We’ve all been there. The aggravation of calling into a customer service line with a problem that you have to explain over and over again, wasting your time and energy. Even nice humans like Betsy and Tony see red during these interactions. Enter Peter Voss and AIGO.
From a garage to a company of over 400 people, Peter has seen it all (and researched it all) over the past 20 years of his career in AI. AIGO is committed to creating chatbots with brains and bringing hyper-personalization into the world of AI. Think about having a dedicated concierge service that actually remembers you and any previous interactions you’ve had with them. How much time would you save as a customer? What if your organization had technology like this for your customers to deliver a better brand experience? For Peter, the answers to these questions lie in the technology at AIGO.
This episode is for tech geeks and customer-focused execs alike – listen now!
Watch the episode here:
Listen to the Podcast here:
About Peter Voss
Peter Voss is a Pioneer in Artificial Intelligence who in 2001 coined the term ‘AGI’ (Artificial General Intelligence) with fellow AI luminaries. Peter is an Engineer, Scientist, and Inventor as well as a Serial Entrepreneur and Leader.
Started in electronics engineering, then fell in love with software.
The first major success was developing a comprehensive ERP package and taking that company from Zero to 400-person IPO in Seven years.
Fueled by the fragile nature of software, Peter embarked on a journey 20 years ago to studying what intelligence is, how it develops in humans and the current state of AI.
This research culminated in the creation of our natural language intelligence engine that can think, learn, and reason — and adapt to, and grow with the user.
Currently focused on commercializing the second generation of our most advanced ‘Conversational AI’ technology called ‘Aigo’ (say I-go) Aigo.ai – Chatbot with a Brian, is the Most Advanced Natural Language Interaction Platform built on human brain-like Cognitive Architecture. Aigo.ai is on a mission to help enterprises deliver exceptional conversational experiences with Chatbots for their customers and employees.
Chatbots With Brains With Peter Voss
Creating A Better User Experience And Human Experience
We have an amazing guest, Peter Voss. He is the Founder, CEO and Chief Scientist of Aigo.ai based in Los Angeles, California. The short version of it is that he’s worked in artificial intelligence for many years and he has created a chatbot with a brain. That’s what we’re going to talk about. We’re going to go into some pretty deep stuff. I warn you, at the outset, I totally geek out in this episode because I fell off the bandwagon on this one here but I’m a recovering data junkie.
Let’s be real Tony you have not recovered. I don’t even see that happening. It was a great interview but it’s also not just a technical interview. It’s about the impact of what AI can do for the world, how companies can protect their brands by focusing on the human experience versus “customer experience.” It’s about so much more than the technology but if you have an interest in the technology, this will also appeal to you broadly. Without further ado, let’s talk to Peter Voss from Aigo.
Let’s jump right in. Thank you for being here. We’re so excited to have this conversation with you. Peter, so our audience can get to know you a little bit better as we already have had a chance to, give us your background, your career path that led you to where you are now.
I started as an Electronics Engineer. I started my own company. I fell in love with software and my company turned into a software company. That was very exciting. It was being able to create by writing code, being able to create applications. I designed a fairly comprehensive ERP system for medium-sized businesses. I built my company around that. We were quite successful. We went from the garage to 400 people and did an IPO. That was super exciting. When I exited that company, I had some time available to think, “What is the big problem that I want to work on?” What struck me is how stupid software is. I’m saying that I was very proud of my own software but if the program didn’t think of something, it would crash, give an error message or something. It didn’t have any common sense. I wanted to understand what is intelligence and how can we make the software more intelligent.
I embarked on a five-year research journey to deeply understand what intelligence is starting with philosophy, epistemology, theory of knowledge. What is reality? How do we know anything? Cognitive psychology, how do children learn? What do IQ tests measure? How does our intelligence differ from animal intelligence? All of those kinds of questions to understand intelligence and to look at what had been done in the field of AI. The combination of that was that I came up with a design for an intelligence software engine that can learn and reason more like humans do. In 2001, I then coined the term Artificial General Intelligence together with two other people. We wrote a book on the topic and that’s to bring AI back to the original dream.
