Gumloop Raises $50M from Benchmark to Scale AI Agents

0:00Cognilytica:Welcome to the podcast. I'm your host, Jaden Schaefer. Today on the show, we're talking about I want to do a startup highlight. It's a company called Gumloop. They just raised$50 million from Benchmark, one of the top tier VCs. And they're essentially their goal is to help every employee turn into an AI agent builder. So look at the tasks you're doing, look at the things you're doing. And how do we create AI agents to automate stuff you're doing, share those to other organizations, share them within your department around your company, to help everyone do more with the company. AI agents. And honestly, this is a pretty ambitious company, but I think they've actually landed some pretty successful plays.

0:36I'm excited to get into this because this is a company that was started in mid-2023, right? So we're talking like post-ChatGPT hype, and they were able to grow, scale successfully, and now they've raised$50 million. So this is going to be a good one

0:51Cognilytica:to get into. Before we get into that, I just want to say a big thank you to everyone. Yesterday was my birthday and I ask everyone if you hadn't already to leave a review for my birthday. And I wanted to say a huge thank you. I had tons of reviews. I wanted to read a couple of them. Some of them are funny. Someone said, I think Jaden does a great job of concisely capturing the latest AI oriented news. He comes across as a generally good guy. And I always look forward to his take on things. That is from Texas Saint. Thank you, Texas Saint for saying I am a generally good guy. I hope someday to become like a super good guy.

1:24But a generally good guy is awesome. I think especially in the field of politics, you probably see this on a lot of other podcasts too, but in the field of AI, like there's so much that touches into politics and ethics and so many areas where people have a lot of strong opinions. So inevitably, I'm going to say things that probably people don't agree with for a variety of different reasons other people agree with. So anyways, I hope that it's still insightful and educational and, you know, we're all learning about AI together. So I appreciate all of you guys coming along. uh one other uh one other i got a four-star review uh from matthew c code or matthew come on bro is my birthday you gotta give me a four-star review like that this is what he said he said provides a good overview of recent news and ai but can be very subjective and repetitive in some topics and comes across as biased for certain vendors uh as far as being repetitive it's kind of hard right because you have like google and open ai and anthropic and they all kind of come up with features and when someone does one feature they also copy it so i totally get that uh how it can feel repetitive sometimes.

2:22Also, as far as being biased to certain vendors, I will say,

2:26Cognilytica:not sponsored by any of the companies I talk about on the podcast currently, although shout out to OpenAI, I won't turn down a sponsorship deal. But in all seriousness, as far as being biased to certain vendors, I definitely, I can agree with that, I probably am, but it's typically vendors that I use the most and that I get the most use out of. So I mean, naturally, I'm going to be biased if I'm actively using them. And I do try to try all of the different AI companies. And so if you know, I'm just seeing that, you know, XYZ video generator, like VO3 from Google flow is just so much better than something like runway, for example, I am going to be biased.

3:02And that's kind of what I'm doing. But I do try to reevaluate because these things change all the time and give you guys the latest on on whatever's improving the most. So anyways, that's where I'm at on all those. I really appreciate it. If you have any comments or thoughts, feel free to share them, write a review. I'll probably cover it in the show if it's an interesting thing that I feel like applies to everyone. All right, let's get into what's going on with Gumloop. This is a really interesting company. It was started by Max Broder Urbass. He's one of the co-founders. And they started back in 2023.

3:31And I kind of gave you the idea, the overview of what they're trying to do. But basically, they're just trying to help automate some of the more tedious parts of a job with AI.

3:39Cognilytica:When they first launched, I think it was very very ambitious, and it was definitely a lot more experimental. And I remember seeing this company come out of the gate. It's, you know, something that was ambitious, but I didn't really know if they're going to be able to pull this off. I think that has changed a lot. I've been impressed. At the time back then, there's actually a bunch of other companies that were doing demos and stuff, and some of them did not make it. But Gumloop has done a pretty good job. They're used by Shopify, Ramp, Gusto, Instacart, Opendoor, a bunch of other of these big players are actually using them.

4:11and they essentially deploy AI agents that are doing complex stuff, multi-step workflows,

4:15Cognilytica:and you don't need an engineer. So instead of having to write the code, you can just build these agents visually, similar to what I'm working on with AIbox.ai, my own startup. And I, so yeah, no, you know, no shocker that Gumloop is doing really well. I see a ton of value and this is an area that is growing very fast as I'm seeing with my own startup. So according to Broder Urbass, because essentially what happens, like what he says happens inside of a company and why they grew so fast, he says that someone is going to build like one of the employees inside of a company builds one and then another team inside of the organization is going to like copy it and they modify it.

4:53Cognilytica:And then basically it just like spreads like wildfire inside of the organization. So it's kind of interesting, right? You, I think you oftentimes hear people creating a tool and then it kind of gets out of the organization and another organization uses it or a template or some sort of resource. And that's how a lot of these companies grow. It's interesting to see that for them, it's like one person in a company uses it and they basically are kind of the person that spreads it to the whole organization. He said that, you know, once someone like this starts building more agents, then suddenly the whole company becomes AI native.

5:24Cognilytica:That's a quote from him. So I think right now, you know, that vision of putting AI into the hands of every employee is what they're trying to do. And that's what drew Benchmark's general partner Everett Randall to them. He's the one that kind of spearheaded the investment into them. Randall also is new to Benchmark. He joined it just in October. He was at Kleiner Parkins before. So, you know, Randall's been at a lot of these top VCs. And he believes that, you know, everyday workers should have a quote unquote, AI superpower. So this kind of I think why they why they made this investment here. And of course,$50 million is a big lead.

