Tobias Dengel: Voice Tech Will Profoundly Change How We Live and Work

This podcast is brought to you by Wells Fargo Advisors. Looking for a firm your clients can't outgrow? Wells Fargo Advisors offers financial advisors full balance sheet solutions with wealth planning, investment products, banking, and lending that meet the needs of affluent to ultra-high net worth clients. Build your practice your way and on your schedule with independent and employee model options designed to meet your needs today and tomorrow. Learn more at joinwfadvisors.com.

What's the next technology that is poised to change the world? Hi everyone, I'm business coach Steve Sandusky for Barron's Advisor, The Way Forward Podcast. My guest today is Tobias Dengel. Tobias is the president of Willowtree, which is a digital product consultancy that was acquired by TELUS International in 2023 for over $1.2 billion. Under his leadership, Willowtree has become a leading provider of premium digital products and experiences for iconic global brands, including HBO, PepsiCo, Johnson Johnson, and many more. In today's conversation, we explore why voice is the next big thing in technology, as big as mobile was a decade ago and the internet in the late 1990s. Tobias says it will fundamentally alter the way companies do business. With that, please enjoy my conversation with Tobias Dengel. In your new book, "The Sound of the Future," you make a big claim, and you say that the coming age of voice technology represents the emergence of a new kind of interface that will redefine how we interact with technology. And here's where I think the big claim is. You wrote that much like the personal computer in the 1980s, the internet in the '90s, and the smartphone in the 2000s, this new interface will have the potential to touch every person, every company, and every industry. So let me know, what is the big deal about voice technology and why should we care about it?

There are so many prognostications that end up being wrong. If you were at CES in 2012, you would've heard that 3D TVs were the next big thing and we would all have 3D TVs in our houses. None of us do now. 3 years later, it was Google Glass. None of us have Google Glass. And then over the last couple of years, it's been the metaverse, which hasn't really manifested itself. And I think the lens you gotta look at technology at is what human need is it solving? And so when Siri started taking off and Alexa started taking off, we started doing a lot of research into why that was. And it turns out it actually solves a really big need for human beings. It's speed of transmitting data to devices. We speak 3 times as fast versus a keyboard as we type, 5 times as fast on a mobile device. And so that's why we all want to use voice. The problem with voice has been that the designers of these voice experiences designed it as a closed world, much like AOL was before the internet kind of went wild. And that's a big mistake. And they've designed it as a two-way voice-to-voice interface. And the problem there is we speak so much faster than we type, but we also read much faster than we listen. So the whole approach to voice deployment has not taken advantage of what makes the voice interface so powerful and amazing. And in the book, we really explain why if you can take any human action, especially an action between a human and a machine and make it 3 to 5 times faster, you're going to change the paradigm and you're gonna change the world.

The human voice has been around for what, 100,000 years or so? So why now? Is there something about the technology that's available today that all of a sudden makes this voice technology the new way that we're gonna interface?

It's really a confluence of multiple different technology shifts. And so one is just the advances in computing power. It takes an extraordinary amount of computing power to process sound and turn it into transcribed words. Google released a report 3 years ago that said they found 2,500 ways in the English language that people just set their alarm, right? So these are extraordinarily complex problems. Step one is transcribing what we're saying. Step 2 is for the machine to understand it. And step 3 is for the machine to respond in the way we want it to. Step 2 is now being solved by large language models, ChatGPT being the obvious one. But that's completely changed how good the system is at interpreting what you're saying and turning it into a command. Step 3 is then processing that and giving you the right response, whether it's doing something or the app gives you an order that you can approve, whatever it is.

You mentioned, I think, Siri and Alexa a moment ago, and that's been around for a few years now, but it's almost turned into a cliché in that they're basically smart speakers where I ask it to play a song or ask it what the temperature is or what the weather forecast is. What could possibly develop, say, from Alexa, or is it just, hey, this was a prototype and we're going to move way beyond that in the future?

