What It Actually Takes to Deploy a Voice Agent to a Fortune 500

1. 0:00 – 0:05

   Intro

   1. HTHarj Taggar0:00

      [upbeat music]

2. 0:05 – 1:16

   Coval’s mission: simulation + observability for voice agents at massive scale

   1. HTHarj Taggar0:05

      Today I'm thrilled to be joined by Brooke Hopkins, the founder and CEO of Coval. Coval is a simulation and evaluation platform for voice agents. They work with customers like Perplexity and Deepgram to monitor and evaluate tens of millions of customer calls per month. Today they're announcing a $28.2 million Series A round led by Norwest with participation from Base10, MAAC Ventures, and YC. Thanks so much for being here, Brooke.

   2. BHBrooke Hopkins0:31

      Thanks so much for having me. Super excited.

   3. HTHarj Taggar0:33

      Could you just give us, like, a quick recap of Coval and kind of what exactly does Coval do?

   4. BHBrooke Hopkins0:37

      Yeah. Coval is a simulation and observability platform for voice agents, so we help you to scale your voice agents over millions of conversations so that you don't have to test your voice agents with real customers in production, and then also when you deploy your agents to production, you know what's happening out there in the wild. And so my background is from Waymo. I led their evaluation infrastructure team at Waymo, and my team was responsible for all of our developer tools for launching and running simulations. And now we're taking all those learnings from robotics, which is actually surprisingly similar to voice agents, and how do you make sure that an agent is getting from point A to point B and all the possible paths in between. We're testing and making sure that those work before production and then once it's live.

3. 1:16 – 2:37

   Why voice agents are the breakout ‘autonomous agent’ interface

   1. HTHarj Taggar1:16

      Yeah, your unique, like, background and how good of a fit working on simulation at Waymo turned out to be for what, um, deploying voice agents in production was one of the big reasons we funded you. Um, I definitely wanna come back to that, but let's start with just, um, voice agents in general. It seems like voice has really taken off as the, um, killer UI and the killer app for AI. Um, why is that? What do you think is going on?

   2. BHBrooke Hopkins1:38

      Yeah. I think what's so exciting about voice is, A, it's the first productionized use case for autonomous agents. It was the first use case where autonomous agents were acting on behalf of users, on behalf of companies, doing things autonomously, uh, in order to get to some objective. But then on top of that, I think voice is going to be the interface for AI. It's the way that you-- It's the most natural way to interface with a text box or interface with a headless agent, and it also meets people where they are in all sorts of different environments. So I think we're seeing voice agents particularly successful with logistics, with healthcare, with all these places where previously software systems weren't as present because voice allows you to have n- like, a single side opting o- into, uh, automating a certain behavior. So, for example, a small doctor's office can still have someone on the phone, and then a, a larger enterprise can then go in and r- automate some process that they're doing in order to talk to thousands of disparate, uh, customers.

4. 2:37 – 4:30

   Enterprise adoption: leveraging existing call infrastructure and expanding beyond support

   1. HTHarj Taggar2:37

      Cool. Um, you have particularly, like, front row seat unique insight into how these agents are being used by enterprises, like in, in production, in real use cases. Um, what's changed over the last 12 months in the enterprise world? Like, it seems like enterprises are especially quick to adopt voice agents, and so I'm just curious, like, has some- what-- has anything made it easier for them to deploy these things, or has anything changed about their attitudes or belief in the technology?

   2. BHBrooke Hopkins3:01

      We're seeing enterprises deploy voice agents at massive scale and more rapidly than any other type of agent because there's a lot more infrastructure already in place for voice. So, for example, standard operating procedures for customer service. You have IVR trees or call flows that already exist. And so the leap from a call to an autonomous agent isn't quite as vast as, say, financial services agent that's, you know, making financial decisions on your behalf. And then what we're seeing with enterprises, and particularly the enterprises that we work with, is that they might start with customer support flows-

   3. HTHarj Taggar3:33

      Yeah

   4. BHBrooke Hopkins3:33

      ... and then move into-- realize there are all these other places in their enterprise that could benefit from autonomous voice systems. So, for example, a concierge to help their users discover more products or meeting them where they are instead of having to go through an application or driving usage, driving adoption, um, automating logistics and back office work, all sorts of things that maybe they wouldn't have otherwise focused on, but now that they have all the voice infrastructure in place, they're able to identify all these different areas.