6:03Cognilytica:They were, I mean, they didn't put the full 50 million in, but they led this round. It's their Series B investment. And this is also Randall's first deal over at Benchmark. So he's making his mark, and this is a cool, this is an impressive deal. The round had a bunch of other people in it. There was Nexus Venture Partners, First Round Capital, Y Combinator, Box Group, and the Canon Project, and Shopify. So it's interesting because Shopify is one of the customers that uses it, and now they're investing. I think we actually see this a lot with software, especially with these bigger firms. If they can go find a startup and they're willing to use the product, they like the product, and probably they're seeing a lot of, you know, internally at the company, it's saving them a lot of time and money and their employees are using it.

6:39They're like, well, that's probably going to be the case with a lot of other companies. And I've actually seen some investors, it's kind

6:45Cognilytica:of their like, it's their strategy to invest in their, invest in their like expense sheet, basically, that the companies that they spend money on, they invest in them. So one thing that I think is interesting. Gumloop wasn't actually actively raising money. But Broder said that, you know, the timing felt right. And Benchmark kind of reached out to them and said, Hey, look, we see you guys are growing, we'd be interested in helping you put together around. And Benchmark, by the way, is the company, I mean, obviously, tier one VC, but they've invested in like eBay, Uber, Dropbox. So they've been into a lot of these.

7:19And so I think because of that, Broder, who's leading Gumloop was like, Look, this is a big, you know, big firm, they obviously have a lot of great connections. And so he wanted to get in with it. Originally, his goal when he was building this company, though, Broder, he said that he was envisioning building just a 10 person

7:34Cognilytica:billion dollar company. That was his goal. He just wanted this to be like super lean. And, you know, with AI agents and especially because his company is an AI agent company, I think that would have been some incredible marketing to be like, look, we reached, you know, a billion dollar, billion dollar valuation and we've just done the whole thing with just 10 people and we've been using AI agents. There's an interesting report that came out. It's kind of went viral yesterday on X, which is that for like getting for getting Anthropic up to like over a billion dollar valuation, there was only one person on the go to market strategy team and he was just using and I actually think up until quite recently, it was one guy running all of their GDM.

8:12And he was just using tons of agents powered by Claude to do everything. And so I'm sure that's like kind of like a marketing thing but i think the guy leaked that out himself it wasn't like anthropic was trying to make a big point of it so i thought that was hilarious it's definitely possible more

8:26Cognilytica:than you'd think but of course uh according to broder in the case of gumloop he said demand from enterprise customers basically push them to scale more aggressively so they had to add a lot of engineers and a dedicated sales team which makes sense right and and i know it's like well why can't the ai do the sales and realistically people usually want to talk to a person on a call and not be sold by an AI agent that you can't really verify and put a face to. Gumloop definitely isn't alone, though, and I think chasing this kind of vision of turning knowledge workers into AI builders. There's a lot of competition, Zapier, N8N.

9:00Cognilytica:There's some newer, interesting agent tools that I've been looking at, like Dust. And, of course, I'm building one AI box. And so I think basically even the major AI companies are moving in this direction. Anthropic recently introduced Claude Cowork, which basically lets you create autonomous agents without writing any sort of code. And like ChatGPT is custom GPTs, which I wouldn't go so far as call those agents, but they are useful tools that can help you kind of automate some things to a small degree. Now, one thing that I think Gumloop is doing very well, and I predict this as a huge trend across all AI companies, is that they are model agnostic by design.

9:34Cognilytica:So basically, every time there's a new update and a new model starts kind of crushing it in the benchmarks and beating everybody else, I mean, typically we see kind of like a cyclical loop. somewhere between like chat GPT and Claude and Gemini and Grok and maybe another random one gets thrown in the mix every once in a while but we see this like every quarter basically cycling through who's leading in the benchmarks and so because of that different AI models are good at different tasks they perform differently Gumloop basically lets companies choose whatever model fits their job best at the moment and they can switch that out I think this is actually smart and it's kind of a competitive advantage that opening eye Google Anthropic they can't really touch because even though they're all going to make these AI agent tools, if you could get the same thing at a company that lets you switch and interchange between them all, there's a huge value there as different AI models are kind of going up and down in the benchmarks and they're edging out the competitors.

10:28Cognilytica:You always want to use the best tool. And some of them are better for different things. And to that point, if you want to get access to over 40 of the top AI models, everything from text, image, and audio, I would love for you to try out my startup, which is AIbox.ai. Not only do we have an AI agent builder where you can describe a tool you'd like to create and it will link together different AI models, fill out prompts so you can automate tasks that you do. We also have a playground where you get access to over 40 of the top models and you can chat with them all in the same thread. So you can start talking with ChatGPT, switch to Gemini, switch to Clawed in the same thread.

10:59Cognilytica:It sees the context and you don't have to worry about paying for subscriptions to a dozen different platforms. You get 11 labs for audio. You get a bunch of cool image generators. So you can go check that out at AIbox.ai. I'll leave a link in the description. it is only$8.99 a month to get started and you get 20 % off if you get an annual plan as well. So it's super affordable, saves you a ton of money. And to be honest, I have replaced my subscription to a bunch of different AI platforms with that, especially for the ones that I don't use like super, super heavily. But yeah, go check it out. Hope that saves you time, money, AIbox.ai.

11:31Cognilytica:Links in the description. Catch you guys in the next episode.