Why Digital Ownership Matters More Than Ever with lilAgents

[00:00:00] The leading issue of agentic AI in businesses right now is ensuring agents act with compliance guidelines. And Denodo applies guardrails across your entire data estate. By aligning your company's data infrastructure under one system, these guardrails perform consistently across your platform. So start scaling your business and start with Denodo. Simply visit denodo.com to learn more.

[00:00:33] What if the biggest risk to your business isn't choosing the wrong AI tool, but building your entire business on platforms that you never truly own? Well, whether it be websites, social media platforms, or customer data and marketing systems and online stores, many companies have unknowingly traded their ownership for convenience.

[00:00:58] And my guest today will explain why he believes AI offers a chance to change just that. His name's David V. Kimball. I don't know why I felt compelled to do that line from The Fugitive there. I think that was Dr. Richard Kimball, right? But seriously, he is the co-founder and CEO of a company called Lil Agents.

[00:01:19] And he will explain how businesses can use agents to reduce cost, escape vendor locking, and build technology stacks that remain flexible as their business grows, and do it all without sacrificing ownership. And with that scene perfectly set, let me introduce you to David right now. So thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do?

[00:01:49] Sure, Neil. Yeah, thank you for having me. My name is David V. Kimball. I am the co-founder and CEO of Lil Agents, which is an ownership-first AI marketing agency. And the ownership-first just kind of means we like getting people on the portable stacks that they own without any of the vendor locking. And I've been doing stuff online for about 20 years now. Awesome. And I'd love to find out a little bit more about that.

[00:02:11] But before I do, one of the things that put you on my radar and set off my tech-spidey sensors was I was reading that you've argued that many small businesses right now don't actually own their digital presence. It's such a good point because they're effectively renting it from other platforms. But the question I've got to ask you is, how did we get here? What risk does that create when a business becomes dependent on a single vendor? Well, it kind of becomes a hostage situation, and they get you in the name of convenience.

[00:02:41] You know, it's like, hey, we're the one platform for everything. You can, you know, host your website with us. Why not just host your domain name and your email provider? And on the face of it, it seems like, well, that's great. I'm probably saving money. You know, it's one place. It's a lot more convenient to go to. And then what you realize is the price keeps going up. It has this habit of doing that every year. Or maybe a feature you really relied on gets removed unceremoniously without warning. And you go, hey, this is like not worth what I thought it was.

[00:03:11] This isn't the value that I thought it was. And so the pain of trying to move off of that, that's the degree of how much control they have over you versus you having control over it. And it's unfortunately pretty common. And so the idea is the more you have under your control, the more you can say, oh, I'm going to like swap out this component. It's kind of more modular. I'm going to swap this out. I'm going to use a different vendor. And you have control. But that unfortunately is not the status quo. So yeah.

[00:03:38] And I think if we go back 15 years ago, a lot of people abandoning their websites to have a Facebook page. And we all know how that story ended because it's just one tweak of an algorithm, one feature here, one change there. And you've lost everything that you've built, your audience and everything, don't you? That's exactly right. If you want to compare apples to apples, if you move off your website, which you own now, of course, there's aspects of that that you don't.

[00:04:05] So there's a sliding scale of ownership where you might rent parts of it and own some of it. When you go onto Facebook, you own basically nothing. You can't even reach the people that allegedly follow your page and they can just completely squash you. And fast forward to present day and AI is now being added to almost every software platform right now. So from your perspective, where is AI genuinely helping businesses save time, improve outcomes? And where is it mostly marketing hype?

[00:04:34] Because there's a fair amount of that out there as well. That's right. Well, I think it's pretty easy to tell when it was kind of an AI chat bolted onto something. It's not to say it's not helpful. Maybe it's really good at like information retrieval or you can summarize something, but that's kind of a very narrow way of looking at what AI can be used for. So I say a lot of the hype is around, yeah, we added like a chat GPT add-on to our service and great, now we're going to charge you more for it.

[00:05:01] Whereas, you know, if you have something that's more sophisticated, that can be really helpful. So for example, maybe you've heard of an AI harness, which is something like Claude Code or like OpenClaw or Hermes. Now these things kind of tilt towards the more experimental, but they are very powerful for actually doing things for you. The way to think about it, in my opinion, is, is it just a chat bot where that's kind of disposable and then like every time you have a new problem, you've got to create kind of reinvent the wheel every time?