   5. HTHarj Taggar4:00

      Yeah.

   6. BHBrooke Hopkins4:01

      And so I think we're going to see something very similar to what happened with web and mobile, will happen with voice, where people start by just putting, like, a piece of paper on the web, and that's HTML, and then you create web experiences. Or people start by just putting a website on a phone and then realize that you can create mobile apps. And we're gonna see the same thing happen with voice, where people start with things that are already happening on a phone, like customer support or, um, logistics, and they're going to branch into all of these novel voice experiences that are much more AI native.

5. 4:30 – 5:29

   The ‘positive vision’ for voice: better experiences, not just labor replacement

   1. HTHarj Taggar4:30

      Yeah. That's a really interesting way of looking at it because I think there's a lot of focus obviously on voice agents as replacing customer support and replacing human labor, not as much focus on the o- the, the more positive vision maybe of, hey, like, actually you can ac- use AI to, um, sell more of your product or help people find more of what they actually want, um, which will just, like, increase the need for goods and products and services to meet that demand.

   2. BHBrooke Hopkins4:54

      Totally. I think this is going to be true, like [clears throat] even take airlines as an example. I think in two years or one year from now, it's going to be unacceptable to call an airline and be on hold for 20 minutes. But imagine if instead you could call your airline on your way to the airport and say, like, "Are there any flights that are, you know, 20 minutes earlier? I made it to the airport earlier."

   3. HTHarj Taggar5:14

      Yep.

   4. BHBrooke Hopkins5:14

      And previously, you would have to go on the web portal, you have to go to the check-in desk, and there's going to be a lot easier ways to interface with really complex information and the complex set of decision-making that is just, um, distilled into, "This is what I'm trying to get to. How can I get there?"

6. 5:29 – 6:47

   What Coval provides: missing infrastructure for scalable voice apps

   1. HTHarj Taggar5:29

      Can you maybe talk us through just how does the infrastructure Coval provides your customers help them kind of go from simple use cases where there's sort of essentially just, you know, there's well-documented maybe support flows to things that are a little bit, like, harder or more interesting to do?

   2. BHBrooke Hopkins5:44

      Yeah. We're trying to provide the infrastructure that allows anyone to scale voice applications. So think about what happened with web infrastructure. It used to be very hard to build a distributed web application. A DDoS attack could bring down any web app, um, you know, high spikes. I even remember websites going down, um, not that long ago, and now with serverless and all sorts of web infrastructure, it's really easy to build a s- really scalable site. And we're missing a lot of this infrastructure for voice. Today, it's still really hard to build voice applications, and we're trying to make it so that any enterprise can build a voice application that scales to millions of users, and they understand what's happening in all those conversations, where are things going wrong, where are the compliance risk problems, but also all of this product information. Like you have customers at your fingertips, and enterprises are always trying to hear from customers, understand the customer journey, opportunities for upsell, opportunities for product adoption. And so we can provide a lens into all of those customer interactions. What Coval is trying to do is how you provide that infrastructure so that any enterprise out there can build a voice application.

7. 6:47 – 8:54

   How voice agents fail: from hallucinated audio to workflow mistakes

   1. HTHarj Taggar6:47

      So when you're working with your customers to help them deploy these voice agents, what are some of the things that voice agents are kind of naturally good at doing out of the box, and where are the areas that they're more brittle and you have to provide more infrastructure?

