Podcast Image: E15: Alexa's $25B problem, AI upends marketing & search, sommelier's tech insights | Kyle Tsai

E15: Alexa's $25B problem, AI upends marketing & search, sommelier's tech insights | Kyle Tsai

Chris Rod Max dive into the challenges facing smart assistants like Alexa and Siri, exploring their limited use cases and struggles to monetize despite massive market penetration. Guest Kyle Tsai (ex-product manager, founder & sommelier) joins to discuss how AI and large language models could reinvent these platforms. The group debates the future of search and digital advertising in the age of AI copilots, potential opportunities for startups, and the need for marketers to adapt to a rapidly changing landscape. Plus, Kyle shares wine pairing recommendations for tech enthusiasts!

Host

Rod Rivera

๐Ÿ‡ฌ๐Ÿ‡ง Chapter

Guests

Kyle Tsai

Investment Manager, Wayra UK

Chris Wang

AI Innovation and Strategy Expert, CXC Innovation

Max Tee

VC Expert, AI Investor, BNY Mellon

E15: Alexa's $25B problem, AI upends marketing & search, sommelier's tech insights | Kyle Tsai

In this episode, the hosts and guest discuss the challenges and opportunities of smart assistants, such as Alexa and Siri. They explore the limitations of current smart assistants and the need for innovation in the space. The conversation also touches on the potential for startups to compete with established players in areas like advertising and search. The hosts and guest share their perspectives on the future of smart assistants and the role they could play in our daily lives.

Takeaways

  • Smart assistants like Alexa and Siri have limitations and need innovation to reach their full potential.
  • Startups have opportunities to compete with established players in areas like advertising and search.
  • The future of smart assistants may involve more personalized and conversational interactions.
  • Contextual advertising and AI-enabled marketing strategies are emerging trends in the industry.
  • The role of smart assistants in our daily lives is evolving, and their capabilities will continue to expand.

Episode Transcript

Introduction and Guest Introduction

Rod: Welcome to another episode of the Chris Rod Max Show. Every week, we discuss the latest developments in technology and how they impact companies and users across the world. Today, I'm joined by my co-hosts, Chris and Max, and we have a special guest, Kyle. Kyle, would you like to introduce yourself to our audience?

Kyle: Of course. Hi everyone, I'm Kyle. To briefly introduce myself, I have a diverse background. I've been a product manager and worked in a ministry. I was also a founder with a partial exit. Interestingly, I'm a professional sommelier as well. I've worked in venture capital and now I'm a CVC investor.

Rod: That's quite an impressive and diverse background, Kyle. It's fantastic because in this age of generative AI and assistants, they are often touted as multi-talented. This aligns perfectly with our agenda for today, where we want to discuss the wave of smart assistants that has been around for about a decade now. We'll be talking about Alexa from Amazon, Siri from Apple, Cortana from Microsoft, and so on. We'll explore what has happened with them and where they stand now in the world of ChatGPT and new technologies that differ from this previous wave.

Current Usage of Smart Assistants

Rod: Let's start by asking who in our group has an Alexa or has used an Amazon device with Alexa, such as the Echo. Max?

Max: Not an Echo, but I have a Google Assistant.

Rod: Interesting. Chris, do you have any smart assistants?

Chris: No, I don't have any.

Rod: Kyle, what about you?

Kyle: Not really. I think it could become a never-ending process if you start buying these devices. You might end up putting IoT devices all around your home. I'm trying to educate myself about them and set the right expectations. I know I'd probably encounter initial frustrations, especially with the current state of IoT device development. So, I'd rather not start for now.

Rod: This small sample of our group already shows the challenge that Amazon faces with Alexa. From 2017 to 2023, Amazon has invested or actually lost $25 billion on its Alexa division. There are 500 million devices around the world in 100 million households. Yet, it's not a profit generator.

Business Model Challenges

Rod: The strategy is similar to the Gillette Razorblades model, where they give you the devices for free or very cheap and try to monetize later. But that hasn't really happened. Kyle, how do you see this in software organizations? When does a loss leader become just a loss generator?

Kyle: I'd rather ask you how you would approach this, and then I'll add to that. I think that would be more interesting than me giving a direct suggestion.

