← Always Be Testing
Always Be Testing
#129 Building Enterprise-Grade AI Tools Safely | Danielle Sakher & Tye DeGrange
00:41:16
Loading summary
Transcript57 lines
- [00:00]
A
Foreign. Welcome to another episode of the Always V Testing podcast. I'm your host Ty DeGrange and I'm really excited to have Danielle Saker today. Danielle, how you doing?
- [00:17]
B
Hi Ty. I'm doing great, how are you?
- [00:19]
A
I'm doing well. It's great to see you again. It's been a moment.
- [00:22]
B
Likewise. So good to see you and thanks so much for having me.
- [00:25]
A
Great to have you. I think everybody today's in for a treat. Danielle is leading go to market for Remy part of Mind Studio, one of the great AI builder tools out there. She's going to tell you all about. So lots to learn, lots to talk about in the AI world and the growth world and go to market. So it's going to be a great, great episode. Danielle, maybe give us a little background on what exactly it is you're, you're building so we could get into, you know, what you're being tasked there and how you're approaching it.
- [00:54]
B
Yeah, will do. Happy to. So we are building Remy. Remy is the platform for organizations to build their software instead of buying their software. So it is a platform where you can build full stack applications for your team, your department, your organization without having to stitch together different services for hosting or auth or databases. Everything happens all in one place and makes it really easy for non technical users as well as developers to build all in one platform. So really excited about the brand new launch. We are just past our alpha stage and I can talk a little bit about more about how that went and what's ahead in beta and yeah, just, just really excited about what's ahead.
- [01:44]
A
I love that. And what are some of those use cases? You kind of alluded to it like you're kind of in this interesting in between layer of like not super like heavy power user but not the beginner vibe coder. What does that look like and what are you finding the use cases are?
- [02:00]
B
Yeah and actually you've, you put it almost perfectly there right is there's these, this category out there in the market of tools for professional coders for developers. The anti gravity is the cursors codecs quad code. Then there's vive coding tools that are really easy to use but are more amateur or you know, hobbyist tools. The lovables, the replit, the bolts. Remy is right in the middle and really designed for enterprises, organizations of, you know, small for sure but definitely in that medium to large sweet spot that need more observability on what's going on within their organization. Who's building what need more Security controls. There's a lot that when Remy's building an app, it's going to do and it's going to output that you're not going to get, you're not going to get from some of these other, you know, these vibe coding tools that are out there in the market and can talk a little bit more about that as we get on along here. But that's at a high level what the market's looking like right now and where we are in it.
- [03:00]
A
It's crazy because there's just an explosion right now of talk about Claude, adoption of Claude, talk about lovable replit and you know you've, you've got to like big names and competition and this, this is exciting times in this world that you're in. Are you, I am curious about the kind of professional level, how is that, how do you, how do you define that kind of being? Like why are those products? And then there's probably a simple answer like how do you kind of define those levels for people in those different use cases?
- [03:31]
B
Yeah, I think like maybe it's best articulated with a story. Like a conversation I had recently with one of our customers or one of our users. She's a partner at a PE firm. They're you know, small to medium sized firm. They've got about six assets under management. The companies that they acquire around like 5 to 50 million in revenue a year. So they've got some interesting port codes that they work with and she's been, she was building out with Claude code a tool that would help the port codes just like up level their competitive intelligence, their biz dev activities without kind of having to adopt a bunch of different tools and figure out how to use them. It would be something that is plug and play and all in one place. So she had spent about a month trying to build this out with cloud code going back and forth and she mentioned, you know, 50 to 100 times different iterations. And I was like, hey, you know, like try like you've already tried this with cloud code. Amazing. Why don't you give Remy a go? Remy got to where cloud code got in a day and you can imagine with like more iterations, you know, it'd get, it'd get somewhere even, even better if you spent a little bit more, more time with it. So that was really exciting to see from just like the speed that's kind of by design, like Remy has, has design principles and architectural principles and integrations that it's leveraging that really produce more high quality, more production grade Products right off the, off the bat. And what she's also going to get, you know, she's in financial services, right? Highly regulated industry. Remy is going to output for every app at the state of publishing. It's going to put out audit logs, it's going to give you a cybersecurity schema framework for each application. You're going to have that level of observability to just using this app. You'll have security and governance audits on the app itself. It'll tell you how it was built and what data it's leveraging and not leveraging and all the details that it is surely going to be asking about in a short order. For any tool that your team is using, you'll have a cake attestation. So many different things that are pumped out of the bat because it knows what, like we know that the, that for organizations right now that it's like a nightmare, right? People are adopting all these different AI tools. They're maybe using personal accounts or, or not. Maybe they are typing, you know, sensitive data in there. And it's just, it's a really difficult management and oversight issue right now. And what, what we've designed Remy to do is be that solution for organizations who need that observability and need that level of control. Which means we've, I mean, we've, we're horizontal, but we're seeing a lot of, A lot of really great traction with financial services with like nonprofits who deal with sensitive information, governments as well. So it's been exciting to see so far.
