Loading summary
Transcript49 lines
- [00:01]
A
Going viral on Twitter is more art than science. Anyway, there is some amount of luck involved.
- [00:08]
B
Welcome back to another episode of Builders. As always, this show is brought to you by Frontlines IO, Silicon Valley's leading B2B podcast production studio. If you're bringing technology to market and want to learn from your peers, we have a library of more than 1200 interviews with Venture backed founders and marketers. Where they talk, all things go to market.
- [00:24]
C
Of course, if you want to launch
- [00:25]
B
your own podcast, we offer podcasts as a service to more than 80 tech startups. The idea there is very simple. You show up and host and we do everything else. Now with all that said, let's jump into today's episode.
- [00:42]
C
Today we're speaking with Abhishek, co founder and co CEO of Yotori Abhishek. Welcome to the show.
- [00:47]
A
Thanks for having me. Of course.
- [00:49]
C
Let's go ahead and jump right in. What does your technology do?
- [00:51]
A
We're building web agents so agents that can take actions and complete tasks on the web. We're pushing both on the capabilities front, so training our own in house models that can complete tasks on the web and on browser based workflows as well as on product experiences. So productizing this technology in different ways for developers, for prosumers and so on
- [01:11]
C
in your life, how are you using the technology today? Maybe talk us through some of the different use cases that you're using in your own life.
- [01:17]
A
Yeah, for sure. So Scouts, which was our first product, it is more prosumer focused there. The value proposition is that these are agents that can monitor the web for anything you care about. So anything from like restaurant reservations, tennis court bookings to keeping tabs of like geopolitical news or like tech news or new open source model releases. It's a fairly horizontal product. You can use it to monitor a whole bunch of things. I personally use it for like I have a lot of news related scouts set up. I have several that are restaurant reservations related. I have several that are monitoring for sales on various like clothing websites.
- [01:52]
C
Yeah, I need one for House of Prime Rib. I have, I always like remember I want to go there and then it's like you know, the day before and it's hard to get a reservation there with court notes, artists.
- [02:00]
A
That's actually one of the examples on Homepage. How's a friendly rule for two? You should check it out.
- [02:08]
C
All right, well you got another consumer signed up. Talk to us about the go to market and how you're approaching GTM here.
- [02:14]
A
Yeah, so when we started we released Scouts made last year. We announced Scouts in June last year. It was in beta for several months before we made it generally available. The initial thinking was that there's a lot of noise around agents, right? Like the entire ecosystem is building agents in the broadest possible sense of the word. And there's a lot of players, a lot of products out there that are basically saying that here's an agent that can do anything for you, but then you try it the first time and it fails miserably. And you're like, no, this can't do anything for me. It can't even do this one simple basic testing query, right? So we wanted to be very mindful of, okay, what is the value proposition? How do we articulate and communicate that clearly to our end users? Which is why with Scouts, it was this specific articulation that these are agents that can monitor the web for anything. You care about monitoring. If you think of it in like, the broader sense of the word, a lot of tasks in people's background that are going on in people's lives in the background have a monitoring nature to them. Like, for example, I know I have a trip coming with my parents six months from now. It's kind of mulling in the background in my head. I know I have to do the research to figure out, like, where we're going, hotels, flights, et cetera. In some sense, that is kind of like a monitoring query that you can enter that query. Scouts will keep aggregating information relevant that might be relevant to you. You can give it some broad stroke specifications of what you're looking for, stuff like that. It can be as broad as this or as specific as, like, get me a table for two at Hazel Prime RIP for this weekend, right? Both have this background always on monitoring nature. So that was the value proposition that we were going with for Scouts. The other part of monitoring is that it's read only, right? So the stakes for mistakes is like, lower than it is for write tasks. Like, it's not the initial version of. It was not meant to buy book reserve things on your behalf. It was meant to just be read, only be monitoring in the background aggregate useful things for you. So kind of like a noise filter or like a noise cancellation for the web to surface relevant things for you. So that articulation landed quite well. We did a good job in communicating that clearly. A bunch of our users by design a horizontal product. So we saw varied usage across the board, across verticals. We ended up signing a couple of enterprise contracts as well that would bottom up in that individuals at companies tried it for generating sales leads and they were like, oh wow, this is really good and I need to make this available to my entire team. And that's how some of those contracts happen. So that was what the GTM motion initially looked like.