[00:05:30] Or is it an AI that can literally do things on your behalf? Can it do things while you're asleep? That's where I think, you know, AI is really helpful for small to medium sized businesses and frankly, any individual contributor or manager for that matter. And so like an example would be an MCP server or an API. Can your AI make use of those things? Can it interact with services? Can it build tools? Can it create files? That's the kind of thing that I would be thinking about when it comes to meaningfully moving the needle forward with AI.

[00:06:00] And I know one of your specialities is helping companies use AI agents to work around limitations inside platforms like Shopify and HubSpot, for example, to name two there. But what are some of the real world examples where you've seen AI agents allowing businesses to solve problems without replacing their existing systems? Because again, there's a lot of hype out there. And hearing these real world stories, I really think, go a long way to helping inspire people listening. Oh, totally right.

[00:06:30] I mean, Shopify is an interesting case because, you know, one of our clients has a 5,000 shop, a 5,000 product Shopify shop. And that's not trivial. And that's been building for over 10 years. And there's a lot of baggage on this website. And so one of the things that we wanted to do is fix the metadata. One pretty common problem with e-commerce websites, especially ones that are kind of owned by smaller businesses, if they want to rapidly get new products out there, what they'll do on the user interface side is they'll duplicate an existing product.

[00:06:59] And then maybe they'll change the title and the description, the photo. But then there's a bunch of fields they forget to edit. And so this kind of carryover baggage from like maybe the title is correct, but the meta title is talking about a different product. And then the meta description is completely inaccurate and out of date or something like a t-shirt size is wrong because it's a remnant of being duplicated. And so this is amassed over the last 10 years for this particular client. And guess what?

[00:07:27] Shopify has no bulk edit option. Shockingly, it has no bulk edit option after all this time. And so what I did was I used cloud code and I connected the Shopify API to it. And all of a sudden that all changed. I could say, hey, look at the title of every product and where that's inconsistent with the meta title that gets sent to search engines that get sent to AI chats and agents do a diff

[00:07:54] and figure out what it should be and a mass update everything. And so it kind of did a dry run. It kind of gave me a little preview of what it would be doing. And of course, I backed up everything beforehand. But maybe we'll get into this a bit more with guardrails later on in our discussion. But you want to do really careful backup and recovery steps, especially as you give more power to these agents. But all that to be said, it made these mass updates.

[00:08:18] And we got a 60% increase in revenue from organic-based traffic over the last three or four months or so after making those changes. And it was very quick to do. And it was very easy compared to manually logging into Shopify and trying to make those changes. And you're a man after my own heart there when you quickly said, I made a backup first. That belts and braces approach to all things IT and tech is so important.

[00:08:45] And of course, we will have many people listening that maybe feel trapped between expensive software suites on one side and then disconnected point solutions on the other. So for those people listening, how should they be thinking about building an AI-powered technology stack that ultimately remains flexible as their business grows? Because so many solutions are incredibly rigid there. But any advice?

[00:09:10] Yeah, I would say it kind of goes back to what we touched on earlier with the ownership kind of being a sliding scale. You're going to rent some aspects. You're going to use some SaaS platforms. The real question is how trivial is it to swap those out with something else? So if you've got like 500 pieces of software, you probably don't need all of them. There's probably crossover. There's probably one platform that's completely over-engineered for what you need it for.

[00:09:37] And when that's the case, you can probably use AI to either build a tool that can do the repetitive thing for you or bridge the gaps between tools that otherwise don't easily talk to each other. Those two use cases alone are a huge differentiator. I would also say you don't necessarily need to use Cloud Opus every time or some crazy big frontier model to do all your heavy lifting. In some cases, you can either create a JavaScript file or a really basic software tool or even a web tool

[00:10:05] that will just run on repeat and use no token. So this isn't about, oh man, I've got to like give up a SaaS tool to pay Anthropic or to pay OpenAI necessarily. And even then, if you want kind of some agentic work on the reasoning front, you can actually outsource to dumber models that are less expensive. So for example, you could use Haiku. You could even use other models completely that are not Cloud or OpenAI like Minimax is really good. Kimi is really good. GLM.

[00:10:33] So these Chinese models are really great for some of these like more menial tasks. Things like scan my emails and find a PR opportunities is a great Haiku or Sonnet task. It doesn't have to be your massive expensive token vaporizing model. And one of the things I love about what you've done here is you've helped companies move away from locked ecosystems and put them on infrastructure that they actually control.