- [06:28]
A
Got to give you props. I don't think cake attestation is the first time that's been dropped on the pod, so you get extra bonus.
- [06:38]
B
Hopefully. Hopefully the last. I'm like. And this is like the. Anyway, not, not my area of interest
- [06:47]
A
in that use case you described for that PE firm. Like, what do you think? And in others as well. Like, what do you think? Is that you don't have to obviously give the whole secret sauce, nor is it possible. But like, what's the, what's the thing that you think led to that better outcome? If you could probably ask for, you know, guesstimate or try to share a little bit more behind the curtain with the audience.
- [07:08]
B
This is an app that I built actually. This is because I've been on a total, like FIFA craze and you know, live near a host city. I was like, where are all the watch parties? I like, want to be with all the crowds and all the good vibes. So I built Like a watch party finder for Toronto. So if anyone wants a link for that, I can share that. We've only got a couple days left here, but in any case. So what Remy did did is. Is consult so many different agents as it was going through the build of this, right? So you can see all the different commands it took.
- [07:47]
A
So you can see the live code.
- [07:49]
B
Yeah, you can see the live code.
- [07:50]
A
If you can see the process it was built, you can select agents.
- [07:54]
B
Yeah, totally. So, and this was like, this was like a fix that I asked it for. Like, as it got along and it put together, even just for one fix, it put together a spec that I could annotate, that I could comment on, that I could refine before it actually went to build it. So, so that no assumptions are being made. Right. And we don't have to go back and forth, you know, a dozen times to articulate the right thing. So lots of. And here's, here's my. Yeah, all the different agents that is consulting, it's. It's done like it has an entire design agent that goes and puts together branding and different, different notes. And it also shows off where's my product roadmap agent and thinks about different things it should build later on and puts actually a full roadmap together. So it's just doing a lot more upfront and it's thinking like a PM instead of thinking like an engineer or a coder.
- [08:42]
A
That's. That's huge. I really, I find that interesting. I like the building block, I like the kind of logical flow and I like the, that you can see different aspects of it. I think that thinking like a PM hits. So that's super helpful. You know, one of the things that you said in the, in the conversation, you know, we kind of joked about the nomenclature, the technical stuff that, you know, some of the audience is going to go, what the heck is that? But it's, it's obviously, you know, germane because you're talking about technology, you're talking about building technology, which can get pretty complex and some of these, you know, vibe coding, entry level things are not going to cut it. I don't know, there's a little stretch because it's like off script, but like when I was at Alex Karp and I think some of the hype that's being created right now around, you know, what are you giving to the LLMs, right? And, and we don't need to go into all the details of that, but is there some aspects of what you guys are doing in Terms of enabling control and privacy. I don't want to put words in your mouth, but is there, how does that, like, you know, controversial or great take, however you want to land on it, relate or not relate to what you guys are building? Because I think it's pretty topical right now with regard to AI tooling and the feedback loops that these people, these builders are giving to these tools. If that's a one you want to pass on, I also understand.