- [04:46]
C
And was that surprising to you that Enterprise came calling or was that expected? And was that a plan to go after Enterprise?
- [04:52]
A
It was. I think the degree to which it happened was a little unexpected. Like all along, when we think of the vision for our prosumer product, it is to be like a chief of staff, an AI force chief of staff that can do web based tasks for you. So we knew all along that this is going to span people's personal lives as well as Bork, and the line is anyway blurry with tools like ChatGPT, Claude, et cetera. Right. But the extent to which individuals trying the product led to them telling their friends, colleagues, teams about it and it leading to some amount of enterprise interest was a little unexpected. We didn't think this initial version would resonate as much.
- [05:31]
C
This show is brought to you by Frontlines Media, a podcast production studio that helps B2B founders launch, manage and grow their own podcast. Now, if you're a founder, you may be thinking, I don't have time to host a podcast, I've got a company to build. Well, that's exactly what we built our service to do. You show up and host and we handle literally everything else. To set up a call to discuss launching your own podcast, visit Frontlines I.O. podcast. Now back to today's episode when you launched Scouts Take us Behind the scenes. What did that launch look like?
- [06:00]
A
Yeah, we basically shot a video of talking to the camera, introducing the product, being very clear in the articulation of what it can and cannot do. We posted it on Twitter, it went viral on Twitter. A bunch of like, prominent individuals picked up the announcement. We had Guillermo from Vercel who reposted it. Soleo reposted it, Scott Belsky reposted it. Yeah. So it got decent traction on Twitter. That drove a bunch of people to our waitlist. It was released in a beta capacity, so we didn't quite have the capacity to like onboard everyone on day one. But over time, we sort of built that muscle, built the underlying tech to be able to do so. Because we were seeing good retention metrics over the next couple of months, we pulled up. So this was initially free. This wasn't even a paid product issue. It was more like an experiment just to see if this resonates or not. But because we saw good metrics, we pulled up timeline on when we wanted to implement some sort of a pricing model. We went with like, a flat monthly subscription. So we built that, and then we actually did the timeline to our GA announcement. And now, like, if you go to the website today, it is generally available, but that's what that looked like.
- [07:01]
C
And are there any numbers that you're sharing with how many people signed up?
- [07:03]
A
So in the week of the announcement back in June, we had about 20 or 30,000 people join the waitlist. Yeah. So that was pretty exciting. Then one retention continues to be at, like, upwards of 80%, which is pretty healthy for a consumer to resume product.
- [07:17]
C
And when you reflect on that launch, is there anything that you would have done differently this time around or the next time around?
- [07:22]
A
I think things are changing very, very quickly, to be honest, like, even relative to back when we announced it last June. So our announcement was primarily on Twitter. That's where we posted the video and it went viral. Twitter feels way more saturated today than it did even, like, six months back. Like, every single day there's like multiple new announcements happening. And how do you like having your post go viral? Seems like table sticks. But even then it might not break through to an end user, to an audience. Right? Yeah. So I don't know what I would do different. And going viral on Twitter is more art than science anyway. There is some amount of luck involved. Yeah. So I don't know if I have an answer to your question, but the platform feels pretty saturated on that launch video.
- [08:02]
C
Like, roughly what did you spend on that? You know, did you go like. Like, I had a friend the other day from. They're like series B now, but he spent like 100k on his, like, launch video. And he was like, worth it. Like, worth every dollar. I have other friends who spent like, 5K. Like, how much did you roughly spend? And, like, was that good value, do you think? In the end?
- [08:19]
A
I thought it was pretty good value. We spent about 25k on it. Yeah. So it wasn't, I think, in the bigger seam. I think it was in the lower range of the spectrum, but yeah.
- [08:28]
C
And then what do you have planned for 2026? How are you going to, you know, maintain this growth and keep growing?
- [08:33]
A
Yeah. We have a bunch of exciting new products that are going through alpha testing right now. We've released it to some users and they will soon be publicly available. So scouts initially could only access public websites, so they would monitor for things that are public information anywhere on the web. The next logical Natural step is for these agents to be able to log in on people's behalf to access their contacts and their websites. So that's what we're aspirational right now. So that's one thing that's going to be coming pretty soon. The other natural extension of this is that agents that don't just monitor but can also do things for you. Right. Like buy book, reserve things, write documents and so on. So not just read actions, but write actions. So that's also something we are alpha testing right now. Those should both be going out pretty soon.