Yeah, I think the analog to Alexa in a lot of ways is AOL or WAP, which came before the iPhone or the BlackBerry, which came before the iPhone. It has made a lot of progress, but it hasn't been the breakthrough that we're all looking for. It, the problem has been what we really need is not for these to be smart speakers, but smart mics. And so the whole terminology of the industry is backwards. What I want to do is ask Alexa or Siri or my app what movies are playing tonight, But I don't want Alexa to come back with 3 movies with 5 movie times each. That's a terrible experience. It takes forever. It's movie phone. It was around 30 years ago. What we want is to talk to our Regal Cinemas app and say what movies are playing tonight, see the movies, and then say, Artie, get me 2 tickets for Star Wars at 8 PM. Alexa is, I think, going to stay around in some form, but more of a smart mic than a smart speaker. And be the interface where it picks up our commands that might then show up on our computers or in our apps in a much more multimodal experience.

Let's take ordering a pizza, for example. So I want to order something from Domino's. I pull up the Domino's app, and it's cool, it's slick, and there's an app for that. And that at one point was the leading edge of technology. So I think what I'm hearing you say is that with this emergence of multimodal voice technology, I'm gonna be able to say something like, maybe it's on the Domino's app, maybe it's on Alexa, right? Say, hey Alexa, I wanna order a large pepperoni and hamburger pizza from Domino's Pizza, the one that's just down the street from me. And I want a large Coke with that and I want it delivered. And then maybe there's a screen there that shows you what your order is and you confirm, yep, that's it. Okay.

Great.

It's gonna be there in 20 minutes. Is that what we're talking about here?

That's exactly right. And it takes a process that in an app might take a minute and a half and turns it into a 10 or 20-second experience. And the hard part is Domino's, there are over a billion combinations of pizza that you can order. And ordering pizza is a pretty simple thing compared to a lot of other things in the world. And I think what the real breakthrough is gonna be twofold. As you're talking to your Domino's app, it's going to real time be transcribing what you're saying. So you can see the words on the screen. That's a big frustration for folks right now is that they don't know if the system is picking up what they're saying. That builds a lot of trust. The second thing is, as you talk, the order is gonna get assembled in the background. And then by the time you're finished speaking, the order will be there. And so conceptually, that the breakthrough is our communication as human beings forever has been call and response. And that's human to human or it's human to machine. Now we're gonna evolve into concurrent communication where real time as you're speaking or, or virtual real time, perceptibly real time, a machine will be doing what you're asking it to. And that is gonna be just an enormous breakthrough because now you can communicate input and output concurrently.

I hate to admit this. I was at a McDonald's here not too long ago and going up to the drive-through, it had what I think was one of these conversational AI systems. So instead of a human being behind the counter there, as I was ordering at the drive-thru, I think it was some kind of AI bot. It seemed to work reasonably well, but I've not seen it there again. Are they just testing that? What's the difference between a customer service bot like that versus this truly interactive system that you're talking about? What are your thoughts there?

Yeah. So these bots, they're being tested all over the place. McDonald's is testing them. They're probably the leaders in investment. Wendy's just announced a big deal with Google doing the exact same thing. White Castle just announced a deal with a company called SoundHound doing the same thing. They are all, I think it's in step one. I talk in the book a lot about what the different waves are of evolution of any technology. Wave one is you basically take what you're doing today and you try to infuse it with a new technology. So that's what they're doing. They've got drive-thrus. They're using bots to make them more efficient, especially in a very constrained world for hiring for them right now. Phase 2 is gonna be to reinvent the system, right? There's no real reason to have a drive-thru experience at all the way we know it because you could be either ordering or most likely voice ordering straight into your app instead of this whole giant screen and device, right? That only exists because people didn't have smartphones. And so then they're gonna start reinventing the process so that the drive-thru just becomes a pickup window, or it's not a window. You just pull in and someone runs out to your car. So you're gonna see businesses start to reinvent their whole processes taking advantage of voice technology.

What would be some other examples of real-world use cases for this type of technology? And if you have anything that might be specific to the financial services industry, that would be great.