- [10:00]
B
Yeah, I mean, here's the thing. Like we, we meet the organization where they're at, right? We have some orgs that we work with that are super AI friendly and understand that, you know, there's kind of more to gain from the LLMs going and doing their thing and running different processes or building applications versus, like, okay, like the competitive, like, you know, there's the, usually the competitive side that you're scared that giving your data to these models is going to somehow lead to some competitive threat down the line with trade secrets or whatnot, or there's the threat that, you know, a competitor is going to just do it and be ahead of you. So it's there, there is, it's a, you know, there's two sides of the same coin there, I think, I think, like, obviously, you know, there are certainly, you know, regulated industries. This is not, not legal advice. Right. Like, you've got to make sure that you're in compliant with whatever your organization guidelines are or your industry's guidelines are. But I think for the, for the most part, we're seeing a lot like the organizations we work with are really excited about the different capabilities that are possible with our platform, with models more broadly. And I think too, that there's a lot that you can do without catching any of your data. Right. Like, you can, like, we've built tools that, like an example, like we have a tool to find different influencers and content creators that we might want to work with and might want to partner with. The only information we've really given that app is just like, who it is we are. And then it's going and doing the work of like, scraping the web and whatever else to find creators that might be a good fit. And any information that we've given it is just public domain anyway. So there's a lot of things that you can build without relinquishing that level of control and that level of, you know, maybe sensitive data that our organizations might be concerned about.
- [11:57]
A
Yeah, I'm, I'm very fascinated by the concept. I think there's, the movement is going to continue to be topical and be important where, you know, companies are going to demand or, or aim for levels of privacy and protection. And I think, I think you're offering some pretty exciting things there. One of the things you talked about a little bit, you know, we, we talked a little about like costs and this has been a huge pain point in conversation even for our team of like, for us it hasn't. We've kept it manageable. But I know a lot of people, you know, there's a lot of news around it. Token usage goes up as price goes up and some of the flavors of, oh my gosh, it's, you're back to almost wanting more to hire more people as opposed to more AI compute. How do you guys think about that and how are you sort of supporting it?
- [12:42]
B
Totally. And candidly, I, I don't know that we figured it out in terms of like a pricing and packaging conversation. Like, I think that in general SaaS is having a bit of a weird moment right now. And that's cool. We'll figure it out. We'll make mistakes along the way and that's fine. Like, there's so many different models out there. There's like companies that are creating whole like credit economies and you know, different things that are so, so difficult to understand and are not standardized. And I don't know how I feel about all, all that stuff but, but all that to say, I think like now with AI, there are things that are possible that were never possible to certain segments of the market, right? Like building an application was not accessible to like the solopreneur in very recent history, right? Like someone who like you'd need to raise money or you need to be individually wealthy or you know, be willing to take on a bunch of credit card debt to, to build an application if you were a solopreneur, right? So what that means now, like with AI, now you can build an app. You know, you can build a kind of maybe a lander with a replit or a lovable. For few bucks you can use Remy and build a production grade application for 50, 100, maybe 200 bucks. That's, you know, that's a price point that didn't exist in this market before. Like that just wasn't possible. That's like a, that would maybe get you like 200 bucks, would maybe be one hour of like a dev's time to like look at your spec or like talk to you and like even consider taking on your project. So, so it's like, you know, it makes the Cost sensitivity.
- [14:24]
A
I'm building so much totally.
- [14:26]
B
And it makes the cost sensitivity question like really just like a weird equation now because like we're comparing like maybe 200 bucks to 5 bucks when like we really should be comparing like, like a few dollar bills to like the tens of thousands of dollars, if not hundreds of thousands of dollars that it would have taken to build like, like think of like an enterprise who'd want to build a legacy tool, like a internal tool or legacy tool, how much that would have cost? Like what if these legacy tools would have costed them, right? Like whether it's people power or hire an agency, a dev shop or you know, what have, you know, we're talking millions, right?
- [15:02]
A
So totally.
- [15:03]
B
It's, it's, it's a totally different equation now. And, and it's begs a lot of questions of like you know, who, who are you for? And then therefore you know, how do you, how do you process out given what level of service you need to offer?
