Podcast #13

Inbenta’s CEO on
the future of AI, Sci-Fi,
and customer service

Jordi Torras
CEO and Founder – Inbenta

About this podcast

Today we interview our very own, Jordi Torras. Jordi is the CEO and Founder of Inbenta.

Having been in artificial intelligence since the 80s, we dissect the past, current, and future of AI. We cover everything from customer service to Science Fiction.

Jordi founded Inbenta in 2005 to help clients improve online relationships with their customers using revolutionary technologies like artificial intelligence and natural language processing.



Explore the power of AI and NLP
for your customers and agents

Interview Transcript

Welcome to the Future of Customer Service Podcast. I’m Andrea Palten, from Inbenta and I will be interviewing customer support and service professionals to see what is currently working well, what issues they’re trying to overcome, and the future success of customer service.

Andrea Palten
Hello, today we have the CEO and Founder Jordi Torras from Inbenta. Jordi, thank you so much for being here. You are on your own podcast. How are you?

Jordi Torras
That’s right. Thank you Andrea. I’m so and so happy to be here and thank you so much for having me.

Andrea Palten
Yes. So, I thought we’d do a little bit different questions this time because since you have been in AI for a very long time, you found your own company, there’s going to be some questions that I have for you that are going to be a little bit different than I ask our customer service guests. So, are you ready?

Jordi Torras
I am. Let’s go for it.

Andrea Palten
Okay, so how did you get started with artificial intelligence anyway, where did it all begin?

Jordi Torras
Well, I started in college with learning artificial intelligence, as I’m a computational scientist, by training from the University of Barcelona. I got specialized in artificial intelligence and that was in the 80s. Can you believe that? That’s been a long time. So, things like neural networks or expert systems, or genetic algorithms, that was the hot thing back then. What I learned is essentially, artificial intelligence is a set of software or hardware techniques that allows us to do things that were not possible to be done before.

This is what artificial intelligence is in this area, this border, where several tasks can only be done by humans, and then computers can take over. So, that is the visual intelligence, which means that it is dynamic, that it’s constantly growing. In the past, when I was studying software that was able to effectively, and with good quality play chess was considered artificial intelligence. Now, it’s just an app on your iPhone, and nobody thinks about that anymore. But that’s exactly what happens. Artificial Intelligence is once it works, nobody calls it artificial intelligence anymore.

Andrea Palten
Yeah, so when you were studying it in the 80s, did you guys foresee what it would be like today and how literally everybody uses AI without even knowing it?

Jordi Torras
Well, one of the areas of AI that I was studying is neural networks. I remember working on that with the Microsoft personal computer with an 8086 with eight bits, and 640 kilobytes of information. It was very limited, and it was very promising, but it’s totally abandoned. So, I would have never said that such a boom would kind of comeback nowadays with deep learning, and machine learning, and all this kind of rediscovery of the concept of neural networks. So, it’s pretty amazing. Other things like expert systems, or genetic algorithms have been abandoned. Who knows? Maybe they’ll come back when new techniques and new software and new technologies are discovered or developed

Andrea Palten
Yeah. So, you graduated. Where did you work? What was your first job?

Jordi Torras
Well, of course, after graduating my preoccupation was to find a job. So, I was assigned, I was working as a consultant, software developer, software analyst for several consulting businesses. I was so blessed to be invited by my older brother, back in the 90s, ready to start our own business. So, we developed that business for several years. We sold that business in 2004. It was a great project. My brother had an offer and all that and said well, you know what, I am retiring. I’m the youngest of five siblings. I’m talking about my oldest brother. So, he said, okay, it’s good. I am retiring. Then I said, okay, we’ve sold that business. Let’s start something else and that’s how Inbenta started.

Andrea Palten
Where did the name come from? Inbenta.

Jordi Torras
That’s a good question. Well, if you look at the verb in English invent, to invent something. In Spanish, we would say, Inventa. But you know in Spanish, B, and V sound the same. So, many Spaniards and many Spanish speaking folks would make that mistake when typing the words. So, it’s actually misspelled. So, the word Inbenta is a misspelled version of the idea to invent or to create with your imagination. So, that’s how we started with a misspelling.

