Dhaval:
This founder built an AI product to replaces your employer with ai. Yes. You heard that right? Your employer, James, is our guest in today’s show, he shares his learning on how he built a product in a hyper-competitive market space. James Clift is a founder of durable. Durable, makes owning a business easier than having a job. He’s a former founder of KarmaHire WorkStory, VisualCV and Holopod. He has been building businesses since 2005.
Welcome to the call. James, tell us about your product.
James Clift:
Awesome. Yeah, so Durable is the fastest way to build a website on the internet. In three clicks, in 30 seconds, you can generate a business website using ai. And not only can you make a website, we’ve got the rest of the stack as well to operate a business, we’ve got a CRM, an invoicing tool a financial account. So essentially everything you need to start and grow your business in just a few click. .
Dhaval:
Wow, that’s very powerful vision where do you sit in your market space? are you serving a specific customer segment?
James Clift:
Yeah, so we’re totally focused on solo operators, so anyone that runs a solo business. So primarily those are service-based companies, so everything from a web designer, marketing contractor, copywriter to more traditional physical service companies like Lawn care, home Services. Plumbing contractors, skilled trades. So essentially anything where you’re trading your hours for dollars or your hours for projects we’re a great fit for if you’re selling goods on the internet or you have a brick and mortar store, we’re not the best solution there, but solo service-based companies is our primary market right now.
Dhaval:
How did you differentiate yourself in this crowded market space? I believe it may be high competition market space. But You have, you seem to have found reasonable amount of success based on what I have seen about you online. How did you create that differentiation for your product?
James Clift:
Yeah, I think there’s a lot of ways to look at markets like most markets are very large on the internet, and for us it was a few things. So one is bundling, so providing all the tools you need to run your business under one login. So that was a big value add from the start. So you don’t have to learn five different tools. You don’t have to pay for five different subscriptions or 10 different subscriptions. It’s everything you need under one platform. And then the other piece is what can you actually do 10 times better than everybody else? And for us it’s the speed of actually getting a website. Out to market. So instead of taking, typically it’s weeks to get a website live, if you’re really good it’s days. If you’re really, really good, it’s hours. We actually do that in minutes. So there’s this order of magnitude that makes that thing faster, or that business process faster. And it actually unlocks a lot of creativity, a lot of, just makes it more fun and playful, not stressful. And I think you open up these brand new markets by just anytime there’s an order of magnitude step change. Something’s 10 times faster, 10 times better, 10 times more intelligent. That creates these huge opportunities. And I think AI as a platform is definitely one of them that we’re seeing. yeah, I think customers are super excited about the speed, the simplicity, and then the bundling aspect of the platform as well.
Dhaval:
Very interesting. So you brought this up, AI and this particular discussion and discussions like. That I host are focused on people who are either interested in creating an AI product or infusing AI in their existing product. Tell us about your infusion of AI into your product. When did you decide that, was it AI first from the ground up? And if it wasn’t, when did you decide to bring AI into the user?
James Clift:
Yeah, I think, so I’ve ran SaaS companies for a while now, so probably about 15 years. And, it’s always, I mean, the goal of any software company is how do you make processes easier and make your products easier to use? So, the long-term vision of us and AI is really, How do we replace your employer with AI and just let you focus on your core competency. So that’s your skill, right? So a lot of the time if you have a service job, you have an hourly rate that you’re then getting marked up for by your employer. So, in a perfect world, you just meet that reach, that market demand. So, hey, you’re a lawyer making, I don’t know, call it 500 bucks an hour that your employer charges you out. You’re making 200 bucks an hour. So that’s a market opportunity for you to go independent and do your own thing. But what the law firm brings to you is a brand customers, some back office services. So the way we’re thinking about that is what can we actually replace? And I think brand is changing a lot. Like the brand matters less, the individual matters more, and the back office piece can be solved with technology. So essentially . Everything can be automated except for the thing you’re really good at, and that’s really how we’re thinking about ai. So from a product standpoint, we built the platform first, and then we built the AI on top of the platform because we’ve got a lot of features and it was always this idea of, how do you make those features easier to use, more accessible, more interesting, and just more intelligent. So the, our customers can just focus on what they’re good at. so that’s always been part of the strategy. Definitely it’s accelerated in the last few months here with all these new technologies and APIs and libraries that have come out, and been super incredible and are moving really quickly. So definitely, the primary part of the strategy moving forward as well.