- [15:14]
A
When you think about like the ideal customers and we kind of, you know, we kind of started a little bit with this in terms of the segments of where it's at. And you know, I'm curious to know where your thoughts are, especially as it's, you know, the market and the tech is moving so quickly. How do you, how do you and the team think about that when you're leading go to market?
- [15:33]
B
Yeah, things are moving so quickly and you know, there's all this different, like new tools that come out every week. There's hype around certain tools. There's a lot of pivoting that we need to do as well in terms of like who are we for? So it's, it's a lot to manage. We've had different people show up as we've launched different iterations of our product. Some of those people are enterprises. We've had New York Times upwork. We have a lot of really great enterprises that we partner with. Some of them are governments. We've been working really closely with the UK government over the years. Some are like enterprise professionals trying to like launch a side hustle. Some are agencies, consultants, like full time entrepreneurs. Like it's really tricky and you want to, we can be for everyone, but it's also really hard to be for everyone because of everything we just talked about, whether it's pricing, whether it's service level, whether it's, you know, the types of features that are, you know, the coming back to Cape. Right. Like, like a solar partner doesn't care about that or doesn't need to care about that for a very long time. It's. We're right in the middle of it. I think of, like, try, like we know that we're building an enterprise product. We know that there are, like, enterprises have showed up and gotten a lot of value out of it. And it's kind of on our team to make sure that we are not getting distracted by other segments of the market that are loving our tool, want to use us, are asking for more, giving us feature requests. We love all that. We have such a strong community. But we also can't let that distract us because ultimately, you know, no matter how AI enabled we are and how many amazing applications are helping us in our work, like, everyone has limited resources. So we just really have to stay focused and not get overly excited by positive signals that are from the wrong people.
- [17:22]
A
Well, what I appreciate, Danielle, is like, I think you have, there's a level at which you've clearly kind of done the thoughtfulness and done the research and talked to the customers. And I think that's, that's where the magic lies with being a PM and being a go to market. And that's what you're touting you're doing, which is huge. So appreciative of that, I can tell that you've done your homework there.
- [17:43]
B
As much as we can talk a lot about.
- [17:47]
A
Exactly. There's always more to do. It's never done. Speaking of growth. Never done. Growth cycles have been collapsing. Right. We've got, you know, we've talked about this a little bit. How do you get efficient roas? How do you grow? As, you know, you think about distribution, you think about go to market. And as I alluded to at the beginning, I mean, you're, you're going up against like crazy heavy hitters here that, that are getting a lot of distribution out there. You know, as a, as a growth leader, like, how do you tackle that?
- [18:13]
B
Yeah, we kind of like if you don't, if you can't beat them, join them in a way. So there's all these different, all these different tools out there. Each one kind of has their. There's a lot that are getting their 15 minutes of fame, so to speak. And it's kind of our job to have a perspective on these things and know how these things work, know how they can complement our platform, know how they are different from our platform and like to, for us to be able to communicate that. And we got to do that anyway, so we might as well share and educate folks because the demand for AI education right now is, is like, I don't know, you might, you might argue it's, it's more than, than the demand even for like the software and the tech. Right. Because if you don't know what this stuff is and how are you going to be able to make a buying decision?
- [19:00]
A
When you were talking earlier, it was very paramount became very apparent rather that you, you, you can kind of make killing and be very useful by just helping organizations navigate this situation and, and letting them know which tooling they need to use when how like your specific tool literally is part of that decisioning and it's, it's wildly under, under utilized at this time in my opinion.
- [19:28]
B
Totally. I'll, I'll finish the point I was making earlier in a second here. But to that point we have at some, at some point our go to market motion has looked like very much like, like an agency or like a pro serve and motion and not because we want to be a pro serve business or that we really want to do a ton of professional services, but almost because it was required to even just to kind of mask all of the things that are confusing right now to organizations and overwhelming really smart people and really smart professionals who just can't make sense of, of everything that's going on just because of the speed and how quickly everything's moving. So we almost had to like just not even sell the tech or even mention the tech at times of just like leading with, with our expertise and that we're building the space now. We, and that we've got you. We'll figure it out. We'll use our technology. But you know, rest assured you guys don't need to, to worry about hiring someone to figure out how to use this stuff or, and, and not that, not that it's not that it's really hard, but almost. It's like, it's almost just. There's. What is it like you get so overwhelmed by the level of information not necessarily coming from us, but just coming from socials or from the news or you know, whatever from conversations with colleagues that you almost can't, can't even make room for even an understanding how to use like very simple and basic tools.