Rod: Well, my take is that these smart devices were always very ambitious. They were far ahead of what they were capable of doing. If we think about Siri when it came out in 2011, it was still very basic in terms of interactions. As a result, they haven't really made much progress in almost a decade. Users now think, "This is not really a smart device. It's more like a set of rules where I say, 'Hey, remind me in 10 minutes,' 'Put an alarm,' 'Play this music,' and so on." It's very command-driven. As a result, it becomes very hard to determine what you can do with the device and how you can make a profit from it.

Market Penetration vs. Monetization

Chris: Honestly, when I look at the numbers, I want to flip the perspective. 400 million devices in 100 million homes - that's more than the population of Germany plus a sprinkle of the UK. That's quite a lot of devices. So from a distribution or penetration point of view, it's reached a critical point.

I think the question of monetization is interesting. When I think about a smart device like that, yes, there's hardware you pay a little for, you lose on every device you sell, and your hope is to make some kind of service layer on top. For Alexa and Amazon, that service layer is definitely Amazon Prime. I wonder how much of Amazon Prime revenue is actually attributed to Alexa.

But I think the core problem, and we see this in this round, is that maybe the use cases are just very limited. The reason we don't have these devices is that we don't find them very useful. One article I read put it well: it's a "$10,000 enabled smart timer." Honestly, I have my phone, and it's enough. I don't necessarily need someone turning on my lights or setting a 10-minute timer for my soup.

Kyle: Exactly. I'll echo Chris' point about this. The biggest issue is the opportunity cost. Think about printers or razors - there's a strong need because clients need to have a printer or a razor for specific use cases. But when you think about the use case for Alexa, as Chris said, you can have an alarm on your phone. You can have a music speaker by yourself. There are so many alternatives with a low opportunity cost that people can use. And there are better quality services compared to other alternatives right now.

So why would you expect clients to pay more for less flexible and agile decisions? My take is that yes, the use cases are limited right now, but we have to take a step back and understand the question: why do we expect people to pay? It's not about answering what model we have to prepare for people to start paying. This is more about customer behavior and training process, trying to understand what the most significant pain point is that clients would need and that has the highest opportunity cost, and then come in to start charging. You can't charge for something that people don't need or already have good options for.

Alexa Skills and Platform Strategy

Rod: Indeed, you're right. There was a report in 2021 from Bloomberg that noted Alexa has pretty much three use cases: turning lights on and off, playing music, and setting timers. One of the things Amazon has been trying to do is offer companies an open platform so others can build what they call Alexa skills. Essentially, allowing other smart devices to communicate with Alexa for interactions.

Thinking from a startup perspective, and putting on the VC hat, would you advise a portfolio startup to build on top of these ecosystems like Alexa or Siri? On one side, as Chris mentioned, there are 100 million devices, more than the German population. But at the same time, you're in an environment that is constrained, and you're depending on someone else's platform. Would it make sense to invest in building something on top of Alexa?

Max: I think it really depends on what you're trying to do. If you think about printers and razors, they give you something physical - cartridges and blades. Whereas for Alexa, it's got to be some sort of service or digital product you build on top of it. But so far, the digital products we consume are either books or some sort of visualized content or movies.

If you want to build on top of Alexa, the question becomes: what are the use cases beyond the three we mentioned? I can tell you from personal experience, I own a Google Home with Google Assistant. And I only do those three things, and occasionally, I tell it "good morning," and it runs through my Google Calendar. It becomes like, okay, I'd like to build something, but what is it for? Why am I building this?

I find it quite interesting because it's been 10 years, and nothing has become very popular with the user base yet beyond those three use cases. I also wonder, in those 100 million homes with 400 million devices, what are they really using them for? Are they all using them for the three things we talked about? Or are there some edge use cases that people are using that we don't know about yet?

If I were to build something, and if it's to service those edge cases which I don't know about, then maybe it's worth it. But the distribution is probably very hard, and I don't think I have the cash to build a new smart speaker from the ground up. I also wouldn't want to lock myself into just the Amazon ecosystem - you'd want to build across different devices.