   2. BHBrooke Hopkins6:58

      Yeah, definitely. I think the interesting thing about voice agents versus, say, like a customer support agent is that voice agents fail in totally different ways than a customer support agent might fail. So a customer support agent might struggle to keep up with new product changes. So products are constantly evolving in enterprises, and so maybe a product was deprecated or changed or the policy changed, and that is something that agents are exceptional at. The moment you change that policy, then it deploys to tens of millions of conversations. The things that agents are-- struggle more with is that they might trip up in more egregious ways that an human agent might not. So for example, it might say just the completely wrong thing, or it might, uh, have a vocal hallucination. Famously, voice agents will accidentally scream, or they'll start to whisper, or they'll change voices halfway through. So the voice agents never, uh, cease to be really funny, but at the same time, if you're on the other side of that, that is, you know... A customer support agent would never accidentally scream during the conversation. [chuckles] And so there's all these things that maybe you didn't have QA for before, and now you have to. And I think also QA used to be somewhat optional as like a nice to have, but ultimately you kind of just assume that things are roughly working. But with voice agents, that's not true because you have the capability for them to access all sorts of systems. There's all sorts of compliance problems and security controls that you have to have in place. And then on top of that, we now have the ability to understand wha- vast amounts of data that's unstructured in a way that we never could before. So even if you continue to have human agents handling all these calls, there's still so much potential for being able to do things with that information. And maybe more specifically to your question is things that people test for are threefold. Did the agent do what it was supposed to? Did it take the right steps to get there, like workflows, the right-- call the right tool calls? And then also audio quality things like, um, background noise, interruptions, latency, all of these things that make voices sound really natural.

8. 8:54 – 10:53

   Building an enterprise evaluation strategy (and what people mis-measure)

   1. HTHarj Taggar8:54

      I was thinking, how do you set a customer up with that? Like, do you go in and tell them, "Here's all the criteria we're gonna evaluate, um, against," and sort of bring that expertise, or do you work with them to figure out what are the things you should be evaluating on and, um, and how do you help them build trust in your evaluations?

   2. BHBrooke Hopkins9:10

      Yeah. For all of our enterprise customers, we work with them to create an evaluation strategy. I think this is one of the things that self-driving car companies did really well, is how do you take a system and make sure that there are processes in place that make you get better and better over time and create this flywheel. Having seen this across lots of different self-driving car companies as well as hundreds of different voice systems, we work with enterprises to help them set up a scalable evaluation system. But then I think we really think about the Coval platform as how do we scale our expertise now that we've-- we're very early in voice. I think when we started in voice, it was really just a few YC companies that were building in voice AI, and it seems kind of niche. And now every enterprise, every Fortune 500 out there, we're talking to figuring out how we can help them scale their voice AI.

   3. HTHarj Taggar9:56

      And now that you've done this with real enterprises a bunch of times, um, what are some of the most common-- When you're sort of going setting up your infrastructure for the first time, what are some of the biggest misconceptions your customers have? Like, are there things they think they overvalue or they think, um, we really need to evaluate against this, um, uh, that turns out not to be so important? Or just what are the things you find yourself repeating a lot?

   4. BHBrooke Hopkins10:17

      I think people think word error rate or transcription is more important than it actually is because really you can have a full conversation and miss lots of words and still understand what that person is saying. Like, if you've ever been on a Zoom call, you know [chuckles] that's, that's true. But really it's about how-- like did you understand the intent of the conversation, and did you get to the final step? The other part that's really hard for agents is maybe that's different than what was hard for human agents. So for example, um, starting over a conversation or saying all the information to begin with in the conversation can confuse agents because they might have a multi-step workflow.

9. 10:53 – 12:32

   Next unlocks: controllability in real-time voice models and better model integration

   1. HTHarj Taggar10:53

      From where we are now, like the next sort of step up in, um, performance and capability from like the voice agents in particular, what, what's gonna be the big unlock there, do you think? Is, is it better models, less, um, lower latency, um, better text to speech, speech to text? Like what, what do you think is gonna happen?

