Coframe raises $25M Series A to make websites self-improving and agent-ready, citing 70% conversion lifts

about Cisco. Critical infrastructure for the AI era. Unlock seamless real-time experiences and new value with Cisco. We're going from the physical world back to the digital world with founder and CEO of co-frame Josh Payne. Building a living internet.

What's going on?

Get ready to go digital. Josh, welcome to the show. How you doing?

Good to see you.

Thank you so much. Doing great. How are you?

So, uh, you raised some money. How much did you raise? I want to hit the gong. I'm excited. It's warmed up.

25 million.

We got I didn't really realize we got this new lighting effect on the gong.

It's good. We got a replay effect. The gong is

as a percussionist, I can say. Great gong.

Great gong.

Boom. From from percussion. That's huge because every time we hit the gong, there's someone in the chat that always gets annoyed that we didn't warm it up, but I always say the first gong hit is the warm up.

It's still raining.

I didn't comment on your gong technique.

I commented on the gong.

Okay. Okay. Okay. That's fair. That's fair. Anyway, we're not here to talk about gongs. We're here to talk about co-frame. Tell us about the company. Tell us about yourself.

All right. Well, high level idea here with co-frame is that the internet is dead. uh you know typically it will take giant teams of people um we hate that it'll take giant teams of engineers, designers, marketers, etc. Um

to build and optimize the user interfaces that we use every day. Yeah.

And that's a huge problem. It's costing us hundreds of billions of dollars in in losses every year.

Um and we're flipping the script on that. We we see a world where

the internet itself can have its own sense of life. um where it can self-improve and become this kind of continuously learning system.

Um

and this actually has two sides. I mean first is on the the interfaces themselves, the websites, ads, apps, emails, etc.

Um that's where co-frame plays. And what we're already seeing is it also happens on the other side with personal agents starting to use the internet more. In fact, uh on the rise and starting to overtake human traffic on the internet, which is really cool. Um, and so we're building for that world and we're trying to bring the internet to life.

How how uh how possible is it to actually route an agent that lands on just a.com in a browser computer use all that uh into something that feels much more like an MCP experience like an API experience. If I'm a if I'm a website operator and an agent shows up, there are some re there are some situations where I don't want that to happen. But there are many situations where the website operator wants to give the agentic experience as much as possible and make it as easy as possible and make it not uh consume so many tokens for their potential customers because at the end of the day there's some they're they're selling something that's physical. They're selling something that that they don't mind if if an AI is buying or humans buying. They just want to sell. uh how how how how capable are you to actually route accordingly? How necessary is it? What does that look like?

Yeah, the the gap is is really not that wide. It just really comes down to establishing a protocol that everyone agrees on and kind of a standard operating procedure.

Um from a technical perspective, super easy. The tools already exists. You know, WebMCP already exists. There's personal agent protocol, which is a new proposed one that we're excited about as well.

Um when you go to any website in co-frame that is powered by co-frame.

Yeah. or own site for that matter and just add agent to it. Uh it will it will basically render an agent friendly version of it that can talk back with the agent which is really cool.

Um it's really just about like establishing that standard, right? And that's what we're trying to push for here.

Um

there's a ton of benefits and you know I guess like this is going to come down to like what what what are the what are the incentives here? Like why would a customer want to do this? Why would a business want to do this? Um you already mentioned token consumption. I would also mention speed like when you go and ask Muse uh which was amazing love Muse Natt Friedman who leads it at le love the whole team there.

Yeah. Yeah. Um but when you go on Muse and you know ask it to go book you a hotel or whatever you're waiting for a while and I think that that doesn't need to be the case.

100%. Yeah. Uh what is actual onboarding look like? Is this a a JavaScript snippet I'm throwing on the front end? Is it a Python package I'm importing? Is it something that lives in the GitHub and is reviewing code? as it goes out like how deep is the is the co-frame integration process?

Yeah, that's a good question. So, we kind of start out we try to start out really like seamless um with customers and that's by design. Um

we'll get in with like a a snippet that they just add to their site and it'll take days. There's this one fortune 500 company that um that we did a pilot with uh recently and have since converted to a big contract and they started out and basically implemented in 3 days uh from from initial contract to on the site and from there boom

uh the value there is that you're able to see

tremendous results really really quickly. So like with that customer we were able to see a 70% lift in conversion in less than a month

which is absolutely nuts. Like

what was going wrong before? I mean that's good for you, but it's like bearish for them. No offense to them.

This isn't isolated though either. I'll say like

they didn't have a buy button or something like 70% is sucking.

Yeah. What what like so when when you're talking to some of these customers is part of the pitch really like hey there's there's new companies out there that are really optimized for agents that are getting like the majority of their new c their new customers from agents and you will literally die if you do not don't like adopt this new paradigm. I think we've just seen a bunch of different examples specifically with like developer infrastructure products where if you're not like you know or people like shipping and paying for whole services without even knowing what they're paying for underneath the hood. Uh there was like a viral post where somebody's like like how did you ship your app and the person's like I don't know. He's like who did you pay? it was like codeex like

and so I feel like it's this is a big

um you know this will be like the probably defining trend over the next 5 years of like what businesses actually adopt this and which ones are like still think that humans are going to make these like the majority of end purchase decisions.

Yeah. What luxury retail

a lot of people are thinking about this.