[00:10:59] And I'm curious, from the conversations that you have with people when you do this, what's the biggest misconception that founders might have about digital ownership? And why does it matter in this AI-driven world more than ever? I think there's a sense that if business owners have a login to something, they feel like they own it. So if they have an account, they go, okay, this is mine. Now that I would say that is one step improved from, oh, my agency handles that. That's definitely an improvement from that.

[00:11:27] Because in a lot of cases, some of our clients move to us from someone else. It's like, well, where do your Facebook ads live? Oh, it's with our agency's account. Or, you know, well, how do I get to your domain? Oh, our agency handle it. That's definitely a position you don't want to be in. So if that's the situation, I would make sure your account is the one that houses your domain or your hosting, things like that. That's kind of bare minimum.

[00:11:51] But then beyond that, depending on the platform, even if you log in and it's under your account, it's not really ownership. Real ownership means you can walk away at any moment, take your site, your data, whatever it is with you. And so I'll give you a brief example. So Bluehost is kind of one of these legacy we'll do everything for you platforms. And one of the things that we tried to do was a client had a website and a domain name registered with Bluehost.

[00:12:18] And all we were trying to do was move the domain name and decouple it. We still wanted the domain name to point to the website that Bluehost was hosting. So it was like a WordPress site that we were going to get them off of eventually. But we wanted to just move the domain name. So the name service didn't change. The DNS records didn't change. It was pointing to the same place. We just moved it from Bluehost to Namecheap. We okayed it. The transfer occurred. And then Bluehost just killed the domain name on its side. So their email went down.

[00:12:46] The website went down for several hours. And we had to say, what happened? We checked Namecheap. It pointed to Bluehost's name servers. So it wasn't Namecheap's fault. We went into Bluehost and they just removed association with their domain without warning, without prompting. So we had to go in and frantically re-add it. Then we had to wait for the SSL certificate, which took way too long. Eventually it happened. And then their email and everything was back. And it's like, Bluehost, why did you just disconnect the domain?

[00:13:15] All we did was migrate it. It was still pointing to your name servers. So baffling things like this occur all the time. And it's because it's all bundled into one thing. And they can get pretty lazy with how they implement things when they do that. And so again, luckily, we were able to eventually get them off of WordPress even better. We got them onto a modern stack that's a static site generator. And they're sitting pretty now. And their monthly usage went to $0 a month from what they were paying before for what we did for them.

[00:13:45] Oh, what a fantastic story. And I'm curious, when you audit a company's maybe marketing or operation stack, what are the most common subscriptions, platforms, or tools that businesses are paying a lot of money for, but seldom extracting full value from it? Do you come across the usual suspects every time? Yeah, I would say so. I would say in general, your website CMS is completely over-bloated and over-engineered for what you need.

[00:14:12] I'd say 95%, especially small to medium-sized business sites, it's a glorified pamphlet. Maybe there's a blog or a form or something like that. But those can be static sites. They don't have to be expensive Squarespace or Wix or WordPress platforms that you're renting. You don't even own them. Like, what's so great is if you use something like Astro and you pair that with Netlify and GitHub, your hosting is free. You're getting something far more performant.

[00:14:39] You're going to get amazing scores on page speed, insights, and GTmetrix, and all these evaluators of how performant your site is. It's going to look better. And you can use AI with it far more easily. It's a huge barrier right now to try to use AI to manage your WordPress site or your Wix site. Of course, they're scrambling. You know, Wix and Squarespace are scrambling to add AI features in. And I'm sure Webflow will eventually as well. But ultimately, nothing has really changed. You're paying a small fee every month to not own anything.

[00:15:08] So that's the first area. The second area to think about is maybe you're on the right platform, but you're on the wrong tier. So for example, with HubSpot, maybe you're overpaying for your marketing contacts. And whereas like one Klaviyo client we had, they used Klaviyo for their email. Over 60% of the contacts they were paying for hard bounced or like never opened an email in the last year or something like that.

[00:15:34] And so we simply, you know, filter those out and they save $12,000 a year from us just doing that one thing. That's insane, man. That is not chump change either, is it? Wow. And I mean, AI agents, they are becoming more capable of handling repetitive work across everything from marketing and customer engagement and operations.

[00:15:56] But what would you say are the best opportunities today for anyone listening in a small and mid-sized business that are interested in using AI agents without introducing unnecessary complexity or risk? And also, to add another part to that question, where do agents still fall short of the human element there? And when should they step in with the right guardrails and permissions? In fact, there's probably about four or five questions in there, but I'll hope you can unpack that for me.

