Hello and welcome to Fireside Chats at the influencer series conducted by the Alchemist Accelerator. These are timely chats with tech leaders about what's on their mind. Let's get into it.
Our guest is a Silicon Valley legend. Amr Awadallah is the co-founder and CTO most famously of Cloudera, a company that allowed you to extract value from big data. Cloudera went public and then got acquired for 5.3 billion dollars. He now has started a new company that he's passionate about called Vectara. We'll talk about Vectara, but Vectara enables corporations to actually build in-house GenAI apps safely. Think about hallucinations. Think about budgets that go overrun. They deal with all of that. Amr is going to tell us more about that. Amr also is a serial entrepreneur. Even before Cloudera, he started a company called Aptiva that got acquired by Yahoo. And at Yahoo, he was a vice president of product intelligence engineering and he also was a vice president at Google of developer relations. He's deeply technical. He was also a PhD graduate from Stanford University with a PhD in EE. Amr is going to talk about a lot of things. So, we're going to talk about why 95% of enterprise AI adoption seems to be failing today. How you as a founder when you're choosing a market can know if you're actually going after a market that is genuinely what you should be pursuing from an AI perspective. We're also going to talk about why Amr started Vectara even though after Cloudera he had no need to do another company from any economic perspective, why he was so passionate about Vectara to start a new company, and also what lessons he's learned about creating enduring fantastic companies and why Vectara is also so critical when you think about the future enterprise stack for having enterprise AI succeed.
So I want to get your take on this MIT report that said that 95% of enterprise AI is failing. What was your reaction?
So first I would say the title of the report is clickbait a bit, like that was intentional. So it's clever of them to pick their title, but again this is MIT and this is a proper research group that did do a very extensive study and survey of many enterprises deploying this technology. And if you really read the report in detail, there is a number of key reasons that they highlighted as the logic behind this high failure rate. The very first reason, and I'll talk about it in more detail later because it relates to what Vectara does, is the lack of technical expertise to actually run these things in production. There's many engineers that know how to build prototypes, but running it in production is a whole different story. So that's a very key reason that they highlight. Another key reason is many of these implementations are coming in because of the board of directors and CEOs saying go do AI. Go do AI is not a strategy, right? You need to like go save money this way, go make more revenue that way, solve this problem or solve that problem. You always have to start with that. When they say go do AI, what you end up with is an experimental budget which we try it out, something comes out and then we throw it away. So that's why it's failing. It's because it's not a real business need. It's a fake let's do AI business need. That's kind of the second reason. And then the third reason is mismatch of expectations. And I blame some big companies without mentioning their names overhyping what AI can do, which is setting the expectations up here. Like AI will come in and do everything in our company. It's like no, AI is not ready for everything. There are a few use cases like document generation, like conversational AI, question answering, customer support that are really prime time right now to deploy it for. But you cannot deploy it for everything. The agentic vision of AI is going to do everything. It will come, but we're not there yet. So because of that mismatched expectations, they launch the project, it does not deliver on what they think it's going to do and they kill it.
And so, if you're a startup founder, how should you triage what's a real opportunity to pursue versus one where you're getting sucked into the hype, but it's not going to be actually a viable long-term opportunity or market to go after?
It's the same advice for any startup you work with, which is make sure you are grounded in a real problem. You're not being distracted by the hype. You are focusing on what problems am I trying to solve. Yes, I know this technology very well. I know how good it is and what it can do. That's great. But now let me find the problem that I can really solve with that technology in the right way.
And how do you know if there's a real problem? Because one of the heuristics that's given in the past is like Steve Blank is a big fan of saying if a customer has a budget for a need then that's a validated market to go after. And in fact that's like the highest on his pyramid of needs is you know does the customer know they have a problem? Can they or do they have a problem? Do they know they have a problem? And then the highest is do they already have a budget? But you're saying that the budget actually can be illusory. Exactly. So what is the way to know if you're actually going after something that's a real production or not?
So, very very good question. So first, yes, the three W's: why do something at the first place? Do I have a problem? Why now? And why this vendor? Why should I work with you on doing this? These are the kind of three W's you want to focus on. The caution with the budget is that many of these new AI projects, it's fake budgets. It's budgets for doing exploration. And these will turn into projects that will churn. Yes, you will get a startup grow up to 20 million revenue very quickly and then it's all gone next month, right? So a very key role that we as entrepreneurs have to do and also investors investing in us have to do is to triangulate with that customer that this is not somebody who's just selling snake oil just to get something going leveraging the hype. No, they're actually solving a real problem that is delivering for that business in a real way and get the word from the business, right? Because for example, we have many of these banners on 101 here saying, 'Oh, we're going to give you this AI SDR agent that will do all the sales for you and give you leads non-stop day and night.' No, that doesn't work today. It's not there yet. If you don't have a real SDR working with the AI SDR, you're not going to deliver. Now, many of these startups selling that, they will go to the investor and say, 'Yeah, our customers love it. We did 20 million in revenue. We're growing really fast.' Now, if you don't go back and talk to the customer to validate what's going on, then you're going to be fooled by the hype that's taking place. And the entrepreneurs themselves, some of us are to blame for amplifying that effect in the same way that the big companies without mentioning names are also overhyping the value of AI.
Okay. So you have to then actually look at where there's real value, real need to be driven. You know that if you can find the business owner to articulate the pain point.
And you can articulate what the outcome needs to be to solve that pain point.