   2. BHBrooke Hopkins11:12

      Yeah, I think controllability for real-time models. So today the way most voice AI applications work is with cascading architecture. So you have a speech to text which transcribes the conversation in LLM and then a text to speech which then says that thing out loud. This is actually what makes, uh, voice so similar to self-driving cars, is that you have perception, what's happening in the world around me; reasoning, what should-- planning, h- what should I do next; and then controls, how should I actually take that action, which maps very similarly to transcription, reasoning, and voice. And so you have this loop where you're kind of perceiving what's happening in the world around you, reasoning about what to do next, and then actually taking that action. And so autonomous systems are all very similar in this way, where they have this reasoning loop. And I think a lot of the advancements that came from self-driving were being able to bridge these different models by passing in different embeddings and different contexts, while also being able to provide expertise and focus for different models. And so in self-driving, there was actually a similar pattern of condensing and then specializing models, and I think we're seeing that with voice AI, where a single model is not going to solve everything, but also keeping all of the models separate is not going to solve everything. So we have to find some way to be able to make each of the steps be able to share embeddings, share context, while also making sure that they're specialized and good at the part that they're trying to do.

10. 12:32 – 14:50

    Waymo to Coval: datasets, developer tools, and why edge cases matter

    1. HTHarj Taggar12:32

       You mentioned a couple of times the similarities to self-driving cars, and obviously, um, you worked at Waymo for a while. Can you tell us a little bit about the work that you did at Waymo? Um, so when was that, and when did you first start seeing the parallels between sort of the simulation work there to the infrastructure you're building today?

    2. BHBrooke Hopkins12:49

       Yeah. So when I started at Waymo, I was building out our dataset infrastructure. So how do you-- One of the really important things about self-driving cars that was pretty different from other ML systems at the time was that you cared a lot about specific examples and less about kind of general performance across the dataset as a whole. Because y- like, for example, a kid in a Halloween costume is a super important example, and it doesn't matter if that's only one in a hundred million miles.

    3. HTHarj Taggar13:13

       Yeah.

    4. BHBrooke Hopkins13:13

       That made dataset creation very different, where it was how do you find these very specific examples so that you can make sure that your dataset is representative of what you're trying to, uh, improve in your systems. And then I went on to lead a t- our team that was responsible for all of our developer tools. So how do you take a dataset, combine it with some configuration, and then be able to run that on distributed compute? And when I left Waymo, I realized that a lot of what was happening in AI sounded very similar to the problems I had talked about and solved a lot at Waymo-

    5. HTHarj Taggar13:43

       Mm

    6. BHBrooke Hopkins13:44

       ... but in a pretty different context. And so I started off building in evals, but realized that was kind of a solved problem, and also people didn't care as much early on about the accuracy of their models for, say, like summarizing a small part of the feature in their web app. But then we talked to our first voice customer, and they were like, "This is a huge problem. It sucks all of our time. You don't have a single line of software written, but we're willing to just pay you to help us figure out this problem." And I think YC talks about this a lot is that pull of people will pay you just to figure out how to solve the problem. It shows that there's enough product pull, that there's market pull, and that there isn't really a good solution out there yet. And then that's when we realized that was actually very similar to self-driving, and that has only proven more true over time. You know, I thought that it was more of a theoretical analogy, but to this day, a lot of our team comes from self-driving, and we're constantly talking about, you know, simulation problems, realism, determinism, how can you, um, be able to simulate, like, only certain parts of the conversation, but maybe I wanna keep the background noise from a real conversation and only sub in synthetic audio.

11. 14:50 – 16:38

    Finding the wedge: from generic evals to voice via intense customer pull

    1. HTHarj Taggar14:50

       Um, I think that-- So that, that part of the story, like you, you're starting out with a, a broader idea and then, um, refining it down to voice agents in particular, um, I think is, like, very, very interesting. Tell us a little bit more about some of the details there. Like, when you applied to YC, kind of do you remember, like, what was the idea then, and then what were the steps to actually sort of change? Like, did you get customers for the initial idea, and then when did you decide to kind of really focus in on voice as the wedge, and, and why?

    2. BHBrooke Hopkins15:16

       Yeah. Well, I think credit goes to you because-

    3. HTHarj Taggar15:19

       Yeah

    4. BHBrooke Hopkins15:19

       ... you, uh, saw the potential in this idea. I think I was definitely in the beginning at kind of figuring out how, like how can I build an evals platform, but there were already a lot of people in the space. And I knew that there was something there with the expertise, having built this at Waymo, and there aren't that many people in self-driving that have really dug deep into the developer tools for self-driving. Um, and so I think kudos to you for seeing that early on. And really talking to customers was one of-- I mean, this is the, the tagline-

    5. HTHarj Taggar15:47

       Yeah

    6. BHBrooke Hopkins15:47

       ... of YC. [chuckles]

    7. HTHarj Taggar15:47

       It's a cla- classic YC staple advice. Talk to customers.