- [20:51]
A
I think you're not alone in that. There's so much data to ingest and so much priorities and change that it's very, there's a reasonable amount of overwhelm and just paradox of choice, which I agree with.
- [21:05]
B
Just to finish the conversation on this piece. Something that has worked really well for us is actually hosting like Free webinars, free talks on, like, making sense of the overwhelm, making sense of the chaos states of the unions on AI. We did this when OpenClaw was released. We did this when cloud code was released. We are now in the middle of doing a bunch of Hermes sessions where I think we've got like our third one coming out in, I think just in a couple days now. And that has one, like, helped us gain authority and get our people out there, get our name out there and be associated with like, hey, like, these are folks who are going to help me and help me make sense of these things and, and almost to, you know, to even get credibility and have people listen to us. We kind of have to talk about what we're building with Remy and, and everything we're doing here. So it also illustrates a little bit of curiosity with folks and who, you know, leads to MQLs and things that we want to get out of these sessions as well. Even though we're not even talking about Remy at the session itself. Right. It's like on a totally different topic. So we are not afraid of, like, other tools getting their moment in the spotlight. It's an opportunity for us to talk about how they're complementary or how they can work together with Remy or just like, help people make sense of them. Because this is what we do. And it's led to a lot of really great leads for us. We've got like, with these Herme sessions, we've got like over a thousand people registered for each of them so far. And it's. And, and like, it did help, like, even just us posting the recording on our YouTube has helped our YouTube do so many numbers. So it's, it's, it's great. And we'll, we'll continue to do that.
- [22:47]
A
Is there a specific hook or something that you think they're getting either pre or during the webinar that's leading to such great. Aside from, hey, this topic has a need. Like, it's, I think it's super topical.
- [22:59]
B
Yeah. I think, like, for folks, they, they'll see like for myself, right, I'll see something in the news, headline after headline after headline. But it's also surface level. Right? You're not actually getting into like, what is this is. And what does this mean for me? The ability to like, sit down and have someone who has expertise and has played with the tool. Whether, like, expertise is debatable. Right. Because all this is so new. Like, who's really an expert. Right. If you've played around with it for like a day. You're like, you're now in the top one percentile. But having someone sit you down for an hour or two hours and just like talk you through what's going on and how, what's important, what's not important is so immensely valuable for people. And that's really the hook is like, hey, show up. All you need to do is show up. We'll talk you through it for an hour and don't take notes. Like, we'll take notes for you. We'll send you all the stuff after you'll, you'll get what you need. Instead of spending like two, three hours watching a bunch of different videos, reading articles, maybe trying to install this thing on your own, failing, going back to it. Like, just don't, don't do any of that. You can do that if you want, but start, start by letting us talk you through it and then you can go and do that experimentation if you want to later.
- [24:10]
A
I love it. And thinking about go to market growth, obviously SEO AEO or GEO depend, how you want to describe it is really critical right now. From what we've talked about. What, what are the types of experiments that you, what are some of the learnings you've had in that arena? What are some of the experiments you've done there and how are you thinking about those, those levers?
