Venture with Grace
00:40:24 · Grace Gong
[00:40:24] S0: How do you feel like the... Each engineering team should kind of allocate their budget after looking at all the data?
00:40:32 · Vitaly Gordon
[00:40:32] S1: Yeah. So I would say there are two extremes, right? One is look, look, sometimes you're you're, you know, a large organization with a very, you know, strict budgeting and you just need to kind of like, okay, you have a very specific AI budget, You cannot go beyond it. And I would say that if that's the thing, you should just now thinking about within the budget, how do I maximize out? And I'll tell you, like, the best way to maximize outcomes is actually to make sure that you don't run out of budget. And let's say if you have a monthly kinda quota, don't run-in the middle of the month because it's very, very hard to go back from, like, using AI to not using AI. Like, most people will, like, probably will, like, do nothing and, you know, and that's kinda the weird part. And so it's probably better to, like, you know, have them use, like, much, much cheaper models, but have them use AI for, you know, the entire period of the month as opposed to, like, just have them to use kind of, like, a more expensive models. Then in the middle of the month, like, they just, like, have no access to AI. Right? So that is... And again, these are, like, strategies that we also work with customers on, like, how to detect that, you know, the forecast of using it goes above their, you know, quota. And then, you know, once it happens, we kind of, like, can redirect some of work to cheaper models to get them back in line so... To make sure they stay on on budget. The second thing, which is, you know, more common, would say, with AI native companies, is, like, as long as you actually, like, producing, you know, great outcomes, you should do more and more. And so that is more about, like, I would say, budget allocation. Like, you... If you have teammates that are just, like, you know, crushing it and they're, like, you know, every dollar that you give them on AI turns into, like, 10 features, you should give them more budget. Right? Like, that's what we should be... What you should be doing. And so, like, this... But but, again, all of these things on large scale, like, you just cannot manually do it. Like, so we kind of, like, created this, like, fully automated system that automates all of these decisions, you know, for, for the companies. And so... And there's, like, you know, some variations in between those two extremes.
00:42:40 · Vitaly Gordon, Grace Gong
But, like, I I would say that, you know, on one hand, you know, maximize outcomes, but stay within budget. On the other is maximize outcomes and add more budget, as needed as long as you continue getting the outcomes that you want. [00:42:53] S0: When you're thinking about, I guess, the competitive advantage between you guys and some of the other players in the similar domain, where would you say is a place that you actually want to build and mold on? Since I feel like one of the things from my understanding, it's in the space, obviously, I think I would assume it has a lot of stickiness as you have to have access to the entire code base and everything's tooling and everything else. But before you enter, what is the journey looks like for you guys to identify the client that actually needs this product? And then is there any clients or is there any specific type of AI company that just don't need this at all?
00:43:42 · Vitaly Gordon
[00:43:42] S1: Yeah, look, are definitely companies who kind of maybe figure out and there's, like, you know, they're very, very specific way of working. Let's say, you know, ramp just released their harness that is, like, highly, highly optimized for the way that ramp works. And by the way, there's, like, a significant number of people that are... Their sole job is to focus on, you know, this thing, and that's that's what they're doing. And so for, you know, those companies, I would say, you know, we less focus on because, obviously, they, you know, they already figure a lot of things out. So we should focus on the companies that need more of our help and not kind of like the companies that very large teams that are, you know, solely focused on solving the problem. Right? So that's kind of like to your question now. Now... But in terms of like what is the what is the competitive advantage is like, look, we've been doing it for quite some time. So we... And the... You know, our argument is is like, look, there is, like, endless advice on the Internet on how to use AI. Like, the problem is not like the... Like, there is not enough advice. Right? There is like... The problem is, like, knowing kind of out of all of that, like, I would argue 95% of at least is complete nonsense. It's completely useless. It's actually, like, more, you know, like, more harmful than useful. Right? And so knowing that what are the, like, the 5% that actually, like, can improve yours and given that, you know, if you actually know it and you don't wanna spend the time, like, sure, you can go and try every single one of these things Or you can, you know, part of a organization like ours that, you know, just we not already tested a lot of these things. Like I said, we like literally run an organization that tests like so many different variants and test every single thing that, you know, is coming in. The other thing is also we are automatically can know how to apply, you know, all of our kind of learnings into your specific organizations. And so I would say probably, you know, you will save a lot of money on, like, focusing on your core business and not trying to, like, focus on, you know, kind of like, what is our core business. And so we're very... We think we're the best at what we do.
00:46:00 · Vitaly Gordon, Grace Gong
And so, you know, like, usually, it will be hard for an, you know, an organization that, like, has a completely different, you know, mission that... To become better than than us at that. Yeah. [00:46:12] S0: Is there a particular, like, kind of, like, data that you would demonstrate to the client that, like, oh, with us, you will save x amount of dollar or like, is there any kind of like standard metrics that you guys typically like kind of like demonstrate? [00:46:27] S1: Yeah, we're like, literally, it's just like we we show them that we're, you know, we... I think there's actually something that is coming up. And there was a case study that is just, coming up on it, but that we, work with an organization. And I think if I recall correctly, their the, you know, their outcomes increased two and a half, x, while their AI, budget actually went down by, like, 20% or something like that. So, like, in terms of, like, actually, again, price per outcome was, like, massively, massively different result, just by, kind of, like, you know, using our system. [00:47:05] S0: Speaking of using your system, so who actually use the platform the most? Would it be the CTO, CIO or CFO? Because I could also see the use case of CFO looking at what's happening here. Yeah. And like kind of cut down the causes and something then.
00:47:05 · Grace Gong, Vitaly Gordon
[00:47:05] S0: Speaking of using your system, so who actually use the platform the most? Would it be the CTO, CIO or CFO? Because I could also see the use case of CFO looking at what's happening here. Yeah. And like kind of cut down the causes and something then. [00:47:24] S1: Yeah. So it's all the above that you mentioned as the CTO, CFOs and other like it's also like developer productivity teams or the people who you know, the AI budget. Like, think about there's like, in every organization, there's someone who's like goes and, you know, closes a large contract with like, whether it's Anthropic or, you know, OpenAI or Cognition or Cursor. And at the end of the day, like, they they wanna see and they wanna show to the CFO the results of that investment. Right? It's like... At the end of the day, all of these kind of kind of contracts are investment in order to achieve a certain goal. Right? There's like... You know, they're not buying those contracts out of the goodness of their heart. And if they can show that, you know, that investment made a large difference, that's not good. And so... By the way, that's not good not for the company, not for the vendor, and that's thing that we help with.