Andrea Palten
I never asked you this, it took me what, nine months to ask you this?

Jordi Torras
All right, you got it. And also, back in time, inbenta.com was available. So, that was another reason.

Andrea Palten
I know, it’s so hard to find that. So, what made you start Inbenta? So, did you see a hole in the marketplace? Did you want to buy something, and you couldn’t, so you made it yourself? Where did this come from? Why did you do that?

Jordi Torras
Well, I started Inbenta to kind of put together creativity and talents into helping customers with one specific mission, which was finding information. That’s how we started and in reality, when we started Inbenta we were not a software company at all. We were a consulting business, and we were helping our customers with search engines, and lexicons and taxonomies, and synonyms, and all these things trying to make existing search engine technology be more intelligent. That’s when we realized a few years into the story of Inbenta when we were, as I said, basically a professional service and consulting business. That is not possible. You cannot take an existing keyword-based search engine and make it intelligent. That is not possible.

So, we had a great team of specialists in search technologies. We had a fantastic team of computational linguists who knew very well about how to model language. Then we envisioned hey what if we create sort of a search engine but that instead of using keywords, it uses meaning? So, a way to search meaning? So, we say, okay, we can build a software that essentially takes the user question, takes the content, computes the meaning, and tries to match that meaning. Not the words, but the meaning of these words. That’s how the idea of our search engine was created and that’s how we basically transitioned from being a consulting business and a professional services company into a software company and more specifically software as a service.

Andrea Palten
So, fast forward to today, who are some of the clients? What kind of solutions does Inbenta serve up to these clients?

Jordi Torras
Yeah, absolutely. So, we started originally from Barcelona. So, as you might imagine, most of all, the oldest customers that we have are in Spain. We extended it to France. So, we have a lot of presence in Europe, with companies like Allianz or AXA or BBVA that are using our technology. We started working in Brazil with Portuguese because one of the things that we realized is actually our software is not subject to borders in terms of countries but languages. Language is really the idea that makes us grow into different territories. Then I decided to go to the US. I came here with my family to start a business globally and in the US. So, now companies like Groupon or DocuSign, or Pinterest are some of our customers and they use our technologies to basically help their customers to locate answers more quickly. That’s essentially what we do.

Andrea Palten
So, let’s talk about artificial intelligence a little bit. So, you’ve been obviously studying it since the 80s, you’ve been working in it forever really? What makes good AI and what makes bad AI?

Jordi Torras
So, I believe that what really makes bad AI being bad, bad, bad is over-promising and hype. When you see this kind of thing out there and in some cases is just well I guess journalism. So, you need sensationalist news out there claiming fantastic achievements from AI? Well, I mean, I could understand that when you see in some cases, companies with some claims. They say, man, that’s not true, this is not really happening. Then I believe that’s kind of more dangerous. It is dangerous, not because it might hurt anybody, it is dangerous because, at the end of the day, it creates a delusion, and then when real companies are starting to use that they say, well that’s not the magic that we were promised. That creates frustration, and that frustration then creates what we call the AI winter.

By the way, you say that they’ve been working with AI forever. Well, basically in the 90s, there was, believe it or not, there was also a boom in AI. After that what we call the AI winter arrived which was a general dissatisfaction and a general impression that well AI would not work. It hasn’t been until recently that AI is getting momentum again, but in reality, this AI winter took pretty much a decade to be totally gone. I believe that the risk is when you go into over-promising on what the technology is able to do. That’s a real risk because you enter in a number of promises, sometimes fueled by science fiction books and movies and whatnot.
So, that’s what sometimes makes this industry a little bit frustrating. But on the other hand, when you see the real progress, and you say, okay, we know we don’t have general AI. We know that. General AI Andrea would be when you have artificial intelligence, that is as capable of any task as a human would be able to do. That is a general AI. Well, that doesn’t exist, period. Specific AI is when you have specific artificial intelligence systems that are able to do a specific task as good as a human would do. That’s specific AI and this is the field that we are, and to see the general AI, we are not there yet.

Andrea Palten
Yeah. So, you touched upon this. You were talking about some misconceptions about AI, that it can solve everything, and we know that it can’t. We’re not there yet, or we might never get there. What are some other misconceptions that people have about AI?