- [24:30]
B
We've done a lot of experimentation in the last few months around SEO aeo. It is really important to us, like, really important. Go to market angle for us right now. We have, it's tricky, right? Because I think like when you let AI talk about your product, there's like inherent risk, right? Like there's, there's going to be things that it gets wrong from time to time. And I think that, you know, in the past has made us scared, for lack of better word, to really like turn on the machine. Like turn on the, the SEO monster that we knew we could build. And I think like, as our messaging got crisp, we got less and less scared to do that. And yeah, like we knew what needed to be built. It just had to get turned on. And what it was that needed to be built is, is like an application that would look at what the top YouTubers in this space, AI YouTubers, tech YouTubers, professional, anyone with kind of like that professional angle that are speaking to knowledge workers and professionals. Because these guys have done the work of knowing what is trending and what audiences want to learn about and hear about. So it would look at what they're talking about. Tweeting about, making videos about and create articles around those topics that sometimes we'll talk about. Remy. Sometimes they won't. If it's relevant, they will. If it's, if it's not relevant, it won't. But there still be like interstitials through kind of like ads throughout the article that would link back and mention, you know, who's, who's, who's behind this blog. And this has just truly overnight, if you look at our website traffic, it was just a hockey. It was here and then now it's here and it stayed here since doing this blog. And we'll have really interesting, we've got listening tools that will find us when people are talking about us or talking about our blog and whatnot. And like really interesting people are reading our blog and posting about it and saying, hey, like these are like this has replaced like doom scrolling for me. So like people are getting value out of it too. It's not just like AEO for the sake of ao. I think that's like, that, that works for some people. But I feel like that's not really what we're trying to do here. Like, we are trying to put out stuff that's, that's valuable and topical and, and interesting for, for, for folks at the same time. Of, of course, like, you know, the traffic is really important and being mentioned in chat tools are really, really important. That's where a lot of people are doing their research now. So that's. We, we, we keep an eye on this like every day. We tweak it like almost every week. One time we tweaked it and it like tanked our traffic and we had to like roll back. Like that was the whole thing. So it's, it's something that is like really top of mind for us and something we're investing in on, on a daily basis.
- [27:23]
A
Replacing your doom scrolling is going to be the title of the podcast. So I love that important work you're doing. God's work, as they say.
- [27:32]
B
Yeah.
- [27:33]
A
So we talked a little bit about partner marketing, but like I'm excited to get your perspective on it. Giving your knowledge of what we do and giving your knowledge of the space and just understand kind of how some of the mechanics around it and how you might think about that. That lover being such a critical piece for SaaS and just what. How are you thinking about it?
- [27:54]
B
Yeah, I mean like I've, you know, I've worked at some really amazing companies. You know, I, I had taken partnership roles at Shopify and Grammarly and Zapier before spending my time here with the Remy team and in those roles, like, I remember I used to spend a lot of time trying to get buy in from like my dream partner, right? Like, this is my dream partner. I'm going to try, like, I'm going to find a way to get that in. And it might take me three months, it might take me six months, might take me a year. But, like, I'm going to be that, like, nagging force and like, you know, there's still a little bit of that in, in this role now. But the nice thing about the time and space right now is that there's like, a lot new opportunities and a lot of like, new problems that potential partners are looking to solve. And I, I just can't wait two years for a partnership announcement anymore. Right? Like, that's like, you know, it's just not like, that's just like not the cycle that, that we're in. So I spend a lot less time trying to get buy in and more time working with partners that where that buy in is already there or close to being there. Not because I'm lazy maybe a little bit, but because we just the speed of things demands it, the pace of things demanded right now. So what I do is I kind of have like a little, you know, matrix matrices can help people sometimes. Frameworks helps people sometimes. So it's like, do you. Are you solving a problem together? Yes or no? And does the partner have a high urgency to solve that problem, yes or no? So I'm looking for the partners that where we can solve a problem and there's high urgency for that problem to be solved. So a couple of, maybe a couple examples I'll share of partnerships we've launched in the last few months and or have been, you know, continue to work in the last few months where that was the case. We have working really closely with the team at upwork. Upwork is like the largest freelance talent marketplace out there. They had a really growing demand for like vibe coding projects, quote, unquote, application building projects. It's like they're one of their fastest growing categories. Like we're talking upwards of a billion