- [09:19]
C
And how do you envision this? Will it be a bunch of different individual products? You know, what does this look like? It sounds like you're building a lot, you're shipping a lot. Does it all come together?
- [09:27]
A
It does, yes. So it is going to be a single product which has all of these capabilities under the hood is inching more and more towards being like an AI first chief of staff that can do a bunch of tasks for you that you would normally do in the browser on the web. So yeah, it is going to be a single cohesive experience that has all of these capabilities.
- [09:46]
C
And if you zoom out here, if you go out maybe three years, I know that's probably hard to do, but do you think it'll be mostly prosumer will be enterprise? Like how do you think revenue will be split here in the future?
- [09:57]
A
Yeah, I mean three years is too long to make projections for in today's day and age to be honest.
- [10:02]
C
Yeah, it's fair.
- [10:03]
A
Yeah. So like one step at a time, right? Like I do think a lot. So most of our users right now are prosumers are individuals who then use it at work and like we end up onboarding more people from their team and so on for the Scouts or the next version of the Scouts product that we are alpha testing right now. It is meant to be prozumer first. It is meant to be individual first. Yeah, like the AI first chief of staff. The phrase that I keep sort of repeating that is sort of the right hook for it. And as you might expect, like a chief of staff is attached to an individual. Now they may use it for work or for their personal lives, but that is how we are envisioning this version of the product. Back in November we also released another product which is our API. So that inner wrapper around the models that we train in house that has a very different go to market motion. It is more developer focused. Many of those conversations are also top down instead of bottom up in terms of like onboarding enterprise customers. So there it's a browser automation API. Basically the closest competitors would be something like Anthropic's computer use agent or OpenAI's computer use agent. Ours is not as broad as being like a computer use API. It is specifically a browser use API. So anybody who wants to automate workflows like that involve navigating websites, filling forms, clicking through buttons and so on can use our apm.
- [11:22]
D
This show is brought to you by the global talent company, a marketing leader's best friend in these times of budget cuts and efficient growth. We help marketing leaders find, hire, vet and manage amazing marketing talent for 50 to 70% less than their US and European counterparts. To book a free consultation, visit globaltalent.co
- [11:42]
C
marketing to consumers and then marketing to developers. You know they are very, very different worlds. Like do you see that as there's almost two different marketing programs then and are there different marketing team? Will there eventually be different go to market teams?
- [11:54]
A
I mean so far there's no marketing team in us. It's just us, the founders to various people. We've seen good cross pollination across the two. Like a lot of developers try out scouts first and they're like oh okay, this gives me an easy way to get a sense for what the product can do. And then like okay, give me an API so that I can hook it up to my system, right? And vice versa. Like there's developers who land on a product or either heard from us at a hackathon or something, but they're totally okay not having the full flexibility and control and the manual work that may be involved in using an API. The web app which is the prosumer product does give some amount of flexibility as well. You can hook it up to your own integrations through like zthir et cetera. And so like they're perfectly happy with that. So like there is some amount of trust pollination but these are yeah like to your point, there's no taking away that these are different crowds on the marketing front. Like we do different things to reach out to these two audiences. Like for example, we've been hosting hackathons. We did our second hackathon in SF last week. So that's primarily to like meant for developers to alpha task bidded us new features on the API front. But as for prosumers, it's more like us announcing features on Twitter or like existing users telling their friends to try it out and so on.
- [13:09]
C
How do you approach hackathons?
- [13:11]
A
So far it's been in collaboration with like other organizers who've been doing hackathons. And we just ended up being like a co sponsor for a couple of hackathons that seemed to have the right theme that made sense for us, but we might host our own pretty soon.
- [13:26]
C
How do you think about just, you know, competing for Mindshare? Obviously there's a couple of other AI companies that may be making noise. Like, how do you navigate that world when it's so insanely noisy?