    8. BHBrooke Hopkins15:49

       But I can't express enough, like, how much we still talk about that on our team today, but, like, throughout the entire journey of the company is talking to customers, trying to understand what are they struggling with, and not necessarily taking what they say verbatim, but trying to understand, like, how is that a window into their world? How is that a window into the problems that they're facing? Um, and so that's how we ended up talking to the customer who ultimately be- ended up being our first customer, seeing that problem happen in, in action. And so Fonely actually was our first customer. They're also a YC company from the same batch. They also just raised their Series A recently, and they, uh, share an office with us now. Uh-

    9. HTHarj Taggar16:26

       Nice

    10. BHBrooke Hopkins16:26

        ... or sh- they're in the same office building. And I think it was really cool hearing them express how their problem was shaping up and really working them-- with them as design partners from day one till now.

12. 16:38 – 18:20

    Recognizing real PMF: procurement momentum and the cost of manual testing

    1. HTHarj Taggar16:38

       But I remember you had, like you did get c- for the initial idea, the broader sort of eval, um, um, you did have early customers for that, right?

    2. BHBrooke Hopkins16:45

       We n- never made any money from it. We had some tire kickers. [chuckles]

    3. HTHarj Taggar16:48

       Okay. Yeah, so that's like, how does it feel as a founder when you've got people, like, interested and kicking tires versus are, like, really pulling the product out?

    4. BHBrooke Hopkins16:56

       It's been really cool to watch Coval grow because now I feel what product market fit-

    5. HTHarj Taggar17:00

       Yeah

    6. BHBrooke Hopkins17:00

       ... feels like, where people are chasing us down to book meetings to push things through procurement. They're putting us on their back and just carrying us through the procurement process. When people are willing to knock down barriers for you, that is definitely a really great sign. But even early on, I think that people were just really willing to use a clearly immature product and early idea in order to solve this problem because it was so painful, and that's because it was really expensive. Calling your voice agent 10 times over and over is really expensive. If you want to have ten six-minute conversations, that's an hour of your time. And then on top of that, it's really frustrating. It's not very accurate. Testing a system 10 times is not enough to then launch it to millions of conversations. Um, and it's important, or it's critical that these conversations go well. And so that was the perfect storm for finding a pain point that people were really willing to, um, dig deep on. And then I think another interesting part of our company journey has been focusing on enterprise. So since the very beginning, we focused on the enterprise use case, even when it was at the cost of maybe smaller startups or, um, wider adoption. And that allowed us to get really deep and understand those customer pain points, so they were able to build a scalable enterprise solution that allowed our customers to scale to mil- tens of millions of calls.

13. 18:20 – 22:29

    Why focus on enterprise early: scale problems, roadmap clarity, and founder-market fit

    1. HTHarj Taggar18:20

       And why did you do that? 'Cause there's definitely been some debate internally here at YC around should you go to enterprises or should you, like, focus on startups. Um, how did you think through that?