A big part of the book is about financial services and a lot of the consumer-facing applications, although consumer-facing is only a piece of it, internal systems are just as impacted. So I'll give you one example. This is from US Bank. The head of digital said, we have 300 different services or features in our app, but there's no way to organize 300 features in a way that's findable. And so for them, voice, especially for what you might term long tail, experiences is critical. So as an example, once or twice a year you reorder checks. If I ask you to go into your bank's app or website and reorder checks, your blood pressure probably is starting to go up just a little bit because you know it's going to be a giant pain in the butt to figure out how to do that because you do it twice a year. What those are the kinds of things you just want to pick up your US Bank app and say, order me new checks, and then it shows up and then it does it. And so The whole concept of customer service directly into the app is going to take off very rapidly. Something like a third of calls into the customer service lines of banks are still to do simple things like check a payment or check your balance, which you will be able to instantly do via the voice experiences in the app. And so that's some of the customer-facing piece. I'll give you an a, not a financial services example, but one from from an airline, Cathay Pacific. So what they started doing is they've got, I think, 8 minutes between flights for the cleaning crew to clean a plane. And so when they do that, they're going through there fast. They've got their two hands, they're taking trash, blah, blah, blah. If they find something that needs attention, they need to stop everything they're doing, take out their phone, open the app, find the different error codes, et cetera. It takes 2 or 3 minutes and in an 8-minute process, that's just a huge problem. So now everything is voice activated. So when they're going through and cleaning the plane and they see a seat that has a broken seat back, they can say, "Broken seat back 13C," and they can keep cruising. And that just shows you an example of how, especially in jobs that require your hands to be on and your eyes to be focused, that the voice experience is super important. And it's being implemented heavily in warehousing right now, in factory floors. Law enforcement's an obvious application where you want both your hands available and your eyes focused on what's going on around you. I can talk more later about this whole safety category of why voice experiences make interacting with machines so much more safe.

One of the things that financial advisors do is they have to schedule appointments with clients for review meetings. And a lot of advisors are using some software today to do that, whether it's Calendly or some other systems where they send a link and then the person clicks on the link and it shows up the different dates and what times are available, and then the person can self-schedule. That again takes some time. Would you envision with this voice technology that somebody would say, rather than going into this app and looking up a calendar and the different times, would just say, hey, I want to meet with my advisor on December 3rd, sometime between noon and 3 o'clock Eastern time. Go check and see if my advisor is available then, or when would be the nearest time to that that they're available?

That's a perfect use case, right? Something that's some somewhat frictioned right now in terms of you gotta go in the system, you gotta look in and you can just say it and it's so much faster. Edward Jones is a big client of ours. We do a lot of work for them, but one of the things that they've gotten really good at is matching advisors to specific clients. And to do that, you need to get information from clients and typing information into any system is time-consuming and slow and people don't like to do it. But if you can get them to just say the information, from the end user perspective, it's 3 times as fast and you're much more likely to get the information you want to then do the matching correctly or to create the right portfolio for the client, those kinds of things.

Let me give you another example. This one is outside of financial services, but it's one that's near and dear to my heart, which is podcasts. So the whole process of trying to discover podcasts that have topics that you're interested in or specific areas of information that you're looking for, it's not real easy to discover that. So is it possible that with this voice technology, because podcasts are all voice, that I could simply at some point say, what are the top 3 podcasts released in the past 30 days that talk about how to use voice-first technology in marketing a financial advisory practice. And then the system would come back to me and it would say, here's the top 3 shows that you mentioned. And then I could ask it to go to the first podcast and take me exactly to the part in that podcast where it starts talking about this. Is that something that I'm just having a hallucination here, or is that possible?

That's one of the best examples I've ever heard someone come up with. And the reason is when a new technology like this hits, you have to think about the second and third order impacts, right? The first order impact is just speed. It's obvious. But the second and third order impacts are less obvious. I always talk about when Steve Jobs announced the iPhone, no one was sitting in the audience that day and said, oh my God, the taxi industry is in real trouble. Right. Gonna figure out how to do ride hailing. What you said relies on and it's a whole section in our book. The fact that search is going to change because it's so much easier to do a long tail search with, you know, the sentence you said describing what you want or multiple sentences was, I don't know what, 40, 50 words. You would never type those 40 or 50 words into a search engine to find a podcast because it's a lot of work and you are worried that if it doesn't give you the result, you've just wasted a minute or two of your time, which doesn't sound like a lot, but that's how consumers operate. Whereas if you can ask that and it's so fast to do the commitment, the time you've put in isn't that much higher. You're much more likely to do it. And then if the system works correctly using LLMs, ChatGPT-like functionality to give you the result, you're gonna start using that. And so the example you gave is the perfect example of why long-tail searches are going to become much more common in a voice environment. And consumers and employees are going to be able to get to exactly the information they want if they have these interfaces. We're working with a big company right now to do a voice interface on their HR system because it allows employees to ask very specific questions of the system that they have a hard time figuring out on their own, just typing searches or navigating documents.