Dhaval:
One thing, one thing that I always hear from other product creators in the space is about finding the balance between building on top of existing AI capabilities. That other companies have created versus building your own AI capabilities? Where do you draw that line in your product?
James Clift:
Yeah, I think it really depends on what kind of company you want to be. Are you really good at marketing? Can you repackage these libraries and build a good user experiences around them? And you can accelerate really quickly? Are you deep technologists? in that case then you should build the underlying infrastructure layer. For us, I think the advantage, and I think. These core models are really, really powerful, but I think you have to train them on your own data sets. Otherwise you’re gonna lose your competitive advantage really quickly. So if you’re just re-skinning chatGPT and it’s a slightly different user experience, but the same data, That’s gonna be a race to the bottom pretty quickly, because everyone can do that. But if you have a user that is unique that you can build data sets around, then you can train these models to be more effective. So we’re doing both, we’re using the existing models, but we’re also training them. Pretty specifically around our category of customers. If you think about a solo business owner, there’s a set of activities that you need to do. Even from a marketing standpoint. It’s okay, you’ve got your website. How do you optimize your seo? How do you create your ads? How do you create your marketing copy, your newsletters? Once you have customers in, like, where do you get more customers from? How do you measure your channels that are effective? When you send invoices, what is the value of that invoice? How does that tie to your accounting system and your customer database? So there’s just a lot of things that we can actually build more proprietary, unique data sets around and workflows and processes that we can optimize. So I think as long as you own that customer journey and lifecycle, then you have the ability to train your model and make it better. But yeah, if you’re just re-skinning an API, I think. There’s some that will do well. So like re-skin the API in the category. I think one or two of those will succeed in every category because it is great tech. And if you’re good at marketing and acquiring customers, there’s opportunity there. But if you’re the, I don’t know, the 20th company to try and build a copywriting app with ai, I think that’s gonna be more challenging.
Dhaval:
Thank you. Thank you for sharing that within your customer journey. Where do you bring ai? Is it from the get-go or is it, how do you find that balance to delight the customers versus, make them anxious?
James Clift:
Yeah, right from the start, I think, It’s just a really cool technology to see, and from what our customers see when they’re getting their websites built by us, you can actually see the copy being written and it’s always wow, this is surprisingly good and really cool. So I think it’s just like exposing them to the magic of this technology as early as possible, but it’s not, but not trying to make it inaccessible. Right. I think the beauty of ai. Is that you see the output, right? It’s not scary because you can see, oh, this is just copy. These are images you chose. This is a marketing strategy that you came up with. It’s more just impressive than scary. and I think, yeah, for us it’s really giving them that front and center and the first part of the user experience that’s really delightful and giving them something to work from that then the human, I think humans are a lot better at editing than creating. A lot of the time. So give them something to play around with, to edit, to regenerate, just to make it fun as opposed to this tedious thing that you have to learn. It’s more of a playful experience, I think is what we’re really trying to build for our customers.
Very insightful:
So making it playful is one of the ways to pull customers in so that they, and also making it transparent so that they see it happening in front of their eyes and it’s a magical experience rather than. Behind the scenes, here’s what we are doing, take it or lose it, or any of that. One thing, one thing often comes up is the thing that you brought up is fine tuning on top of existing LLMs. Do you have any tools for other, do you recommend any tools for other AI product creators in this space who want to embark on that journey? What are some of the quick ways that they can fine tune on top of existing LLMs?