    2. BHBrooke Hopkins18:28

       You definitely don't want only enterprise customers because then you're not learning from the fastest moving AI companies. And at the same time, moving up to enterprise allows you to have a lot more focus and consistency. I think the thing I really like about working with enterprise customers is that they have longer term roadmaps and longer term visions of where they're going, and that allows you to re- work really collaboratively. But I would say, like, not moving too fast upmarket, but also I think moving up market can be a really, uh, strong product signal. With our product, it's interesting because it's actually a problem that you really start to face once you scale. Certainly, small companies have this problem of, "How do I test my voice agent?" But getting hundreds of engineers to work together, like we work with Fortune 500s where there's 100 plus people from their, their company in our platform, and so getting those enterprises to allow hundreds of engineers to work together on a system and make it better over time, one, it's a unique point at which you hit this problem of scale, but also it's a really good, uh, fit with my background of I spent my career at Waymo where I was building tools to help hundreds of engineers work together to scale AI systems. And so there's some element of f- founder market fit as well. Like, those were the types of systems I was used to building and knew about building. And so, um, sometimes I think in building a company, you also just have to figure out, like, where are your strengths as a founder and build. I think there's a lot of founders out there that are very open source native or they're, um, you know, very natural fit to startups or a natural fit to, like, different segments. So, for example, if the models become 100 times better, our simulations become way better, um, or our metrics become way better. And no matter if you have AGI or not, um, even let's say humans answering phone calls or running these workflows, you want to know whether or not it's going according to plan, according to how you define your product. And so I think that's a really useful litmus test is if these models become 100 times better, simulations are better, metrics are better, and you still need to be able to scale systems and understand where things are going wrong acro- across tens of millions of conversations. The other question is, how do you stay up to date with whether the things that you... How should you be thinking about agents? How should you be thinking about product within the context of agents and whatever the next thing is? It almost makes you sound like an enterprise when you talk about it like that, but I think even at an early stage startup, you're- we're thinking about this every day. For example, all of our APIs are very agent native, and we aim to make everything that you can do in our platform, you can do via our APIs. But these are things that that would not have been obvious five years ago to do. That would have been actually an anti-pattern. And so I think there's a lot of things that you have to start rethinking and just see, like, what's working for other people and just stay really, uh, creative.

    3. HTHarj Taggar21:08

       Yeah, it's funny. I hadn't thought about it, but yeah, it, it is an enterprise way of thinking about things, which seems odd, but it's totally necessary for founders now because everything change... The, the world moves every three to six months. Like, you can't, you can't just do the old thing of, like, focusing on, like, product and customer. You do have to, today, you do have to think about what might the world look like in six months' time.

    4. BHBrooke Hopkins21:26

       Totally. And you have to hold the thing, have strong opinions loosely held has never been more true of your strong opinion two months ago might completely change. Like, I think six months ago it seemed to me that real time was never going to be, um, something that everyone uses because of the lack of controllability. But now speech-to-speech models, I think, are showing more and more promise of, um, being really controllable and scalable. Other things like security, you know, early stage companies I think didn't think about security nearly as much, and now you have a bunch of open claw agents. Like, how do you manage security for your-

    5. HTHarj Taggar22:00

       Yeah

    6. BHBrooke Hopkins22:00

       ... engineering team of 10? That's crazy. [chuckles]

    7. HTHarj Taggar22:02

       Yeah. We were just having this conversation internally at YC. It's like you want basically, there's just a trade-off where a bunch of us are just using open claw personally, and it's, like, freeing and amazing, but then you try and bring it into a work context, and you've gotta have some trade-off between, like, security and functionality. But once you've kind of tasted the good stuff, you don't really wanna cut. It feels like coming into work and using this thing that's really secure but can't do half of the things that you want. It's quite frustrating.

    8. BHBrooke Hopkins22:26

       You're telling me I can't just put open claw on all of my emails?

    9. HTHarj Taggar22:28

       Yeah. [chuckles]

    10. BHBrooke Hopkins22:29

        Why? [chuckles]

14. 22:29 – 25:35

    A new category: agent testing + observability, and why validation time is growing

    1. HTHarj Taggar22:29

       Basically. Something else I'm curious about is I feel like the whole concept of just evals, um, obviously has been around machine learning for a long time, but they, it has only sort of entered, like, mainstream startup talk recently. How do you think about, like, the market you're in? Do you think of Coval, you're in the, the eval market, or is it broader than that? Like, uh, is it a new market all entirely that you're creating? Like, just what, what's the right terminology to be using?

    2. BHBrooke Hopkins22:54

       Yeah. I think it's really a new paradigm of maybe has, uh, inklings of testing and observability, so something like Datadog and Applied Intuition, but for AI agents. And at the same time, I think we're seeing a, a change in where do developers and operations people spend most of their time. For example, it used to be that you spent a little bit of your time planning, some of, a lot of your time building, and then hopefully a little bit of time testing. Sometimes not testing.

    3. HTHarj Taggar23:22

       Yeah.