Potential Use Cases and AI Integration

Chris: I want to challenge that because I think we're talking about very obvious use cases, and they're not really good use cases. We talk about the timer, right? But if we were to go beyond that, I mean, there are quite a few use cases you can leverage. Think about a digital companion, like in the movie "Her," or something more applicable, such as audiobooks for your kids when they go to bed. We have these Toniebox devices where you put a figure on it, and it plays an audiobook.

My point is, I think there are use cases that Amazon or any of those voice assistants could explore. And I believe that some are already dabbling in this. We talk about AI today, and I think that actually expands the field of use cases. That's something that people can build upon.

Max: I think if I qualify that a little bit - I was talking about the past 10 years of use cases. They haven't quite come out yet with the Large Language Models (LLMs). However, yes, it will go further.

Alexa did come up with something with Capital One, if you remember. This was probably just in the U.S. Alexa partnered with Capital One, allowing you to ask about your bank account balance or even transfer money. They were working with Capital One, but that quickly got scrapped, or the uptake wasn't as high as expected. So I guess I'm questioning why it wasn't successful. Perhaps there's something to do with Natural Language Processing (NLP) and data. But I do agree with you, Chris. There are things we're thinking about, but they haven't quite manifested from a usage perspective yet.

Transforming User Expectations

Rod: The things we're discussing here revolve around use cases as well as data. One of the challenges that Alexa and other assistants are facing is that after a decade of an installed user base, people have been trained to understand that with Alexa, they can only control their lights or give simple commands. It's perceived as a relatively simple device. Kyle, how would you advise on this transformation? Not only do organizations need to shift to developing these new use cases, such as what Chris was recommending like having Toniebox-type devices or a digital companion, but they also need to retrain their user base that has been conditioned over almost a decade on how to use these devices. Now you're saying, "Actually, this is not how you're supposed to use it. You should explore all these other features." Do you have any mechanisms or ideas in mind for how this can be done?

Kyle: I think my approach here would be to separate it into the client side and the corporate side. Looking at the client side, I think clients have to go through practice, and there are a lot of educational processes that need to happen to get to the level they want and expect right now.

It's important to segment the market because the current market where Alexa is deployed is huge, but the needs are different. The extent of using it for a smart timer right now is probably the majority. But there are still some early adopters, although the tech isn't really early anymore, some real Alexa enthusiasts. I believe they're actually leveraging those devices quite well.

We need to segment that market to study them and understand what exactly these people need and what needs haven't been figured out yet. Maybe we haven't explored the kids' market, the pet market, or the elderly market - they all can have different uses and functionalities in the future. We've been hearing a lot about how Alexa can work with robotics, and those robotics can actually carry out physical instructions. This could potentially bridge the gap between virtual services and physical demands.

Looking at the corporate side, Amazon is a huge company. They've really been working on building that ecosystem with Amazon Music and all different kinds of services. At the same time, they also need to figure out the B-track strategy for all these services. What kind of services do they want to deploy in Alexa first? This way, people would have a pre-concept about what to expect in terms of new features. Rather than just dumping new features or products and telling clients to get used to it, I think having a segmented approach to the market and understanding the B-chess strategy is going to be crucial. These aspects have to be aligned together.

Innovation and Change Management

Rod: Absolutely. On the corporate side, Chris, you've been involved in innovation, especially for industries where their customers have been trained to associate the company and their services with a specific set of premises, and this for decades. Suddenly, these companies also want to have new revenue streams that maybe force or invite the users or customers to see their companies and their products in a different manner. What frameworks or methodologies do they have available to initiate this process of moving away from very specific, targeted use cases and models towards something more innovative and novel?

Chris: Honestly speaking, Rod, my simple answer to this is time. I think we're in a very long transition period right now. There's this new technology, AI, that everyone is trying to leverage. Looking at the stock market, we've seen this bubble and faded expectations. I think if we were to follow Gartner's hype cycle, that's really on point. Now, we're sort of in this trough, and everyone's trying to figure out how they can really commercialize AI. It will take time because people take time to change, and they are very resistant to change. Honestly, you can come up with the smartest framework, but in the end, it's really time that yields everything.