When the term AI was coined years ago, it was about building thinking machines, machines that can think, learn and reason the way humans do but that turned out to be difficult. The field of AI has been one of pursuing one problem at a time or narrow AI. We coined the term Artificial General Intelligence, AGI, to recapture that original ambition of AI. That’s what I’ve been working on since 2001. The various companies are splitting my time between R&D development and commercializing the technology. My company, Aigo.ai, is the second generation of commercialization of the technology. We’re creating intelligent conversational AI or how we like to call it a chatbot with a brain.
We need to step up to improve customer service.CLICK TO TWEET
Peter, can you give us an example of exactly how this works for your customers? We want to dive more into how it solves problems for your customers and why this is different. Take us down the path of why this is impactful for your customers.
We are focusing on working with larger enterprise companies. Pretty much every large company we talked to, either has chatbots or has tried to implement chatbots. Universally, they are unhappy with the results, both from a customer experience point of view and from a cost-saving point of view. The simple reason for that is that the chatbots that are out there have no brain. They have no intelligence. To be specific about that is they don’t remember what you said two sentences ago. Never mind what you said in the previous conversation you’ve had. They don’t have any deep understanding of what you say.
They just extract keywords, key patterns and then they respond to that. It’s sort of a stimulus-response, you say something like weather, “You probably want the weather report.” You could also say, “I hate Uber. Don’t ever give me Uber again.” “Here’s the Uber app.” They don’t have a deep understanding, memory, the ability to reason or common sense. Those are very severe limitations. If you had a human personal assistant that didn’t remember what you said two sentences ago or picked up on keywords, that wouldn’t be very useful. That’s the frustration that large companies have with trying to implement chatbots. Quite a few of them have given up on it.
It’s the definition of knowing your customer from a technology standpoint. Applying what we talk about on this show, really knowing your customers. I’m curious about as you’re running this company, how do you instill that culture of customer centricity within your organization so that it flows out to your customers as well?
That’s often said that the company culture comes from the top and I’ve always been that way. To me, producing a quality product has always been my main motivation. Maybe some investors don’t like to hear that but that’s what drives me more than the profitability. I believe that the quality of the product comes first to me and then the profitability happens as a secondary thing. For example, we have a VP of Customer Delight, so that’s how we think about it. We obsess about doing that. Our technology, as advanced as it is, is not at a human level. You have to say, “What is actually possible?”
It’s a fine balancing act to say, “What do you want to automate? What can you automate successfully and what can’t you automate successfully?” Also, very important is how can you integrate? How can you hand over from artificial intelligence? From an intelligent chatbot, how can you hand off elegantly to a live agent where that’s necessary? It’s integrating the system deeply into the customer’s ecosystem, both in terms of the backend, knowing as much as possible about the customer, being able to use it and then being able to integrate that with other channels, whether they’re existing FAQs or help channels.
I’d love to dive a little bit deeper into this idea of the chatbot with a brain. With my experience back in 2007, I started working with text mining platforms to look at the feedback that was coming through surveys, emails and those types of correspondence. There’s a real challenge that I found with those systems and I apologize to the audience if I get a little too technical here but it’s an important piece. Most people don’t understand that in language, there are the content words, which make up 98%, 99%. It’s the nouns, the verbs, the adjectives, descriptions.
There then are the functional words, which are the personal pronouns, the articles, the conjunctions, the things that make the sentence work together. The biggest challenge in a lot of these mining platforms that then eventually have moved into AI is being able to understand and value those functional words. If I pick out Uber, weather and those key content words, I may have no real understanding of the context of the sentence. How does your chatbot potentially look at the whole context of the sentence, the meaning of the sentence, not just those keywords?