dollars in volume per year. Amazing demand. But on the supply side, like no real objective way to show freelancer competency in that space. So we, what we've done together is we've gotten a bunch of their freelancers certified in AI develop, AI product development, and they are certified by us. So it's one. Do they know how to use our platform? Are they good at it? How have they shown a number of different projects that they've worked on and then they get that certification and now people who are browsing upwork and they've got you know, over 700,000 clients on there who are always looking for, for, for freelance talent. They can filter by folks who have that level of certification that is you know, outside and, and, and, and certified. That's like one example share maybe one more of a partner who had a problem that we were able to solve really quickly. One is, was, was with HubSpot. HubSpot, you know, kind of existed as this like mid market tool for a while, right. But given you know, I guess maybe with AI or you know, I think a number of factors. There's so many more like smaller, more entry level teams who now might be ready for something like, like a, like a HubSpot. Whether it's because of the demand, like how important email marketing is now or you know, what, what have you. So they've been working with a lot of founders, SMB startups and the problem for these, that demographic is like empty state CRM, right? If you spend all these, you know, hundreds, thousands of dollars for the CRM and you've never used a CRM before and it's empty, you're going to churn, right? You're not going to be getting any value out of any of their great powerful features. So we built an application with them that would let founders in like one click based off their ICP import, a bunch of contacts that can fit that ICP to their HubSpot instance directly. Right. So that solves the problem of empty state CRM and really improves the stickiness of that HubSpot user. So they were willing to. Two examples of companies who were willing to one had a big problem that was affecting business outcomes and then two had a really strong urgency to solve that problem. We were able to come in, build something really cool together, get it out there and pretty quickly. Always, always can go faster, but pretty quickly.
- [32:27]
A
That's amazing. That's very interesting. It's good to hear that. I love the specific examples and knowing I think some of those tech integrations there's so many cool things to do with partner and I could go on and on about it, we'll get into that. But the technical integrations and those similar use cases and I really appreciate the matrix reminds me of the Eisenhower matrix which is a great one and very, very good work there. Kind of coming down the home stretch here we obviously the POD being always be testing, you know, talk about Real time experimentation. And I'm wondering like what, how are you thinking about that and what, what are some of the big maybe aha's or surprises you've had maybe over the last month of like experimentation and what you've seen?
- [33:12]
B
Yeah, absolutely. I think like this Remy alpha in itself is like, you know, you could, you could argue it's one big experiment. Right? We, we're really surprised and you know, we're, I think we discussed this a little bit in the pod, but Remy is brought to you by the folks who built Mind Studio. We've been in the AI automation workflow automation space for upwards of three years now. And to me it was really surprising just like how quickly people got it. Like we were out there launching this, you know. Yeah, there's like, you know, we're still in the AI space, like we're, we're still doing similar things, but this is like a totally new product. Right. And I was really surprised how quickly people got in and how much demand there was even for like an alpha stage product, especially with all the noise that's out there. We just wrapped up our alpha, a thousand users in the alpha and these users collectively spent over $300,000 in tokens in compute and inference to build their applications. Right. Like that's, that's a lot of skin in the game from a wide variety of individuals. These range from like organizations and teams to founders who are trying to get a new product off the ground. And it's really cool. And I'll flip you just like really quickly one of our internal dashboards just so that you can get a visual here. This is all like unidentifiable customer data. But you can see here we have users, these are individual user accounts, might represent larger teams, but who are spending tens of thousands of dollars in some instances, really just whether it's on one app and they're going really deep with it or developing apps across their organization for a multitude of use cases. So this is so, so, I mean, I don't know if surprising is the right word, but really validating for us as we're in this alpha stage and I think will set us up well for what's ahead in beta.
- [35:14]
A
Oh, I love that. It's so cool to see and I love seeing the visuals and I love seeing real time and to be able to say, hey, this is a test that we've got some learnings on is extremely appreciated and powerful. Bringing it all together, it's been great. There's been a lot of things talked about in Terms of like beta and all the testing and data driven thinking and talking to customers. If there's some things that you kind of think about that maybe didn't go so good in the data, but you're kind of like, okay, how do I make sense of this? What did it teach us about our audiences? What are some of those. Maybe your favorite failure.