The Apple-Google Dynamic and Opportunities for Startups

Rod: That's an interesting perspective. Another company that's trying to play with time and this shift is Apple. Just in the last couple of weeks, we had this big announcement that Google was considered a monopolist. Part of the second-order consequence of that is that they might have to review their agreement with Apple, where every year they pay roughly $20 billion to Apple to be the default search engine on Apple devices.

This might be a blow for Apple in terms of the revenue they're receiving, but the silver lining is that now that Apple is transitioning towards generative AI services, they might not need to rely on traditional Google search anymore. Maybe they can use Siri as a modernized version that can generate these answers.

Max, now that there's potentially the possibility that we'll be able to choose our default search engine, that maybe we'll be able to choose our default copilot on Apple, that these defaults will no longer be there, but rather we'll have options - do you think this offers a window of opportunity for startups to try to compete with these entrenched players in different areas such as advertisements, search, and so on?

Max: I think in the long term, yes, definitely. In the short term, I think it would take some time for people to get used to different ways of searching. What I find interesting from this perspective is how Apple has positioned themselves for consumers. It's all about being locked into that Apple ecosystem. You have everything Apple, from an Apple Watch to an iPhone to an iPad. I don't have a Mac, but I used to, and it kind of just locks you into that space.

If I think about how Amazon approaches this, they kind of just drop devices from different angles. They're connected, but not in the same cohesive way as the Apple ecosystem. So if I think about how Google will be affected by this ruling, one thing about Apple is they've built a very good ecosystem.

If you're a startup and would like to play in that space, I think yes, as long as you're willing to pay the Apple tax, so to speak. Apple is not just going to sit there and let you run one of the most profitable business models without them doing it themselves. Even Amazon has a search function - every time you search for a product, you go on Amazon. Every time you search for some general question, you go on Google.

So you could see specific searches for specific use cases. It kind of reminds me of the early days of the internet, where you had multiple different search engines - Yahoo, MSN, Google, and so on. To answer your question, I think yes, there's an opportunity if you're willing to pay the Apple tax and navigate that ecosystem.

The Transformation of Search and Advertising

Rod: Thinking about this transformation of searches and how it's happening, there was a story from Wired discussing how the new wave of generative services is also changing not only how we search, but also how advertising, which is the primary revenue stream for Google, is being done. Traditionally, we just type "comfortable mattresses," and then we get the search results with some paid results mixed in.

But now with AI copilots, these results are being summarized. We no longer need to click through a list of best mattresses; instead, we have a summary at the top. Also, in the browser's copilot section, when we ask about mattresses that are comfortable for singles, for example, we can get tailored results.

Now that these markets have been upended, what opportunities do you see for both established players (like agencies and large organizations) and for startups? What do you think they can do to make the best of this moment, where we see that these technology giants are a bit stuck in trying to innovate and figure out where they fit, while also facing regulatory restrictions? This seems to be opening a window of opportunity for those who are more nimble and agile.

Kyle: I know a bit about the marketing agency ecosystem. Typically, a marketing agency would have a client they work with, either a big company like a bank or various other companies. These marketing agencies would work with what we call a publisher, and the publisher owns different distribution channels. One of the distribution channels could be Google

More from the Podcast

Post Image: Oscar Rovira from Mystic AI

Oscar Rovira from Mystic AI

In this captivating episode of AI Product Engineer, Oscar from Mystic AI takes us on a thrilling ride from the world of mechanical engineering to the cutting-edge realm of AI and machine learning. Discover how his passion for autonomous vehicles and artificial intelligence led him to England, where he honed his skills and embraced the power of Python.

Rod Rivera

๐Ÿ‡ฌ๐Ÿ‡ง Chapter

Post Image: E19: AI's Next Leap - OpenAI's O1, Robot Dexterity & Spatial Intelligence

E19: AI's Next Leap - OpenAI's O1, Robot Dexterity & Spatial Intelligence

The Chris Rod Max Show explores OpenAI's O1 model with improved reasoning, advances in AI-powered robotics, and Fei-Fei Li's new spatial intelligence startup. Chris, Rod and Max break down the implications for technology, business and society. Plus, updates on Klarna's AI pivot and SpaceX's milestone private spacewalk.

Chris Wang

AI Innovation and Strategy Expert, CXC Innovation