So here's a frustration that we all have. There's an app for everything, right? So we've mentioned McDonald's, we've mentioned Domino's, we've mentioned Edward Jones, we've mentioned all these different companies, and every company has their own app for you to get your information specific to that particular company. So with voice technology, is it gonna be similar in that every company is gonna have their own version of the system where I'm gonna have to use their version as opposed to some big system out there like the internet? Like we all log into the internet to get access to everything else. Will there be some like overriding voice technology system where I can just— I'm going to say log into that for lack of more imagination— where then I can access everything? Or am I going to have to separately log into each different company that I want to use their product or service and then I've still got this whole issue with all these passwords that I've gotta keep track of. How do you think about that issue?

Yeah. So it's a very complicated issue. And obviously Siri and Alexa wanted to create, and Google Assistant and others wanted to create this overlay of a single assistant for everything. Turns out that's really hard to do for two reasons. One is discoverability. No one knows what the system can do and what it can't do. I don't know Does my bank have an Alexa skill or not? Is it any good? Does it— I don't even know what it does, and I can't— it's not very easy to figure out because it's all voice and I can't see it. The second thing is authentication. Yes, we have to log into our own apps, but if we were to have a system that accesses everything, it's gotta authenticate into those other systems on our behalf, which is in some ways even more complicated. What we think is gonna happen is The voice layer is gonna sit on top of websites, but mainly your apps. And so basically what we're gonna see, and it's already started, Spotify did it, DoorDash just announced it 2 months ago. Waze has been doing it for a while because obviously so much usage is in the car. But we're gonna start seeing these little mic icons in your apps and it, you're basically gonna start commanding your apps via your voice. And it's just gonna be a very easy way to interact with the brands you interact with. Every day because for the most part you've pre-authenticated. You might have to authenticate once a month. You can use Touch ID or whatever it is. The authentication process on applications is much more advanced than it is on the web. And I think the frictions there have come down a lot. We think the primary voice interface is going to be sitting on top of apps, but it's going to cause basically all apps to have to be redesigned and reinvented over the next year or two.

Okay, so that's really interesting. So on my website and pretty much everybody else's website, I have the search bar. So if someone wants to know what's on my website, they just type in what it is that they're searching for. So what I hear you saying is rather than the search bar, or maybe in addition to, we're going to have that mic icon, we click that and then we just speak what it is that we're searching for and then it's going to search and find that. Is that basically what you just said?

Exactly. Yep. Okay.

And then we talk about big data. So it sounds like we have to be even more cognizant of how we are capturing data, how we're organizing data, how we're structuring data so that it can be accessed through voice technology. Would that be fair to say as well?

Yeah. And I think that's where LLMs play a big role. Right? Because LLMs have gotten really good at processing trillions and trillions of pieces of data and making them findable, searchable, and then creating responses from that data. That's been the big breakthrough. In late September, ChatGPT announced what it said was its biggest innovation since it launched, which is making it voice-powered. And so that's, you finish writing a book long before it publishes. We were finished writing 4 months ago, 5 months ago, and we predicted The next big innovation for ChatGPT is gonna be voice powered, and it happened to coincide right about with our book publishing. Sometimes it's better to be lucky than good. But the point there is that this marriage of conversational AI or voice technology and generative AI is what's gonna unleash all this innovation.