There are a couple of things involved. One of them is deep parsing. You don’t just do pattern matching. The way other chatbots without a brain work is they have a categorizer, which is a large statistical analyzer that you train with lots of example words and then it takes the input. It simply says, “Which of those possible outputs does this relate to?” It’s categorized. It can only fit one of maybe a few dozen or a few hundred slots. In the case of Alexa or Siri, it might even be a few thousand different slots but whatever the input is will always trigger that one intent. There’s no analysis of the actual sentence, the input sentence. Statistically, these words are not even a sentence. It’s just these words. Statistically, it is most likely to correlate with intent number 47 or whatever it might be.
What we do is analyze the sentence and do a complete deep parse of it. Get all of the clauses, the subclauses, how they are related, picking up negations, resolving pronouns and other indirect references that you have. It’s not good enough doing a parse by itself. You also have to do it in the context of the conversation because to resolve pronouns or indirect references like that, they’re often referred to things that we said 1, 2 or 3 sentences ago and also taking the context into account. What are you trying to do at the moment? Had you tried to place an order? Are you talking about an existing order? If we’re talking about troubleshooting a modem, have you already tried to reboot your system three times? Whatever you’re saying might have a very different meaning or different implication. It’s the deep parse of deeply analyzing the sentence within the context of the conversation. It’s necessary for us to then figure out what you should be doing, what is the next step that you should do.
Two things come to mind as I listened to you say that. The first one is an experience I had with my cable provider, who will remain nameless. That very experience of I had rebooted my modem. I have gone through all the troubleshooting. I got to the end and they said, “How satisfied are you?” I wasn’t very satisfied because it didn’t solve the problem. It put me back into the same loop again to start the whole process over. It’s not very satisfying. I love this idea of the conversation and want to touch on that for a second because it’s been my dream since 2007, looking at this feedback of customers. One survey comes in or an email comes in and none of these tools took the whole conversation in mind.
We built processes manually. We would use the system to tag and then we would feed it back into the system again with those additional contexts tagged, if you will. We would be able to look at across time. This is an interesting place that I’d like to explore. We could look at conversations over the course of 6 months or 1 year with specific groups of customers that had behaved. Maybe they canceled their product or service. We will look back. What did they say over the last six months? We start weaving those conversations together. How would your platform do something like that to give this intelligence either to an analyst or more likely to a customer service agent so they can do the right thing at the right time?
There are a couple of things here. The big companies all have a lot of data, have a lot of computing power so for them, it’s all creating buckets of what is your profile, what bucket do you fit into, whereas ultimately you want to have hyper-personalization at the individual level. While statistics may be important to make general business decisions, from a customer interaction point of view, you don’t want to put a customer into a bucket and say, “That’s the profile. That’s how I should treat them.” Ideally, you should say, “This particular customer has this particular history. His modem is in the kitchen. It’s this model. He’s called three times. It was replaced a month ago or whatever the case may be. You have 2 dogs, 1 cat and 2 young children. You live in Oregon.” It’s something quite unique.”
The brain that manages the conversation consists of, you could think of it of three layers. The core layer is general intelligence that knows about people, places, how to start a conversation. There’s general knowledge on how to hold a conversation. The second layer is the knowledge that is embedded, that is trained for the particular company and application that you’re doing. If it’s about modems, what are the different products, what are the different names, what are the rules about them, which are obsolete, which are new, whatever you might know about them? The outer layer is unique to every single customer. That remembers the conversations that you’ve had before. It’s also via APIs that will have information that the backend system might have, that information from other interaction channels or whatever might be in the backend.
That is how we can provide hyper-personalization at scale that every individual user essentially has their own Aigo brain. To get back to the question asked, “How can service use it?” Even though each user has their unique brain, we can extract information across all of the different conversations that have been had. Where are smooth conversations happening? Where do the conversations get stuck? When do you need to transfer to an agent? When do people hang up? When do you need to retry?
You can add to the knowledge base that you have to the ontology and to the skills that the system has to handle those or if they should be handled by a live agent then to recognize it and hand over. The handover to the agent will be with a summary of whatever has already happened in the conversation. One of the absolute, the most frustrating things in big business is you call in, they need to transfer you to another department and you start from scratch, your authorization, validation and your history. We all know that especially dealing with banks.