Exactly. Are you solving it because it's a true pain point that they have that they are trying to improve their efficiency in a certain way and you're really delivering that for them, or are you doing it because the board of directors told them go do some AI right now? So understand the rationale for why they're doing it. It's the three W's that you're saying before.
Exactly. Okay. Why do anything? Why now? And why me? Yes. Those are the three questions.
And if you were to, I know you have your own startup now, so maybe this is going to be the tea. But I'm curious if you were, because you also have such an, you're sort of legendary within Silicon Valley. There's so many things that you've done. If you were going to start a company right now, if there was a startup founder that wanted your advice on which markets to pursue, what would you pursue?
The advice I always give is we don't know. That's the whole game with entrepreneurship is you don't know what's going to work in the future or not. Especially that's doubly so and probably true today with how fast AI is evolving. That said, my advice will be the same: make sure you're picking a problem that you have an affinity with, you're passionate about solving that problem.
Yes. And that's a really important first point. So it's independently what the customer need is. You should operate from an internal compass of something you care about.
Yeah. Because if you don't have that, as you know very well, there's so many ups and downs and punches and you're going to fall on the floor 100 times. The only thing that's going to keep you having the persistence and the resilience to stand back up is your love and affinity with the problem.
So that's number one in my book. If you have that then you are now on the right path at least.
And did you have that at Cloudera?
Okay. And did you ever have an experience where you didn't have that?
So you've always had the passion. I'm a very passionate guy. So I know I'm curious to persevere. Yes. Now maybe for my first startup. So actually your first startup you did, the one that you sold to Yahoo. What I want to key in on is a lot of founders, you know, they're in the business of selling visions and they can also be self-delusional. They can sell a vision that they actually don't really believe in, but they're saying that they believe in it because they need to rationalize it for whatever. So I'm curious how to really know if you really care.
Yeah. So you're right. Unfortunately, there are some good scam artists that can portray such a vision and there's some famous examples in Silicon Valley.
Yeah, there's extreme examples like Elizabeth Holmes or Sam Bankman-Fried. But I think the more pernicious ones are actually less explicit. It's really when a founder is pursuing something because they think other people are going to, you know, it's like a hot space to go after or a hot market to pursue or they can get into an accelerator because of it, but they actually don't really.
That's your job as an investor. Of course, that's our job as an investor, but I feel like it's an even more of a disservice for the founder themselves to be too skilled because they can trick me and they can trick investors and trick themselves. Why trick yourself? Why do that to yourself?
So, when do you know, I guess how did you know, how do you know when it is something that's an authentic passion to pursue?
It's how do you know when you love somebody? When you see a girl and you really want to be with that person, it's because you just relentlessly think about that. There's something like define love, right? Like something like your neurons fire up. And that's what you've gone through your different paths, but that's it. When you were at Cloudera, you were just relentlessly... I never questioned my love for the problem of big data and solving that problem. Same way right now, I never question my love for solving the agentic deployment problem in organizations. But I'll go with my first startup which I started in 1999, shopping that we sold to Yahoo when we were one year old. And before we made the decision, let's go, I'm not sure actually I love this problem. I've been doing this for 9 months right now. I'm not excited about going to work. That's how you know. You wake up and you're not excited about it. I'm being forced to do this. It's not fun. So there's many signals. My definition is love. Do you have passion and love for the thing that you're working on? And there's many studies from Stanford and other organizations that show that the number one attribute that defines successful entrepreneurs is grit. And I love this word grit. It's such an awesome word because it captures two significant concepts together: perseverance. Like you're going to be punched down 100 times and you're going to persevere. You're not going to give up. You're going to keep going. But you're only going to get that perseverance if you have passion. Right? Because if you don't have the passion for that thing that you want to persevere around, you're going to give up and you're going to fail. I assure you. And one of the hardest jobs that we have to do as investors, I do some angel investing every now and then myself, is having that conversation with the entrepreneur to see that true light in their eyes that is true love versus you're faking it, dude. I know you're faking it.
So the first thing is fall in love with the problem.
In the AI space, a problem.
The problem, not the AI, not the technology. You actually, that's a really good nuance. So it's not getting in love with what you're seeing on the agentic side or the autonomous side. It's a problem, a need that you want to solve.
Exactly. Okay. And then what's step? Is there another step after that?
Of course team. Team comes right after that. If you don't have the right team around you and hire first your co-founders, be very, like 70% of the companies I think fail because of founder dynamics.
Yeah. Surprisingly OpenAI survived those founder splits. I know. I think it's a question if it's survived. I mean there's been a lot of children that have come out of OpenAI that we'll see how the long term plays out. But I cannot stress that enough.
And what did you learn? You were with Mike Olson, is that right?
Yes, amazing co-founder. I love Michael. Mike is a fantastic soul. Jeff Hammerbacher as well.
But how did, so can you share any unintuitive insights to know when you have a great co-founder or when you're...
It's the same when you have a great spouse. It's the same analogies. What it means is you will be able to have significant debates and fights with them, argue about things relentlessly and even point to each other's mistakes and so on, but never be upset at each other.
Yes, I think that is what I've unified in the love for the mission that we're after.
So you all can agree, you all know who owns whatever parts of the business and what we're good at and you all also complete each other. You complete each other. You can also have strong disagreement so there's honesty. You're not holding anything back because I think that's one of the hallmarks. The first moment you feel like you want to conceal something from your co-founder, that for me is the sign that you're going to... you have to have radical candor.
You have to have radical candor but then also be aligned on the vision or the purpose. And also do you think it's important to be aligned on the CEO and who makes the final decision or does... maybe not. I don't want to put words into your mouth but so that's okay.