James Clift:
Yeah, I think that’s the challenge now is that space is so new that, tools are popping up every day, but nothing’s really. Taken over as the category leader yet. So I’d say just the more you actually, I mean, prompt engineering is a career now, so being able to figure out how to write good prompts for open ai, I think is a really interesting way of thinking about it. We’re building a lot of that in-house as well. Like our own like model optimization is all gonna be in house at this point. I think when there’s third party services, we’ll definitely look at them as well. But I would say just use. Underlying technology. and then under, the more you understand it, I think the more you can optimize it as well. So play around with it, test it out. If you wanna read the papers, they’re actually really interesting as well. I think from a. Like a pure software product. There’s not a lot that I’ve seen yet. I think it’s all coming out and people are building it. There’s prompt analytics and stuff as well. That’s really interesting. But for now, I think, yeah, that’s where the expertise comes in, right? Is how deep can you go and how deep can your team go on these different models and training and all that.
Dhaval:
Now as a founder, are you heavily involved in the product at this point or have you gotten to a point where you’re focusing more out, focusing out? Or are you focusing in on the product and you have a separate product manager leading the product?
James Clift:
It’s, myself and my team leading the product. So we don’t have any product managers or project managers right now. So yeah, we’re still really heavily involved in all product decisions. We’ve got an incredible UX and product designer, incredible engineers, incredible marketing folks. So I think it’s more, it’s an, all the whole team’s working on it right now. I think it’s led by, the group right now and at the end of the day, it’s me that’s making the final decisions on product here.
Dhaval:
Any learning lessons you wanna share around fine tuning and prompt engineering when it comes to building on top of LLMs. How has that experience been with your engineering team and what do you recommend others to, what are some of the practices you recommend other PMs or founders to model if they wanna establish that strong partnership, with their Technical partners, to fine tune and create a strong AI product that creates differentiation. What are some of the best practices for that partnership? What does that look like?
James Clift:
Yeah. I think it’s really just having this culture of experimentation and iteration and testing. It’s really just volume. How much can you test different models and outputs and questions and, gathering feedback from users as well. So having a mechanism to gather feedback around those prompts. And then having your data. So if you say, Hey, we’re. We’re doing this right now as an exercise. we’ve got thousands of categories of business being created and we’re saying, okay, we’re really good at creating lawn care websites, but we’re not so good at personal trainer websites. Why is that? Maybe the images we’re choosing are off, or the prompts we’re using don’t apply as much. So it’s really just training your own data set and having that data available and also the simple answers look at the output and decide if it’s good as either get the feedback from the user, get the feedback from yourself. But I think being, pretty honest about like, okay, this is actually good or not. If not, what can we change and I think that’s what we’re finding, with some of our categories. Like, wow, this is totally off. And then you really dig in and evaluate what went wrong there, what you can fix. But really it’s just volume iteration. And just, yeah, playing around with the outputs and the options. It’s pretty much it. Put in the work basically. No simple answer. ,
Dhaval:
tell us a little bit about the size of your team and a little bit about where you are in terms of product. Have you raised capital? Are you bootstrapping? Are you looking to raise capital? How many users you have? What is your revenue like, and what is the size of your team.
James Clift:
Yeah, so we raised capital in May. We’re eight people now. Raised a 6.2 million seed round from some awesome investors. We are at about 200,000 websites built, in the last three months. So we launched in October, and that’s just been doubling month over month or tripling. So yeah, lot of users, lot of, a lot of, we got some revenue coming in the door now we’re not disclosing that. But it’s growing very rapidly as well. So yeah, all systems go right now. And the plan is to grow the team right now as well. I’ve got a few new hires coming on board pretty quickly, so lean team right now. But yeah, definitely operating very efficiently and shipping product very fast. Right now,
Dhaval:
are you a hundred percent remote hybrid? Where are you hiring?
James Clift:
Yeah, we’re remote by default. So we’ve got, team members in South America, Vancouver, San Francisco. And then we do have small offices in those locations as well. So kind of your hub so you can come in, do work, or work from home as well. So we design our whole culture to be remote first, so a lot of async, not too many meetings, but some crossover time for collaboration. and then, yeah, we’ve got a space you can go into and work, but it’s not required it’s optional
Dhaval:
as an AI product creator, did you have that AI domain knowledge or did you find partners who could fill in your entrepreneurial gaps for that area?