We want everybody to have intelligent conversations with AI.CLICK TO TWEET
My sister and I have this thing where if we’ve had a bad experience with customer service tech support, whatever it might be, we end up calling each other to talk us off the ledge like, “I’m a nice person but this bad experience turns me into an upset, frustrated, human.” That’s horrible for a brand, for people to engage with a company that you’re giving money to and then end up feeling so horrible about the brand. The market is huge for what you do. I don’t think there’s a single person on the planet that has not had either a great experience or a horrible experience but either way, they remember it. I’m interested in your thoughts on how that impacts brands and how your clients talk to you about that piece of why what you do is so important.
There are two elements to it that our customers and corporations are interested in. One is their customers have a great experience. That’s important. Often they come to us and say, “That’s the most important thing.” When push comes to shove, how much money can we save? I hate the word containment. They want to contain a conversation. It doesn’t have to go to an agent. It’s finding the balance between those two things. Fortunately, quite often once we get involved with a customer, we find that there are fundamental holes in their system that need to be plugged, that we can’t overcome.
A perfect example is in my previous company, Smart Action, which focused on automating phone calls, whereas, in Aigo, our focus is on chat. Unfortunately, we found that a very large number of corporations simply didn’t have the mechanisms to transfer the information that our IVR with a brain got to transfer that to an operator. To me, it’s still incomprehensible that companies come to us and say, “We want to improve customer service,” they didn’t put the effort in to transfer that. It wasn’t even a matter of cost because we were so keen on providing that customer experience. If we said, “We’ll provide the technology to you free of charge,” we want whatever we’ve gathered in our IVR conversation we want that to go to the live agent. We still had a sizeable number of customers that had so much inertia. There wasn’t enough of a priority. I find that frustrating when companies won’t do relatively simple things to improve customer experience. People engaging with us come to us because they’re frustrated with the limitations of dumb chatbots. We’re stepping up to improve their customer service.
It’s funny. I had an experience recently. I had a doctor’s appointment. This doctor I’ve been going to for years. He was out and there was another doctor that was taking over for him that day. I had not met this doctor before. One of the first things he said to me is, “Are you spending a lot of time at the lake with your boat?” He started chatting with me about stuff that I was like, “How do you know this?” It’s because the previous doctor took good notes and transferred that information. What that did was build instant trust. A) they care enough to write down those kinds of things but B) it made me trust that they are in this for me and that they took the time to transfer that information. Being able to do the same through digital means is it’s the same result. You have that built-in trust like, “They don’t want me to have to repeat my story three times.” That’s one of the things I love most about the work that you’re doing and the level of frustration that you’re saving the world by being able to figure this out is significant.
That’s a great story. In fact, often people still think of automation as being inferior to talking to a live agent but if it’s within the capabilities of the system, it’s a much better experience because there’s no wait time. You can’t always staff a call center above 100% to keep them under 100%. An agent will always be immediately available. Very few companies can afford to do that. They have peak periods. With automation, you have an immediate response. You have massive scaling and instantaneous scaling. It’s 24/7. A lot of companies can’t justify having 24/7 but even more importantly, Aigo intelligent chatbot can have photographic memory essentially, know everything that is known about the customer that can help to deal with whatever service request there is.
It’s almost like having a dedicated concierge service of somebody who remembers what your previous interaction is, which in a live situation you can’t ever get that. You’re not likely to get back through to the same person on the call center. Even if you were, they wouldn’t remember the previous call. With automation, you can achieve a much better service, quicker taking into account the history and then a very consistent experience because that’s another thing that call centers struggle with is training everyone to the same level of quality. Typically, it’s some kind of a bell curve or something in a way you have very few good agents and most of them aren’t that great.
I love this because you’re speaking to with the right technology, how you can create an excellent experience for the customer. It’s the hyper-personalization and that memory that matters. It gets to the heart of our show, really knowing your customer. Really remembering your customer might be a whole another episode that we have to do. I’ve got a question. I’ve worked a lot in the financial services industry and I know that would be a target market for your application. One of the challenges that we saw happen quite regularly that would cause this massive influx into contact centers, chat, email, phone calls would be when there’s a data breach. There’s one several years ago like Home Depot. Their credit card got breached or you could take any of them that we’ve seen Target. All the major brands have been through it.