ChatGPT just now integrated the voice into it, but there were other apps before that because I remember when ChatGPT 3.5 came out back in, what was that, November of '22. Not long after that, I actually did a podcast where I connected a voice app to ChatGPT and had an interactive conversation with ChatGPT. And it was really fascinating because I played the role of a consumer and the interaction with the AI was about, do I have enough money to retire?. And so I asked the AI to play the role of a financial advisor, and I'm a consumer who wants to know if I have enough money to retire. And then I asked the AI to ask me, what questions do you need to know from me so that you can tell me whether or not I have enough money to retire? And so we went with this back and forth for about 20 minutes, asked me a bunch of questions. Ultimately it came back and said, yes, I have high confidence that you have enough money to retire. And then I said, okay. Given the information that you have, are there other questions that you need to ask me to determine what is the likelihood that I'm going to run out of money before I die? And then there's another dialogue. So it's really interesting, and I'm a financial guy, and I thought, this is actually pretty good. And it was that interactive dialogue that I thought was pretty cool. So that, to me, gave me a glimpse of the potential of this type of voice interactive technology, which I thought was really cool.

Now, that's a great example. And I think the next iteration of that is twofold. One is the responses you get from ChatGPT might be on a screen because it's so much faster for you to read the response and to listen to ChatGPT's voice. But the second is you're using a general-purpose large language model in ChatGPT 3.5 or 4 for that. Most companies as they deploy this kind of technology are gonna use much more customized LLMs just around financial services that are, A, much faster and much better trained just in financial services and much less expensive to run. And so the combination of that voice and those specialized LLMs is gonna create the final version of what you were experimenting with.

There's always talk that when it comes to new technology that either A, it's going to eliminate jobs. And I think for the most part, people realize that every new technology, yes, it does disrupt jobs, it eliminates certain jobs, but then it also creates a whole lot of new jobs. And if we look at where we're at today, despite all the technology advances that we've had over hundreds of years, we still have relatively low unemployment. So I think most people would say, I'm not too worried about that. But then a second thing, particularly for financial advisors, is if we have this interactive voice technology where an investor can have a conversation with an AI, just like the example I gave here about do I have enough money to retire, which is a common question that financial advisors help a consumer solve. What are your thoughts as far as humans being comfortable having conversations about money with an AI as opposed to a human being? Is it a situation where people will over time get comfortable and the advisor will be cut out of the equation? Or as most people in our industry seem to think, it's not going to be AI only, it's not going to be human only, it's going to be AI plus human is better than either one of those individually. What are your thoughts?

I think that's the right thinking, at least for the foreseeable future. It's gonna take a while for these voice and AI-based tools to get the trust of the end users, and there will surely be missteps along the way that reverse some of those trust mechanics. Long-term trust, cognitive trust is ultimately based on does this person or this machine do what it says it's going to do on a repetitive basis. And so as AI gets better at doing that, more trust will naturally develop. I think in the short term, medium term, 5 to 10 years, it is gonna still largely be humans in the loop. And same thing goes for medicine. We spend a lot of time in the healthcare field working with hospital systems, et cetera, on how this is gonna manifest itself. And it's gonna be a support system for doctors and nurses and caregivers. Step one, and then slowly over time as it gets so good that it's 99.99% accurate on certain things, it's going to replace the human being. And then the natural progression we've seen is that there will be more and more things for humans to do in different ways, but it will displace certain humans. The, the number one area, of course, is contact centers or call centers where in the near term there will be less need for humans and more of that will be digitized, but there's always gonna be second and third order problems that you need the human beings for. So it'll be an interesting evolution. I think what we've learned in the last 30, 40 years is we can't ignore those as a society, right? The globalization backlash that's happened over the last 5 years is because, yes, there were net gains in jobs overall because of globalization, but that didn't change the fact that very many people were displaced and lost their jobs, and I don't think we sufficiently took that into account. And those people are angry, and we're gonna have to deal with the same thing with the proliferation of AI.

Yeah, and that's such a good point because there's a disconnect in that the technology can move much faster than humans' ability to keep up, to retrain, to learn the new technology. And so there's always that gap between technology's way ahead here, and we displace humans who haven't had enough time to really catch up to what the technology is doing. And so the ability for government, for society, to try and shorten that gap or somehow make sure that we don't end up with a lot of displaced people, which creates all kinds of problems, is certainly a good point. Do you think the powers that be, let's call it, are paying enough attention to the potential ramifications, the second and third order effects of that, as you mentioned?

I, I would say definitely not. I think this has to be a classic case of federal, state, and local governments working together because a lot of these programs have to be administered at the local level or using the community college system, et cetera. There are so many jobs that are evolving due to AI and emerging due to AI that are not super technical, But they do require training. There are things like prompt engineers that didn't exist basically a year ago that is now proliferating or conversational designers. There are going to be new jobs, but historically, the education system and the government system has been, as you point out, 3 to 5 years behind industry and behind the private sector. And we gotta figure out how to shorten that loop.