    4. BHBrooke Hopkins23:23

       And now that has-

    5. HTHarj Taggar23:24

       I think it's kind of famous amongst engineers to, like, look down on testing.

    6. BHBrooke Hopkins23:27

       Yeah, exactly.

    7. HTHarj Taggar23:27

       It wasn't seen as something that, um, I, I at least in my circles, I think the elite engineers felt like testing was something that they should-

    8. BHBrooke Hopkins23:34

       Yeah. It was like, "Ugh, like we have to."

    9. HTHarj Taggar23:36

       Yeah.

    10. BHBrooke Hopkins23:37

        And now I think that's completely changed, where you spend most of your time planning and in validation, and building basically, and is going towards zero, right? Where, like, it's so easy to build and so fast. At least on our engineering teams, that's where we're spending most of our time is that planning phase, and then how do you roll it out? How do you validate it? How do you make sure that it works continuously at scale over time? I love this analogy of, uh, software engineering is the integral of programming over time.

    11. HTHarj Taggar24:03

        Hmm.

    12. BHBrooke Hopkins24:03

        So it's pretty easy to make something work once, but then to make it work over time is the challenging part.

    13. HTHarj Taggar24:10

        Yeah.

    14. BHBrooke Hopkins24:10

        And I think that's still true even with agents and even with, you know, all of these coding agents is how do you make sure that something is continuing to work over time and it's maintaining, uh, all of its functionality.

    15. HTHarj Taggar24:22

        Yeah. And there's always, like, however good the base models are, um, there's always gonna be, like, customers will always want some comparative advantage. They're always gonna want something slightly better than, like, their competitor has, which is just, like, room for you, right? 'Cause as, as long as you can always add something on top of whatever, like, comes out of the box, then there's always gonna be value.

    16. BHBrooke Hopkins24:41

        Yeah. And I think we also are constantly thinking about how do we scale our expertise and be a tool that not only provides, like, an operations layer, but also scales our expertise on how we think you should scale your voice agent. I think Linear is a really interesting example, where Linear is actually incredibly easy to vibe code, right? Like your own task automation. But I think what Linear does really well is it tells you how to run an engineering team well.

    17. HTHarj Taggar25:05

        Mm-hmm.

    18. BHBrooke Hopkins25:05

        Um, so it provides you all the tools before you even think about them. They're very opinionated about how you should structure, like, sprints, how you should structure teams, and they also do a really good job of designing a product that's really delightful to use, and you can just spend all of your day mana- using it. It's agent native, et cetera. And so I think- It's really more about how do you become an expert in this particular area and be scale expertise of your team, and also by using this product, we're leveling up X

15. 25:35 – 30:45

    Founder journey: solo-founder rationale, YC’s bar, and Coval’s next roadmap

    1. HTHarj Taggar25:35

       Can I shift gears here a little bit? I wanna talk a little bit more about you and your founder journey. Uh, I mean, Waymo is like, must have been such a fascinating place to work. It's like changing the world. What gave you kind of like the confidence, what was your thought process to leave to pursue a startup, uh, and especially pursue a startup as a, um, solo founder?

    2. BHBrooke Hopkins25:54

       Yeah, I knew I wanted to start a company probably since I was little. Um, I think both my parents are entrepreneurs. My grandfather started his own bank.

    3. HTHarj Taggar26:04

       [laughs]

    4. BHBrooke Hopkins26:05

       So I, I definitely had role models there of when I quit Waymo to start a company, my dad was like the most excited for me out of anyone, which I only now realize is a special experience to have parents that are so supportive of that. But I really wanted to start a company because I love the creativity aspect of it. When I started Coval, I actually wasn't intending to necessarily be a solo founder, but I just was really excited about this idea and kind of just kept going until I found a roadblock. It's been really helpful to be a solo founder for Coval specifically because it's a very technical product and also requires like category creation and kind of talking to enterprises and helping them to understand how should they be evaluating agents, how should they be thinking about voice agents. I'm curious on your side, I know that YC generally does not fund solo founders, and there were only six solo founders in their batch. Why make the exception?