This is not going away anytime soon. How does your system adapt to that? What we found was on day one, people are calling in about, “Was my card breached? Was my information breached? What do I need to do?” Over the next week or two weeks, the conversation evolves. People are calling in about different things. Once they feel calm about their card, then it’s, “How do I reconnect it to this account and that account, so it automatically pays?” How does your system potentially adapt to those sudden changes and then migrate with the cloud as the cloud moves across the horizon?
I have a very good client. With conventional chatbots, usually they are siloed. We’ve spoken to some banks. They have separate chatbots that get implemented for different things. These chatbots don’t talk to each other. They don’t have a common brain anything or they don’t have a brain. With our approach, it’s essentially that we’re talking about the three layers. That middle layer has all of the information about the company that’s relevant, business rules, it has APIs to the backend system and so on. That section is used for all of the different types of conversations that you might have. What that allows us to do is to very easily and quickly add functionality to the system. You already have that corporate brain as it were in your chat infrastructure. That’s already in place so to add some functionality in terms of saying where you’re breached or not and how you answer that question. As it evolves over the days, it’s very easy for our system to add that functionality.
In fact, some of the things we can teach a system in natural language, even to teach another skill. That’s one of the other advantages. You can get the system to adapt to emerging requirements. It’s getting that initial integration going that’s the biggest job. In terms of data security, people are increasingly nervous about the SaaS model. We’ve seen a couple of nasty incidents. Our system has been deployed behind the customer’s firewall. It’s not a SaaS application. We provide the chatbot brain technology to our customers that they integrated into their existing infrastructure. That may still be, in many cases, a cloud service but either their private cloud or a cloud that they’re already securing, they know how to secure. Our chatbot gets implemented behind their firewall. At least we did not add into the potential holes that you can have.
Peter, clearly you’re a visionary. What does the future look like for you and for Aigo?
We want everybody to have intelligent conversations with AI. We think it’s high time that people reject these dumb chatbots. It’s not necessary. It’s a little bit of an accident of history that people are okay with that. The big companies have a lot of data. They have a lot of computing power, so that’s a hammer they’ve got so everything looks like a nail. They say, “We can provide a chatbot with this statistical big data technology.” It’s the exact opposite of what you want. It’s not the quantity of data. It’s the quality of data but because these big companies are driving it, Google, Amazon and IBM, that’s the technology that they have and know. That’s what was provided. Enterprise and end-users should not put up with that. There is a better solution available. What I see in the future is that we will insist that every chatbot has a brain, every IVR has a brain.
Going out further, this is a separate topic. Let’s look at things like Siri and Alexa as personal assistants that are there. Do we ultimately want these dominant systems that are owned by a mega-corporation, fulfilling their agenda rather than the user’s agenda or do we want personal assistants that you own, that you control, that is hyper-personalized to you and are intelligent? We see ultimately intelligent chatbots being available to individuals to help them in their lives. They own it, control it and it’s theirs. We call that a personal, personal assistant. In fact, it should be a personal, personal, personal assistant because of three different meanings of the word personal. Personal, you own it. Personal, it’s personalized, customized to you and personal, the information it has is private and you only share it with the people you want to share it with. That’s the future I see that people will insist on having these intelligent personal assistants and that they have more control over them as well.
I’m glad Betsy asked the question because I wanted to ask that question, too. Knowing where your technology can go, it seems to be a logical step for individuals to be able to plug into a company’s chatbot with their own personal, personal, personal assistant, rather than having to start from scratch every single time. It’s an interesting and compelling thing to think about. We have covered a lot of ground. Is there anything else that you want to talk about that we haven’t specifically chatted about yet?