I wanna go back to something you said a little bit earlier when you were talking about what basic needs does voice technology meet? And one of them was you mentioned speed. And I think in your book, you talk about a total of 6 of them. So what are a couple others that you think are important that we should be touching on?

Yeah. So I would say the second one that's really important is safety. Look at the Deepwater Horizon accident in the Gulf of Mexico 10 years ago. You look at the Boeing MAX 737 MAX accidents, they were all ultimately driven by human beings' inability to communicate with a machine, a device. And voice is the perfect cut-through for that. After the Deepwater Horizon, some researchers at Stanford created a voice-based system that would've allowed that accident to not happen. Voice-based cockpits are used right now, and we don't know if they're deployed or just in testing, but both by the United States Air Force and the Russian Air Force. And those are examples of where you're creating it. If you think about it, we've interfaced with machines forever using knobs, buttons, pedals, keyboards, our fingers, right? And that's not a way that humans naturally communicate. They naturally communicate with each other using their voices. And if we can cut through that and allow the humans to use their voice to tell these machines what to do, it's going to be not just faster, but much more efficient. So that's a very big category. I think the other category that we haven't touched on yet is inclusion, where using voice technology, we can bring people into the digital world that do not have access. And it really falls into two categories, folks that have disabilities. So there's a whole section in the book about an app we built called Vocable that allows users with no speech or significant speech impediments to use their eyes to communicate with the device based at pointing at words. But the voice part of it is that a third party in the room, a caregiver, can be asking the patient something, and then the app using ChatGPT-like tools comes up with likely responses. And then using eye tracking technology from the app, the user can just respond very quickly to the caregiver. And so it's completely changed the experience with, for folks, for example, with ALS to be able to speak. And the second thing is either the illiterate, there's still a billion people in the world that are functionally illiterate or speakers of very small language groups. I think India as well, hundreds of languages. And so many of those small language groups, there's no digital experience in that language. There are just too few speakers for you to create apps or websites or anything in that language. And using voice-based translation tools, all those users will now have full access to the digital world.

Tell me about your company. So what exactly is the role that your company plays in this voice technology space?

We've been a digital products company. What does that mean? We build apps, websites, et cetera, for clients. And we got super interested in voice, as I said, 4 or 5 years ago as we saw the proliferation of Alexa and Siri. But we realized really quickly that building Siri or Alexa skills specifically wasn't going to be super interesting because they just weren't important enough for our clients. But we knew voice was going to be a big deal. So we have gotten really focused on this combination of generative AI and conversational AI to create these new experiences, which largely we think will be sitting on top of apps. Sometimes they'll just be direct experiences or literally on a phone that you call into a customer support center. And the agent on the other side is all a voice slash GenAI experience. But that's really what we're helping clients with is this entire customer experience, right? Because a customer often starts with a website or an app, and then eventually they run into some trouble. So they might talk to the app, they might do a phone call, and all that has to be tied together because there's nothing more frustrating than when if you call into a call center and the first thing they say How can we help you, Steve? And you're like, I've just spent an hour trying to rebook my flight. I wish you could know that I'm not starting from scratch right now. So tying all that together is, I think, the next generation of customer service.

I want you to close your eyes for a minute, and that's you, Tobias, not the people listening to this that are probably driving somewhere. But if you close your eyes and you just think about 5 years into the future, what can you imagine this voice technology being 5 years from now? How will our lives as human beings be different because of this voice technology?