- [13:36]
A
Yeah, it is. And this kind of goes back to what I was saying earlier. Like, platforms like Twitter do feel extremely saturated. So far, what we found to be the biggest bang for the buck is even if we have a small set of users initially just impressing them, blowing them away, sort of like gives them something to tell their friends and colleagues about. Right. And word of mouth has been a big channel, big source of acquiring new users for us. So that I think is topmost for us. Beyond that. Yeah, like just maintaining a quick cadence of shipping, making sure every new user who signs up or existing users who come back to the product again, they feel like this is a product that is alive where like new things are getting added on like a weekly basis. Just doing as much as we can to create the surface area for users to be telling their friends and like bragging about the product in different ways.
- [14:27]
C
What have you seen people use it for or talk about that you're most excited about or maybe surprised you were like, oh, holy shit. Like it can do X, Y, Z.
- [14:34]
A
Maybe this is like a little bit of a meta point. But like, one thing I was initially surprised by is how good users were at writing prompts relative to us, the developers, or even the researchers who are building the model that powers these products. Like, users write really, really good prompts. And a lot of that is because of like other competitors in the space like ChatGPT, Claude, etc. Have kind of trained a lot of prosumers at like, okay, write highly detailed specific prompts to get high quality outputs from these models. There is garbage in, garbage out. So like, if you don't write a detailed prompt, if you give it generic query, then you are going to get slop and you're not going to get most out of these months. So that was one thing that surprised me. Like, even for this relatively new product, which has a pretty specific value proposition of like monitoring, users were writing really, really detailed specific prompts in terms of like, use cases that I was pleasantly surprised by, I think the sales lead generation and go to market is not something I had expected. It as strongly as it did. So that would be one. There's also been a ton of marketing teams who are using scouts to monitor either specific clients that they're working with or. Yeah, like monitoring for mentions of, like, any popular individual or company out there. Ebay search has been surprisingly. Yeah, like that particular use case has been quite surprising. Like, ebay has a ton of rare collectibles in their Aptria products, But search on ebay, for whatever reason, may or may not. Yeah. Is probably not that good. And so people use scouts to monitor or search ebay for rare collectibles that they're interested in. And that works quite well. We've had a ton of people use it for like, real estate monitoring. So like monitoring for house listings for like rent or sale. And scouts works quite well on Zillow, Redfin and the likes. So that's also been a popular one. Yeah. Most of these, I hadn't tried these myself before we launched the product. So these are all just discovery through users trying it.
- [16:31]
C
And especially given the state of the housing market in San Francisco, that must be booming. It must be booming right now for that use case. Final question for you. And I know the future is kind of hard to predict, but what's the big picture vision here for everything that you're building? What does it look like? I don't know if we want to go out three years, maybe go out one year. What's the big picture vision inside your head?
- [16:49]
A
Yeah. So I think by and large, interacting with the web hasn't quite changed in 20, 30 years. Right. Like there's a browser, you open a website, you click buttons, fill forms, navigate web pages. Right. What we are betting on is that the technology with AI and with agents, and some of which we are also developing in house and contributing to that ecosystem. What we're betting on is that the technology is there for there to be an inflection point in what that interaction with the web looks like, where people are not manually browsing websites themselves. They're operating at a slightly higher level of abstraction, just specifying what tasks they want done. And there's web agents that's handling it on their behalf. We're in the business of building that technology and like making a few experiments on what product experiences around that technology might look like. If we can make a dent in reimagining what interacting with the web looks like for developers, for prosumers across the board, that is sort of the vision that would feel extremely gratifying.
- [17:50]
C
Amazing. Love it. All right, my friend, this is where we're going to end things before we wrap. For those listening in that want to follow along with you, where should we send them? Where should they go?
- [17:57]
A
Check out utility.com that has all the information about us, our team, our products, and so on. Or you can find me on Twitter. My handle is abhshkdz.
- [18:06]
C
Amazing. Thanks so much, man. Great to meet you.
- [18:08]
A
Thank you. Great to meet you as well, Brett. See ya.
- [18:13]
B
Well, that's all for today's episode of Build, brought to you by the Frontlines. If you want more amazing content like this, visit Frontlines IO, where you'll find the library of more than 1500 interviews with founders, marketers and other GTM leaders, where we unpack the tactical lessons from their journey. And of course, as always, if you do want to launch your own podcast, we'd love to have a conversation with you. Visit Frontlines IO Podcast as a service. Mention that you listen, mention you love the show, and we'll give you a 10% discount.
- [18:41]
C
Thanks for listening. We'll catch you on the next episode.