James Clift:
Yeah, it’s definitely, it’s always been an interest, but not an expertise. So it’s something that I’ve like it’s been a hobby, but not part of any business that I’ve done before. So it’s really just, yeah, finding the partners and the technologists in the community that, is on the cutting edge of this stuff. And yeah, there’s a few out there that are great. and I think like one, one of my early investors is called South Park Commons. So they’re a community of pre, before it’s a zero to one stage. So before you actually start a business, it’s a place to experiment with different ideas. And that’s a really strong AI community that we’ve really drawn on for hires and expertise and just learning as well. So I think having a community to learn with is super valuable. and just, yeah, again, being curious and being obsessed with it to some degree, always comes in handy. And I think if you’re gonna bet on a technology, it’s a pretty interesting one for the next big cycle of software and technology and yeah. What the world might look like.
Dhaval:
What was that process like for you to bring someone with an AI expertise on board? Did you meet them at South Park Commons, by the way? Great Community. South Park Commons is the. Where I have gotten a lot of guests from, and yeah, great community. So is that, would you say you, you attracted your technical partners there or what does that look like? How did you pull them into your vision?
James Clift:
Yeah, I think it’s just, again, volume, like having as many conversations as possible. My CTO I knew from my last business as well, so I worked with him before. And he’s got his own network as well, so I think when you get one. Really excellent person. They typically bring more excellent people. So that’s the value in hiring with and partnering with amazing people. And then just, yeah, getting the, doing all the stuff that seems not that important. So going to the demo, like the little demo days, like getting your name out there, just getting as much exposure on to what you’re doing. people who are interested will reach out as well and want to communicate.So not building in a vacuum, but really. Really getting your name out there, putting yourself out there, showing your work. So even if it’s not totally done, just ship something cool. And that leads to a lot more than you’d expect. It’s the, you can’t predict it, but typically the more you’re out there, the more people will discover you and the more conversations you have, the better people that you will work with that will buy into what you’re doing.So it’s not trying to convince someone of what you’re doing, but. have a good thing that you’re doing and talk about it, and then people will be convinced. The right people will be convinced anyway.
Dhaval:
Yeah. There is, I think there is this famous quote from novel, which is if you share what you’re working on, you’ll automatically attract people who are interested in that.
James Clift: Exactly.
Dhaval:
Very interesting. Very interesting. What I wanted to ask you next is the future vision of your product. James, where do you see, where do you see your product going?
James Clift:
Yeah, I think I touched on earlier, but essentially, can we replace your employer with ai? So how can we abstract and automate every part of running a business except your core skill? I would love to see. Someone who has a skill sign up for durable and they have a job scheduled in their calendar tomorrow, they show up, they do the work, their invoice is automatically sent and paid for their accounting’s automatically done. So you have this total control over your schedule and flexibility. We tell you how to price your service how can you make a business as easy as renting your house out on Airbnb? So really just leaning in on the automation and the AI and just letting someone focus on their core skill. And there’s so many problems to solve there. So we’re starting with the, the marketing piece. So even from a, like on a website builder standpoint, there’s a hundred features we wanna build to make that process much better, much easier. And then once you have your website, how do we help you market? So automating your advertising, your customer outreach, your marketplace connections, your lead generation, your follow ups, then it goes into scheduling. So we’re kind of working our way down, the operations stack of a solo business. But yeah, the end goal I think is really, and what we’re seeing in the world is hundreds of millions more people that are, that are their own boss, right? They’re running their own business. They’re more in control of their lives, they’re pricing, their freedom, their flexibility, and they’re living happier lives because of technology. And that’s really where, what we wanna build for.
Dhaval:
Thank you so much for sharing this vision, James. I wish you all the best in achieving that vision. I’ve done service businesses and I know how difficult it is. You are doing some amazing work there to help individual service business owners achieve their full potential, right? That’s where the innovation comes from. So thank you for taking on this vision and thank you for coming onto this podcast and helping us learn from you. Thank you. Yeah,
James Clift: Thanks. Awesome. Appreciate it take care.