I think there will be a voice layer to the entire world around us that will be second nature to all of us, just like the internet is second nature today. If you go back to the late '90s, I was at AOL at the time. I remember the executives from Coke coming in and being like, we don't understand how the internet's ever going to have an impact on Coke. And then the next phase was a realization that it would. And what did everyone do? They started an internet department. Like, these are our internet people. Now the internet's everywhere. You can't imagine anything happening without the internet. And I think the same thing is gonna happen with voice to some extent. It's gonna be generational. This is a true story. On Sunday night, my 10-year-old's favorite team, the Dallas Cowboys, were playing, and I sent him to bed at halftime. And he was super upset, threw a tantrum, walked into the kitchen. He's in 5th grade and asked the Google Home Assistant, hey Google, tell me what boarding schools take 6th graders. And Google responded, Worcester Academy. And he said, hey Google, write me an application to Worcester Academy. And so he's a voice-first kid, right? When he asks, when he interfaces with the television, either through our cable remote or his Apple TV remote, it's all voice. He just turns it on, asks for what he wants to see. It's that the next generation is gonna expect everything to be voice first.

This is purely speculative, but Steve Jobs, of course, brought the mouse. He didn't invent the mouse as an interface, but he certainly popularized it. Do you think if he was alive today that he would be at the front end of this voice technology? And what might he be thinking about how to use voice technology as an interface that perhaps we're not placing a high enough priority on today.

In the book, one of my favorite interviews was with Adam Shire, who's the co-founder of Siri. And he tells us, and, and we detail it in the book, how he founded Siri and then how he got a call out of the blue one day from Steve Jobs to buy Siri. And it was because Steve Jobs was so focused on voice being the platform of the future. His idea was to launch Siri, make it applicable to the 10 or so apps that Apple ships with the phone, which they did, and then very quickly make Siri available to app developers to use in any app. Steve Jobs then died. One of the most interesting things about Steve Jobs' death is there's stories all over of people that are near death that hang on for one more event that they want to see. I think the most famous is that both John Adams and Thomas Jefferson died on the 4th of July, 1826, on the 50th anniversary of the signing of the Declaration of Independence because they were both just so focused on being alive for that last day. Steve Jobs died the day after Siri launched, and he was so focused on voice. I believe if he had lived, he would have pushed the voice agenda much further. Than it ended up getting pushed. We'll get there. It just might be 5 or 10 years later than we would've gotten there.

Let's talk about his nemesis, Bill Gates. So as I was preparing for our conversation here today, I was in Microsoft Word, and Microsoft Word has a microphone, and it's had a microphone for quite a while. So I would click on the microphone and I would start speaking my notes for this conversation, and it would transcribe it in real time. It actually does a pretty good job. Of transcribing it. What can you imagine that Microsoft could do with the Office suite of applications that goes beyond just transcribing what I'm speaking to it into a Word document?

Microsoft bought Nuance, I think, 2 years ago at this point, which was one of the absolute leaders in voice technology. And I'm not sure, but I'm pretty sure that most of the technology that they deploy now throughout their suite is all based on the Nuance technology suite. It's really about the marriage of this transcription and then the ChatGPT-like logic. And I think the transcription works pretty well, and it, especially for folks that don't have a quote unquote accent, et cetera, it works arguably 98, 99% of the time, which does make it much, even if you have to go back and correct things, much more efficient than typing. But the problem folks run into just in a Word doc, for example, is if, if they were trying to do extended doc, just the formatting, et cetera. Now that's something that the ChatGPT-like LLM logic is really good at. And so this is again an area where the voice technology and conversational AI married to the generative AI is going to solve a ton of problems because if you dictate most of a paper, ChatGPT or tools like it are going to be really good at turning that kind of rough draft into an almost final draft, and then the human will come in and do the final draft. But that process should take a fraction of what it takes today to the human being to write a whole document by hand. And so that's where I think that evolution is going in. It's going to, you're going to be able to tell the system that you want an Excel spreadsheet that has blah blah, blah, blah, blah, blah blah blah and it and pulls data from here and there. And then you'll probably have to fix it a little bit, but it'll. All the annoying stuff where you're like writing formulas and getting everything just right is going to get automated voice first.

There was a time when a small startup could completely disrupt an industry, and I think obviously that can still happen today. But it seems like today when a startup has a great idea— and I think OpenAI is a perfect example who came out with ChatGPT— but then it essentially gets purchased by Microsoft. They make, what, a $10 billion investment in it? And maybe this is more of a larger technology question, is this technology going to be co-opted by the Microsofts, by the Googles, by the Metas? Is it going to be controlled by these big companies, or are we going to see a new flowering of startups that will become the next generation of those big companies?