Jordi Torras
Well, I believe that maybe not in the last months but maybe last year, there was a lot of information circulating out there that AI could be dangerous, dangerous to humankind. Somebody said some brilliant mind said well, that’s the biggest danger that humankind is facing. I believe these claims are preposterous for the reasons that I tell you. Assuming that an AI can actually come and become dangerous because the AI might want to survive or my one way might see humans as competitors, or these things that are the story. Thousands of science fiction movies are far from reality. I believe that misconception really comes from confusing intelligence with will. Maybe you are intelligent, but existing or not, it doesn’t really matter.
What if we have an AI that is able to think but it has no particular desire to live or not live or simply doesn’t care. That’s the kind of AI that we will build that will liberate humans from doing repetitive and boring tasks and it will give us the freedom to use our fantastic human brain to do things that are really great, like art or music or inventions or science or so many aspects where we humans are pretty good. We can live the most boring stuff to be done over and over again by AI because guess what, AI will not get bored and it really doesn’t care. It doesn’t feel pain. It doesn’t feel joy. It is just a machine. So, maybe we’ll have to get used to the idea of intelligence without conscience, intelligence without feelings. It’s weird to think that way but we’re going to see that more and more.

Andrea Palten
I love all those sci-fi shows like Battlestar Galactica is one of my favorite ones. It’s always the same thing. AI gets smart and has feelings and has will and they destroy everything.

Jordi Torras
Correct, Correct. I love horror movies and I love these movies where there are ghosts and houses get haunted and all that. But I don’t believe in it. I would never feel terrified by that because I don’t believe there are any ghosts. But still, I like movies. So, same thing here.

Andrea Palten
So, let’s talk about customer service and how AI helps customer service. A lot of our listeners are in the customer service space, either Director, VP, that type of level, or they are customer service agents. How does AI help their lives and their departments out?

Jordi Torras
Absolutely.

Andrea Palten
Actually, Inbenta’s AI. I’m going to talk about the solutions that Inbenta has that helps customers?

Jordi Torras
Absolutely. So, very much during the 90s, we saw the proliferation of call centers, or now will we call it to contact centers that sometimes is a place physically crowded sometimes with folks in it, answering phone calls. That’s kind of the idea when we think of a call center. And for many years, that was exactly what it was. Then what happens is, every person working in this call center, has a computer, has a phone, and gets calls. Now, every person uses that computer to do something with the system. They change the password, they assign a refund, they change your address. They do things and in reality, if you think about it, the number of things that every call center agent is able to do is limited. You call the call center for a big company and you ask sorry, what is the average speed of a cheetah? They will say well, I don’t know. I’m not here to answer these kinds of questions. What they’re able to do is essentially understand what a customer wants.
We hear their human brain and translate that question into something that the system can do, the system that they work on because that’s what they can do. There’s nothing else that they can do for a customer that cannot be done with that terminal. So, if you think about it, if you think in these terms, what is happening is the only thing that a call center agent is doing is transforming human language, which is questions and concerns and remarks from their customers into digital actions that the system is allowing them to do. So, that’s what they do and, and our AI is saying, well if that’s exactly what they do can the machine do that? In order to do that well, it seems easy, but it’s not because everybody calling speaks English, Spanish, or whatever natural language that they do. It’s not easy at all for a computer to understand that.

A computer can understand the comments and the menus and whatnot that the call center agent is doing but is not understanding what the real question is. Our system is doing exactly that translation, taking user questions and transforming those user questions into digital actions, or at least doing this for 80 or 90% of incoming calls. So, actual call center agents will be able to do more difficult tasks and tasks that require more value-added and a human touch or more complexity. So, this is what we do for our customers, we essentially help them find ways for their customers to find information and get effective very quickly, online with a phone, WhatsApp, text message, and any digital channel.

Andrea Palten
Yes, love that. Yeah. That’s a really good explanation. I love that. Tell me Jordi, what do you think is going to be the future of AI? Where are we going with all this? Maybe let’s talk first in the customer service space, what’s the future there? Then let’s go into fantasyland, where is AI going in the future-future? But I want to talk about customer service folks first. What are they going to be expecting? What’s the future there?

