offered from IVR and the voice recording is offering their choice and we'll take this, so it does sound intuitive and you mentioned eight minutes to go through the entire process or something like this and versus what kind of, from initiation to resolution, whole claiming, a completed claim, how much of a difference we are talking about?

 

José:

After the technician leaves as I mentioned before, you didn't finish the process, you didn't get the PDF, you declared the insurance, you made the claim, everything is fine. In the other case, you have to wait until let's say 24 hours afterwards, you would receive the claim by email, so this is huge also.

So these eight minutes, we are trying to make it lower this time with the next version, because this has been hugely evolving I mean, it's very easy to make it evolve, to test some experiences, evaluate them, make some changes to make improvement to the user experience and one of the things we saw it was kind of cumbersome for the users and I experienced it myself in the accident I had, it was that you have to, for example, the declaration of what happened in the core version, you have to write it down.

So you have to write, I was going through this avenue and I turned left and whatever, this process now, we're training an AI for speech to text, so they would be able to speak what happened and this would translate in text in order to put it on the claim system, so the declaration in the claim system and print it on the PDF afterwards at the end of the process.

We tried this at the beginning, one year and a half ago, and this was not possible because the noise around, because you are in the middle of the street so, but the maturing of technology, AI and particularly voice recognition technologies has evolved so much that now you can do these kind of things.

 

Jean:

This is wonderful, the user is in the same place, similar level of noise and all that and now just technology is simply gotten better to separate the noise from the voice and then transcribing it into written text.

 

José:

You get to train the system because you get audio and then you train the system in terms of noise, background noise, and how to tune the system to take out this noise and leave a cleaner voice.

 

Jean:

I'm curious, are you seeing decline in terms of the handover to Asians being requested, how is that looking?

 

José:

Just very, very low, I mean, it was almost non-existent, so if we realize that the user at some point has some struggle with the process, if there is the first attempt it fails and you offer a second attempt and it fails, then you send it to the agent.

So this is what we are looking for in what we do is to complement the work of the agents with the work of the automation, so that you can go and come back and go and come back, whatever you need, I mean, as a user, you're following a process at any point, you may be needing a clarification or something like this, so you may go to an agent and come back and continue the process.

 

Jean:

So far, we've been talking about changes that's happening on the users end, the customers end, but now you opened the back door a little bit here where the agent can have the visibility into the process that is happening of those users and when they get stuck, they can intervene at any moment and then put the process back into automation and things like this. So tell me a little bit about the changes happening on that end the agents end in the back.

 

José:

In that sense, I see like a big moments kind of, you have the call center industry that was used to phone calls. They're kind of trying to add to their solutions, but they are not that advanced in what is conversational, but through chat, through social networks, these kind of things, so they are trying to move to this place of conversational because I mean, everything is moving there.

And on the other side, you have companies like us that were bringing this app by design based on this new experience, on this new demand from customers, so I think that's an opportunity, so we're kind of rebuilding kind of a call center experience in that sense, but based in different things, different interactions in a pictures position, sharing the history.

 

Jean:

When you look at this particular success, were you thinking of where else it can scale up to?

 

José:

I think there are two ways to scale this, I mean, kinds to replicate the same use case in other big insurance companies, for example, this is one case or using all these learning to provide all their use cases similar to these, maybe not exactly the same, for example, with another company with the LBC, La Boliviana Ciacruz, this is the biggest insurance company in Bolivia, we're doing the outdoor inspection.

So they request, their customers to make an inspection of their vehicle every year and the inspection takes, you have to take pictures of the different parts of the vehicle and send some information. The most complicated thing maybe is that companies don't really know all these possibilities they have. The biggest challenge we have is to make them learn that they can do that, that they have these possibilities, because I don't think they know that, I mean, I know a big insurance company, maybe they don't realize that they have these kinds of possibilities.

You also have this kind of the problems of AI, there is too much information and maybe they imagine some advanced things that you still cannot do because they, as I mentioned one year and half ago, the speech to text was not working very well in noisy situations, but now it is, so maybe there is a lack of reality check, it's kind of, or you go to the magic use case that is not really possible today.

So maybe they find out some information that realize this is not possible and you have the reality today, so there's something in the middle where our experience is that, for example, applying AI not for, let's say trying, for example, one example is companies may come and say, I want a chatbot that would look like a human, respond like a human, and that's not what users are looking for anyway. If you try to do that, you will be spending a lot of money, not having any, let's say, any increasing conversion or whatever and you will leave frustrated, both the user and the company.

If you, rather than that, you focus on using technology in what the technology is good for and identifying that it's not easy. It's not easy for us that are in technology and it's even harder for companies, so I think they're kind of lost in this kind of in all of this information and they don't know what's really possible and what's really not possible. So we're using AI in things that are applicable and real today for solving problems that are specific, not trying to do that magic of speaking like a human.

 

Jean:

This was never done before kind of thing.

 

José:

Yeah, these kinds of like magic and I think you will get frustration in, for everybody.

 

Jean:

I absolutely love this and that's just such a nice wrap of the first segment that we are having right now, because I do think I hear two different things in what you just said it really jives with me is that I think some people were in a kind of an early part of the hype cycle where you kind of a try something because it was hyped and then you have disappointment.

The incremental improvements that happened from that because we happen to be in a very, very fast moving industry that people are not looking at it as a real opportunity where now you finally become something very, very useful and I absolutely love that.

 

Sign up for our newsletter

This article is part of the Connections Now newsletter. You can sign up here to receive it every month.