I think we're going to see a flowering of startups, and I think it's going to happen in two ways. One is on the technology side, where the investments that have been made in the last 12 months in AI-based startups are, it's the biggest wave since 2008, 9, 10 in mobile. And so I think you're gonna see something, big things emerge from there. And I think it's gonna be largely industry-specific solutions is where most of the money is gonna be made, where it's not a general purpose tool like ChatGPT, but it's something targeted at the financial services industry and maybe a very specific piece of the financial services industry and dominates that area and completely digitizes it. Very analogous to emergence of the internet where all these startups emerge to some of them unsuccessful, pets.com, some of them highly successful. And so I think that's gonna happen. And then the second layer is always companies using this toolset to disrupt their marketplaces and disrupt their competition. And that's where smaller companies can outcompete bigger companies in their space if they use these tools more quickly in productive ways. And I think in every technology wave we've seen that. We've seen a lot of companies who are adopting the new technology more quickly are able to gain market share.

But then will we end up with these handful of behemoths today in the tech space that end up buying all of these companies and continuing to control everything, whether it's the Amazons and those types of companies?

I think there'll be a big piece of that. I, this winner-take-all world that has emerged, I think is going to continue because they have such ability to cross market and to create these ecosystems that are just really hard to break down. So however you view the world, fortunately or unfortunately, I think that's going to continue. I think generally, It's interesting, Amazon is being charged by the government now for running monopoly-like operations in certain areas, and their defense is based on the fact that it's a benefit to consumers. And I think they tend to show that. So it's a cost for their competitors, but a benefit to consumers. And I think that makes it much more complicated than the implementation of the monopoly laws in the '70s or even going back 100 years where it was clear that the monopolies were disadvantageous to the consumers because they were price gauging. That's not necessarily happening. So we could do a whole nother podcast on the cost and the value of these tech behemoths to society.

For sure. Tobias, I think we will wrap it there and want to give you one final opportunity here to share any comment, thought, idea that perhaps we haven't touched on yet.

Yeah, I think It's just such an exciting time. We had our summer interns that left a couple months ago, and I told them I was so jealous of where they were right now because they were basically starting their careers at a time like 1998 or 1999 or 2008 or '09 when it comes to mobile, that over the next 3 to 5 years is going to be an innovation wave like we haven't had in a decade. And they're going to have opportunities that are once in a lifetime. And they're young, so they're gonna be super quick to understand and to embrace this voice-first ecosystem. But I encourage everyone to start thinking about it because it's gonna hit us real fast.

And do you encourage those interns to go out and start their own businesses, or do you encourage them to go work for one of these startups?

I always say you gotta pick your own path. My path was I worked for a big company, AOL, for 7 years before I became an entrepreneur and I had to learn how to do it. There are obviously plenty of examples, Zuckerberg, Gates, Jobs, who just did it on their own. So it's who you are and whether you feel like you gotta cut your teeth for somewhere else or you feel like you're ready to roll.

You mentioned AOL. So we'll, we'll ask one more question here. What did you learn from Steve Case, the founder of AOL?

I think the most important thing was Persistence. AOL became this incredibly important company in '95, '96, '97, but it started a decade earlier and it took blood, sweat, and tears to get it to the breakthrough stage. And I think those values of just grinding it out and being patient are so important for any business leader. And he was of course a tech visionary, but the secret sauce was just grinding it out and believing in it and motivating the team around him to get there, which they did.

And if folks wanna connect with you or learn more about what your team is doing, what would be the best way to do that?

There's two ways. Find me on LinkedIn, Tobias Dengel. I'm the only one out there. Or look up Willowtree online at willowtreeapps.com. Excellent.

All right, Tobias, thanks for being on the show.

Thank you, Steve.

All right, that's all for today. Make sure you like and share this podcast through your favorite social platforms, and for more great podcasts, visit us at barons.com/podcast. Take care and be safe.

This podcast was brought to you by Wells Fargo Advisors. Looking for a firm your clients can't outgrow? Wells Fargo Advisors offers financial advisors full balance sheet solutions with wealth planning, investment products, banking, and lending that meet the needs of affluent to ultra-high net worth clients. Build your practice your way and on your schedule with independent and employee model options designed to meet your needs today and tomorrow. Learn more at joinwfadvisors.com.