    5. HTHarj Taggar26:56

       Yeah, it's pretty... It's definitely rare. Um, first it's like actually it has to be technical. Then I think like as a solo founder, you need to be able to both build and sell, and it's easier to find people who can kind of do the building part and then learn to do the sales part during the batch, like as you, as you did, um, than people who can do the sales part but are gonna have to learn to, to build. So I think that's, um, sort of the first big picture criteria I look at. But then in your case in particular, it's like, it was just like such exceptionally strong founder market fit. And so even by, I mean, it was early days, like you mentioned, like it wasn't completely clear kind of where things would land and what the specific idea would hone in on, but, um, it was just clear that like you had like a very unique set of expertise and skills that would transfer over in interesting ways. So I think combination of being able to build, having like real valuable domain expertise, um, and I just, it just seemed like you were, you were like going to do this either way. I think that's another thing. Like sometimes you see p- solo founders and it's like you just needed like an extra level of like grit and determination to do the thing, and you can just get a sense for when you're interviewing or you're talking to people, like some people are like, huh, like I feel like if we don't fund them, they're just kinda like gonna go back and-

    6. BHBrooke Hopkins28:07

       Yeah

    7. HTHarj Taggar28:07

       ... go back to school or go back to their job, or they g- they, they'll, they're gonna hang around and not get much done. But with you, it definitely felt like this is gonna happen either way, so we just like would like to be a part of it.

    8. BHBrooke Hopkins28:18

       That's awesome. Well, I think that was an accurate read, 'cause definitely at the time-

    9. HTHarj Taggar28:21

       Cool

    10. BHBrooke Hopkins28:21

        ... I was like, "I'm gonna build this either way." And so-

    11. HTHarj Taggar28:23

        Great. [laughs]

    12. BHBrooke Hopkins28:23

        ... this is going, the ship is, uh, the train is leaving the station.

    13. HTHarj Taggar28:27

        Yeah. That's basically, that's actually the t- that's the exact phrase we use internally, is like if it feels like the train is leaving the station, um, then you just wanna work with those people. And so sometimes it's like they're single founder, sometimes they're like still in college. We don't really, despite like the, um, perception on Twitter maybe, we don't actually like funding people who are still in college and are gonna drop out. But every now and again, you meet people who are just like, it's happening either way. And so, um, yeah, you look for those people, then you back them. I'm sure another big part of this round is gonna be investing in the product and the product roadmap. What, what are some of the, um, things you're personally most excited about that's on the Coval product roadmap for the next year?

    14. BHBrooke Hopkins29:02

        Yeah. I mean, I think as everything goes more agentic, how do you build agentic evals? So that's the big question, is doing it well. It's easy to build a script that kind of goes in a loop and, uh, iterates on your agent. To actually build a self-improving agent or systems that help do reinforcement learning or prompt optimization or just optimization of your system, that's really hard. And so we're looking at how do we build all of these systems across the board to help you understand exactly where things are going wrong, get really high signal out of it, and then do something about that. And then I think the biggest next stage for us is just how do we take what's worked for many Fortune 500 enterprises, for some of the fastest-growing startups and hyperscalers in AI, and replicate that to every Fortune 500 that's building voice agents.

    15. HTHarj Taggar29:47

        Cool. All right. Well, I'm excited to see it happen. Um, maybe just to close out, uh, for everyone who are watching who want to start a company themselves one day, any sort of closing bits of advice for them, um, as they think about that?

    16. BHBrooke Hopkins30:00

        Yeah. I think to be a s- just being really obsessive is one of the most useful things, and being really curious of, you know, for every, you know, across the board, there's no way to really... I think I thought that there was a way to become ready to be a startup founder, but really you just have to jump into it and be really curious and work as hard as you possibly can. And, uh, you can make up for a lot of shortcomings by w- just working harder and moving faster than everyone else.

    17. HTHarj Taggar30:27

        That is shockingly true. [laughs] Cool. I think that's all we have time for today. Congratulations again on, uh, raising the Series A, Brooke, and we're excited to see where Coval goes from here.

    18. BHBrooke Hopkins30:36

        Thanks so much, Harj, for all your support. [outro music]

Episode duration: 30:46