- [35:53]
B
Yeah, totally. And Ty, I don't know if you're gonna like this one, so we can, we can cut it or go in a different direction as we go along here. But I'll say this right. Like, I think like something that we did too early was spending, yeah. Spending big money on some big influencers. And I think like it was, you know, maybe we thought it'd be like a surefire win, right. Like we would drop a bunch of money, we'd go viral. Right. And then, and then we kind of sit back and relax and let that, that hype take us, take us away. We've had really great success with influencers and content creators in the past. YouTube specifically has been a fantastic like YouTubers have been a great channel for us specifically like dedicated content. Not so much, not so much ads and, and, and mid roll and that kind of thing. But dedicated videos have been really fantastic for us. But I think maybe it was a little bit too early with Remy. I think like investing more of our time early on in nurturing that, that alpha audience. Bringing on like handholding, which is what we're doing now is really hand holding organizations to join our beta as like our founding pilot partners and really, really handholding them, giving them a lot of support and preferential treatment, helping them really find all those different use cases there. I think that would have been a better use of our time and resources versus trying to jump hype change for a product that was still an alpha, still had, you know, still had like it was we and we communicated still had, it was still bumpy at times, wasn't perfect. I think that that dollar value could have spent, been spent elsewhere.
- [37:38]
A
Yeah, I love the example because it's like that we're of the mindset of like oh look, things work and things don't. We want to compete with the best of the marketing levers. And I think it's, it's health healthy and effective to obviously, you know, share openly those and I think as you alluded to, right. Sometimes some things are slightly earlier but sometimes maybe bigger on those tests can sometimes not always be better in the influencer game specifically. So I have no doubt that when you know, we talk we get into details of influencer and reviews and partners. Like we'll find a way. We'll find a methodology. We'll find a. A valuable piece that. That wins and works. So but it, but it's not for every business. It's not for every stage. And I think you're appreciate the transparency and I think that's. That's our game too. To be like let's let those cards land as they do. You'd rather compete in, you know, the hardest of mmm models to make sure that things are working and as efficient as once you get sophisticated enough to even get there. So it's all good.
- [38:42]
B
Totally. And I like, I think like, and I think like actually and I actually think working with professionals on this are really important because they can help let you know when. When is the right or wrong time and, and maybe help you avoid some of these, these failures or some of these iterations and, or if, if, if you know, depending on like maybe there are certain things that you can do at one stage that actually is effective. Right. So I think like we kind of threw a bunch of money at. At some big names that or so. Or that we thought were big names too. I think so much of what's out there now is gamified and it's really actually hard to like even with a discerning eye, it's really hard to be able to make sense of what engagement is real and what's not. And you see these big numbers and your eyes light up and you're like oh wow, like this could be really great for us. And then it's not. You like spend $10,000 and you get six new users out of it. True story. So it's like not, not always.
- [39:45]
A
Totally.
- [39:45]
B
It's like not always the best going in blind. And it is actually really helpful I think to have a bit of guidance in, in navigating this space for sure.
- [39:53]
A
I love it. Let's just going down the home stretch, get some personal to get to know Danielle a bit. The, the. The trip stuff was awesome. Like love to hear your perspective on those. The dream trip or the recommendation?
- [40:06]
B
Yeah, my dream trip is Egypt for sure. It's been on the list for a while. I just think like history and just like the pyramids, man. Like do I need to. Do I need to elaborate? So, so cool. And then for folks I recommend this was. It was so beautiful. Right. For folks I who haven't been to Mexico City yet, I think it is such a great spot. Whether it's for just pure vacation or Digital Nomad or some combination of the two. Such a rich and vibrant city. So much to do there, from nature to city life to food and everything. So I really, really recommend it. And go. Go for a while. Don't. Don't skimp out on that one.
- [40:50]
A
It's high up on my list. Danielle, you've been amazing. It's been a pleasure. You covered a ton, and I just want to thank you for coming on. I'll talk to you all soon and until next time. Thanks, Danielle.
- [41:01]
B
Thanks, everybody. Thanks, Ty.
- [41:03]
A
Bye. Bye, Sam.