There are lots of things that we can talk about but our focus is on quality service to customers and how we achieve that. This is what we breathe and live. We love doing it. As much as I like the technical aspects, understanding intelligence and building intelligence systems, to me, it’s nothing if it can’t be implemented in the real world and help people. For our enterprise customers, it’s helping them in two ways, offering a much better service to their customers and saving them a ton of money at the same time. That’s a real win-win. I’m excited to be doing what we’re doing and hope that more people find out about what’s possible and that they will start insisting on having intelligence in their conversational AI.
Peter, thank you so much. What’s something that our readers, Tony’s network and my network can do to support your efforts moving forward?
Spread the word. Let us know if you have a company that could benefit from an intelligent chatbot. Contact us. Let more people know that you don’t have to put up with dumb chatbots anymore.
One of the things that we like to do on our show is to give our guests an opportunity to give a shout-out to a charitable organization or some organization that’s doing good in the world because we all need more of that. Is there an organization that you’d like to tell us about and that we can support by virtue of spreading the word?
There are many worthwhile charities but the one I particularly like is Institute for Justice that maybe not as many people know about. It fights for the common man injustices through government regulation, overreach or eminent domain abusers, where big organizations will use government or their power to do injustice to individuals. They’re doing a terrific job.
Thank you, Peter. I know Tony has been geeking out on the technology. He’s a little more tech-focused than I am but we’ve both enjoyed this conversation so much. We’ll be following to watch where this goes. I honestly believe there’s not a person on the planet that won’t benefit from the work you’re doing. We know you’re very busy. Thank you so much for agreeing to be a guest on our show. We look forward to seeing what you bring to the world in the future.
Thank you so much. This was fun. Thanks for having me. I also look forward to having seven billion people have intelligent AI personal assistants.
That’s going to be an awesome world.
Thanks, Peter. We’ll talk to you soon.
Thank you. Bye.
Betsy, this interview with Peter was phenomenal. The biggest takeaway I have is this distinction between general AI and narrow AI and the real value it can bring to the customer experience but also to the world as a whole. This is so much bigger. We talked about in the context of customer experience but it’s the human experience and what can happen as general AI spreads out there and as it’s applied in so many unique and different ways.
What it brought to mind for me is when you and I were writing the book and we were talking about human experience, not just customer experience, yet how that approach can also be doing exactly what he’s doing with AI but the idea that everybody can learn from what they’re doing at Aigo to provide that great human experience. I’m with you. I hate to admit how frustrated I get when I’ve had a bad situation with customer service or tech support. I’m a calm, reasonable person. If I can get that frustrated, imagine the energy level of the world, how much better it will be when things are a lot more seamless.
In the future … We will insist every chatbot has a brain.CLICK TO TWEET
As a side note, in 2021, one of the practices I put into place is eliminating annoyances. It’s fascinating because what I have discovered is it’s not just thinking positively. It’s getting rid of those things that are draining my energy and my emotions, causing me to focus and pay attention to the things I don’t feel good about and the unpleasant things of life. By eliminating them, it has changed who I am and it changed how I look at things. I find joy in the little things in a much different way than I had before. I bring that into this conversation because if you look at the billions of hours that are spent every year waiting in lines, dealing with customer service issues, being frustrated, then calling your sister and having a conversation about how bad that was, we all do it. We waste our life. We only have 70, 80 years, whatever it is on this Earth. If we waste years of it accumulated over time on unpleasant experiences, dealing with things as a customer, that’s a real shame. There’s a real injustice, to put it bluntly. That’s one of the reasons I work so hard in this space because I believe we ought to be living every moment of our life the best way possible. I’ll step off my soapbox but that’s my passion here.
I’m with you. It points to the bigger why. I love when companies have that bigger mission that is a positive impact on the world. If you looked at this on the surface and say, “It’s an AI company,” you don’t get to that bigger why and what it means for the world. I love what he’s doing. He’s clearly passionate about it. He wants to have a big impact on the world. I enjoyed this conversation. Tony, thank you. It’s always a pleasure having these conversations with you. I feel so fortunate to get to do this with all these interesting guests and to get to do this with you. We’re going to keep doing this. Thank you to our readers who are supportive and keep tuning in. We appreciate it. Until the next time, we will see you on the show. Have a great day.