Jordi Torras
Well, imagine Andrea that you call because you have to call center for a specific company and immediately you get an answer. You get an answer over the voice or a chat or your phone, and whatever you want and immediately someone is answering your questions, is understanding what you say and giving quick and effective answers and actions to what you want without any need to wait or any need to repeat or any need to be transferred to another agent. In fact, you get those answers in a way that is indistinguishable from what an actual human would do. If you would really not know and not care if you are talking or chatting with a machine or a person because your call was resolved. Imagine that this is the kind of future that we are building at Inbenta. We want to make this customer service for AI indistinguishable from a human. So, that’s a service and that’s the future when it comes to customer service that we want to see.

Andrea Palten
Yeah. It’ll make life easier for the customers, for customer service agents, for everybody. I love that.

Jordi Torras
Exactly. Exactly.

Andrea Palten
Let’s talk about outside of customer service. Where do you think the future as a whole of AI is going?

Jordi Torras
Well, let’s say this. Let’s say okay, the intelligence of a fly. Let’s put it that way. So, the intelligence of a stone is zero. We agree on that right.

Andrea Palten
Yes. I agree.

Jordi Torras
Okay, okay, the intelligence of a human is 100. Let’s put it that way. Okay, in this space, from zero to 100. Well, if zero is a stone, hey, computers are more intelligent. We are making progress. So, we are like at one or two then you start analyzing other organisms like a fly a mouse, a dog, a chimpanzee, a human. There are different, different numbers on that. We, at this point, are at 10 very much. That’s exactly what we are. First, there’s no general AI as I said. They are only able to solve specific problems. In reality, those problems are not that complicated for a human. Let’s talk for a bit about customer service.

Again, it is known that the industry of call centers is packed. Now, a lot of employees work in a call center temporarily, they want to work on something else. They don’t plan to work in a call center forever because it is boring and that’s exactly the solution. So, what is a boring job? A boring job is a job that does not require your full attention, your full intelligence, your full capacities. You will only use for eight hours of work or four or whatever you work, just a smaller part of your brain, and then the rest has nothing to do and that’s when boredom shows up. That’s a job that you don’t want to have.

So, let computers take that part and let’s free humans for working on what they want. So, that’s kind of the future of AI that I see is in which areas humans get bored. Let’s remove them from our existence forever. That’s what AI comes into effect. If you look at how humans evolved over, I don’t know, a few hundred, thousands of years, started using animals to do our work. That was a great invention because you know what a horse or a cow these guys have 10 times our force. So, that’s let’s use them. But then the Industrial Revolution arrived, they say, well, now, we can actually get better than these horses and we can have one machine that is so much more powerful than my horse. We still today call about horses to measure the power of a machine. So, still today. So, we measure in horses. But now we have the capacity to create machines and software and technology that is able to do what we still have to do with our brain.

This is where AI comes, replacing things that we have to do over and over again and freeing us from doing other things. So, I believe that AI will be the next Industrial Revolution. I don’t know how that’s going to be called but it’s when we will be able to build systems that really do what we don’t want to do because it’s boring. I believe that will multiply. Imagine all the brains in all the humans in the world being free to do whatever they want, to create, to experiment to do arts or science. What will happen is it will effectively accelerate the progress of the human race. That’s what I believe that AI is not here to, drag us down, it is here to help us to progress even faster.

Andrea Palten
That’s so beautiful. AI is not here to drag us down but to progress even faster. We’re going to end with that Jordi because that was a really good soundbite. Thank you so much. I really appreciate your time.

Jordi Torras
Thank you. Thank you Andrea. I really appreciate it.

Thanks so much for tuning in. This podcast was brought to you by Inbenta. Inbenta Symbolic AI implements natural language processing that requires no training data Inbentas extensive lexicon and patented algorithms. Check out this robust customer interaction platform for your AI needs. From chat, bots to search to knowledge centers and messenger platforms. Just go to our website to request a demo at inbenta.com, that’s I N B E N T A.com. If you liked what you heard today, please be sure to subscribe to this podcast and leave us a review. Thank you.