It's like it's something that a lot of people are thinking about. It's not a core driver like it's not on, you know, it's not a lie down on a budget yet. Um, so this is really

for us, we see it as positioning and it's something that we've actually talked about and posted about for like the last, you know, two years plus. Like it's not it's not new either and it's it's not going to be on the budget for another at least year, I think. Um,

however, it's growing incredibly incredibly quickly. You know, the usage met metrics of these personal agents. I think by 2029 the majority of traffic in the US for commerce is going to end up being you know agent personal agents. That's my prediction. Um and so people are going to start thinking about this. Um but

when we go and talk to to companies to try to work with them,

it's it's more focused on what are the goals and needs they have today. um and just trying to have a practical point of view and positioning ourselves as a partner that's going to enable them to be ready when that transition happens because we do support this and we are going to be able to allow that seamlessly um when when it does

some of this uh some of this product the value that you're bringing it feels a little bit like old school enterprise SAS are you just doing subscriptions is there a consumption based pricing model because I'm sure there's AI involved but how are you thinking about positioning the uh the cost benefit to prospective customers.

Yeah. I mean, honestly, we probably should be doing uh performance-based pricing.

70%.

Yeah. Let me take half of that beginning. I mean, there's others like like Dropbox 158% lift. Like Replet 410% lift. Cash App 60% lift.

Whoa.

And these are all the ones that we talked about like many many many more. Like we're seeing a lot of great results. And if we were pricing on performance,

our revenue would be a lot higher. uh you know it's obviously you know um very attractive. Uh we're starting out keeping it simple. It's just like we'll come in try to evaluate the ROI going to be be able to provide you um usually targeting like 10x plus maybe even more and then we'll just have a flat rate for the for an annual agreement. Um

old school

and uh that's how we do it today. But I think down the road it's probably going to end up being interesting for customers. I could see a performance-based approach becoming really successful.

Yeah. Yeah. performance base would make a lot of sense. And then on your side, I imagine that you'll be grappling with like the well, if we throw a lot more compute at this, we might get better results, but then also if we just wait a couple months, like the models will get way cheaper and our margins will be better. So, there's a lot there there's a lot that you're going to have to puzzle out, but uh it's amazing that things are going so well. And congratulations on the on the series A. Glad to see.

How big is the team? How big is the team?

Uh we're about 45 people now. Um I don't know. I don't know about team size as being a key metric though. I think that what's more interesting is revenue per employee. So that's what I'm really pushing for.

Yeah. But but it's, you know, it's just helpful to get a gauge.

Yeah. It's just interesting to see in the modern where something like this grows.

Yeah.

Yeah. Exactly. Yeah. Exact, you know, this is said a lot, but

companies going public with less than 100 people, even less than 50 people, that's going to be a reality. Yeah.

Pretty soon here, I think with uh you know, the rise of what we're seeing with productivity and it's awesome time to be building. It is. Yeah. Thank you.

Next round. Nice.

Next round.

Yeah. Eventually, I mean, I think that this eventually ends up looking

essentially like a lab. I mean, we're, you know, one of the things that's been driving a lot of our growth is we're training this transformer based on all this data that we're seeing coming in.

Sure.

Uh, incredibly rich user interactions.

Yeah.

Obviously data that we're allowed to use and, you know, certain companies has agreements that we can't and so on. But,

um,

that that's that's turned out to be extremely powerful. Uh, and I think that this eventually all of our people are just going to be focused on making that model better.

Yeah. Amazing.

Thank you so much for great to meet you.

Great. Have a great rest of your day. See you soon, Josh.

Goodbye.

Let me tell you about CrowdStrike. Your business is AI. Their business is securing it. Crowd Strike secures AI and stops breaches. Jordy,

uh, Lulu, uh, is highlighting. Uh, so Shaquille said, "Note that Nvidia spokesman John Rizzo has put out multiple statements in recent weeks that we now know are false. Unclear of course if he believed them to be true at the time." John said Jensen's interactions with Miss Hang are not unusual or unexpected. And Lulu says, "No serious comm's person would ever lie like this. Rule number one is never lie to the press. Never lie to anyone. I assumed he believed this to be true. If so, he was deceived by his leadership and made to be a useful idiot and should resign on principle. Palmer says, "There is another possibility, Lulu. One that dominates my group group chats. Jensen is sleeping with all his customers." Um, and so when they say his interactions were not unusual or expected, they mean he is on a rampage seducing thousands of AI hardware execs with a glance.

Uh, anything's possible.

It is. Uh, it is a timeline and turmoil day for sure. For sure. Um, well, we'll follow the story more tomorrow and I'm sure there will be more updates on the investigation and what's going on in the stock market. Uh, fortunately, it feels like uh if you tried to catch a falling knife over the weekend, you're looking pretty good. You grabbed the handle. You didn't cut yourself because the market is firmly up today. Uh the Dow Jones up 2.15%, S&P 500 up 2.35% and the Nasdaq up 2.74%. So some some good news amid a lot of turmoil. But uh the last the last story, uh Richard Sutton, this guy, he needs to read this blog post. He says, uh brute force computation is not enough for intelligence. Clever new methods are needed. Uh the science of AI is far from finished. I I would recommend to this guy Richard that he reads this he reads this blog post. It's called the bitter lesson. Uh and it c kind of refutes exactly uh his his take here. Um but uh you know um you know maybe he knows something that we don't know. Of course he was the was the progeniator of the bitter lesson and it's it's interesting to see him not be bitter less and pilled in this particular moment but uh we will we will see where it goes. Anyway, anything else? Anyways,

chaos out there.

You're just looking at the timeline, all the Jensen stuff.

No, just just everything.

Everything.

Axe.

Oh, yeah.

Jensen.

Yeah, lots of stuff.

And it's just Monday.

Yeah, just Monday.

Start of the week.

Uh we will

be back in action tomorrow.

See you tomorrow.

We're live from the studio every day this week, which we're very excited about. Mhm.

And uh we hope you have the best Monday afternoon of your entire life. It is possible. I know it sounds fanciful, but it's possible. So just go and do it. And without further ado, flashback.

Leave us five stars on Apple podcast and Spotify. Goodbye.

Love you.