[00:16:27] Yeah. Well, to address the first part of your question, there's not really a great turnkey plug-and-play option for small, medium businesses right now to adopt AI. Because quite frankly, you need to be a certain level of technically proficient to really run these things. Now, again, you can kind of do the bolt-on chat. That, you know, there's not really any technical knowledge there, but that's not really moving the needle. If you want to utilize something like Clawed Cowork or Clawed Code or something like Hermes or OpenClaw, you're unfortunately going to need some technical knowledge to get it to work.

[00:16:56] And you're going to need to babysit it more than you think. I think where I'd position myself in the spectrum of AI hype bros, you could say, to complete AI skeptics, is I acknowledge the power and the potential of AI, as I'm sure you do. But I'm also going to acknowledge that you have to be technical right now. Just this is how it is. I would say a company like Marbleism might be getting closer. Marbleism is an organization that, you know, you kind of get seven AI employees, quote-unquote, and they have different specialties.

[00:17:25] But they still hallucinate. They're still confidently wrong. They still do things like accidentally delete all your social posts, drafts. And I know that because I spoke with users at one of their events about this. And so that's why I think right now one of the reasons for our name being Lil Agents when we were founded a little over a year ago is we predicted small agents that have very focused, you know, tightly defined roles will be what are successful.

[00:17:52] As soon as you add kind of more responsibilities and more ambiguity, you're going to have to babysit it a lot more to get it to do what you want. And you need to define context. There are all these steps required. But if you could do something really small and specific, that's where it's going to be a more plug and play, set and forget, that kind of approach. So that's kind of the first part of your question. I would say the answer is no right now. And I'm hoping that we can build the thing that will change that game. The second part, though, is where do they fall? Well, I gave you some examples. They hallucinate.

[00:18:21] They make things up. Anyone that's used AI for any reasonable amount of time have these experiences. And so that's one problem. The other problem is as we give AI more control, there's not a lot of knowledge or a lot of, what would you say, I guess, standards or protocols around guardrails. So they can read your data. That can even be a problem depending on governing bodies that you're affiliated with depending on your industry. So that, you know, reading on its face isn't even safe. Reading data alone isn't safe.

[00:18:50] But assuming that you can read data and be compliant and be fine, there is the deleting. There's like the destructive actions. They could do deleting things. They could do things like modifying and sending things on your behalf. Those are all super dangerous if you don't have really careful guardrails on them and backups. We mentioned backups earlier. That's one example. But if you send an email, you can't unsend an email. So if your AI agent is going to be sending emails on your behalf, there is no backup for that. Sorry to say.

[00:19:18] And I think so much of what you've spoke about today will resonate with people around the world because they now understand the importance of digital ownership and probably locked into a lot of relationships or solutions that they hardly use anymore. But as you said a few moments ago, to get out of that, you do need to be technical. And I know, little agents, you build websites, brand marketing systems, custom tools that belong to users, not a platform or agency.

[00:19:46] But tell me how you help businesses and the kind of questions that you get asked and how you make it possible for anyone to finally break free and own and manage everything themselves. Yeah. The first step is definitely taking account of what accounts they have. Do they have a website? Are they running ad campaigns? Are they running ad campaigns? What's like their current footprint? And then we get an idea of, OK, well, you have all these things. How many of them actually live in accounts that are yours under your name?

[00:20:15] And how many of them are under Joe Blow, who helped you one time five years ago and you haven't talked to him? And, you know, a scarily high number of folks fit in that camp where they're sort of like, OK, we're going to pay someone to do this. And then they're going to, you know, literally run it all. Well, the ideal case, of course, is to either create accounts on their behalf and then give them the keys, but then make yourself an admin or the other way around where maybe you're an admin of it. And then you're making sure that they have admin access to whatever you have.

[00:20:42] You don't ever want to be in a situation where you're a standing duck. But unfortunately, a lot of the clients we talk to are in those situations. So it's always the first step of having the awkward conversation with their old vendor or maybe doing it on their behalf or maybe they have the conversation. Hey, can I have a login for this thing that's technically mine? And then once all that's sorted, we look at what they're spending, what their needs are. And nine times out of 10, we go, your website's overengineered. This CRM plan you're on is way too expensive. And we kind of have an operations layer.

[00:21:11] One of the things we don't really advertise very much is our operations acumen, but we have a really strong background in operations. So we get them, we kind of map their old stack to their new stack, and then we walk them through doing that. And the end result is, there's some case stays on our site if you're curious, but usually people are literally paying $0 a month for their entire, they're probably paying for the domain name, $15 a year or something like that.

[00:21:35] But these are small numbers in theory, but for a small, medium-sized business, if you pay up front to actually own it, and then there's no subscription cost, that's a massive weight off their shoulders in a world where everyone's inundated with subscription services and everyone pays more than they realize. That's a real help for the folks that we've worked with. Wow, it sounds like a genuine call to action for anyone that's looking through their spreadsheet and bank statements.

[00:22:02] A long list of subscriptions there, not to mention the annual ones that come out when you least expect it. But if I was to give you a virtual crystal ball and ask you to look into the future, what do you think will separate organizations that successfully use AI to maybe streamline or create a more competitive advantage from those who simply just add more tools, more subscriptions, and ultimately more noise? Do you see a gap developing there? Oh, absolutely. The way you put it is actually right, I think.

[00:22:32] I think previously it was, okay, we have a problem, what subscription service is going to solve it? So instead of thinking that way, think in a broader system, like a company-wide system. So how can we use AI to bridge these problems that we have? What's our kind of current environment? And can we be conservative about the adjustments we make and use AI to kind of fill in the gaps to make that happen? And then one more thing I want to add to this is the vendor lock-in thing isn't exclusive to non-AI solutions.

[00:23:00] It's very prominent still in the AI world. If you optimize your setup for Claude, use ClaudeMD, everything's stored in Claude, well, what happens when Claude increases their prices and they force you to use their API instead of their pro plan? You're in the same situation you were in before in pre-AI land with vendor lock-in. So you need to make sure whatever system you create, whatever harness you end up using, you future-proof by making it portable and you make it vendor agnostic.

[00:23:29] That's so important even in the AI world. So I think the ones in the next 10 years that will really be successful is they will be so portable. They will be so modular. They'll be able to adapt to any change coming up and pivot to any platform they want on a dime. And they'll have complete control over all those things. They won't be at anyone's mercy. They won't be, you know, constantly having their stuff held hostage. They'll have backups of all their data. They'll have data just seamlessly going from one tool to the other.

[00:23:58] And they'll be able to get so much done with all that compounding knowledge and that intelligence. Such a great point. And I think it's a tech model we've seen many times before. If we go back to the introductory prices of the cloud to tempt organizations to leave their data centers, put everything in the cloud and then ramp up the prices little by little. And AI is obviously going to take that same playbook. I think that seems inevitable as well.

[00:24:23] And before I let you go, though, I have a Spotify playlist for all my podcasts on the Tech Talks network. And sometimes that's my guess if they've got a song that means something to them. Maybe it's when they're coding. Maybe it inspires them. Maybe it's their walk-on song at a tech event. All I'm going to ask is what song would you like to add and why? If I had to pick one song, this is such a cool question. I think it'd have to be Ghost Love Score by Nightwish.

[00:24:51] And it's a selfish add a bit because it's a 10-minute song. So I'm taking up more of your playlist than is probably fair. But it is worth it. It is a masterpiece. If you like bombastic, over-the-top, and you like movie scores, and you like metal, you will love it. Oh, man. What a great choice. I will gladly add that, all 10 minutes of it.

[00:25:12] And finally, for anyone listening wanting to find out more information about your work, how you're helping organizations, especially cut back on all those subscriptions, where can they find you? There are two .coms and two handles to think about. If you're curious about what Lil Agents is doing, that's L-I-L, Lil Agents. You can go to lilagents.com or that handle on almost any social platform and see what we're up to on your favorite one.

[00:25:35] And then if you care about what I'm doing, which is a little bit different, davidvkimble.com is my name, and at davidvkimble are my handles. And, Neil, I just super appreciate you having me on. This was a fantastic conversation. I'm really, really pleased to be here. Well, thank you so much for agreeing to sit down with me today. We did cover a lot there from how to use AI agents to bypass platform limitations on everything from Spotify to HubSpot and so many others.

[00:26:02] And also showing some very real stories of companies that have migrated off some of those locked platforms and onto stacks that they actually own. Digital ownership, massive talking point right now. And thank you for bringing it all to life today and even having a chance to leave us with a soundtrack by Nightwish 2. Thanks for joining me. Thank you, Neil. Take it easy. I think today's conversation was a powerful reminder that AI isn't just about adding another tool to the tech stack.

[00:26:31] When used thoughtfully, it can simplify operations, reduce unnecessary ones, and give businesses far greater control of their digital future. So thank you to David for highlighting how so many organizations have overlooked this until it's too late. And the difference between using a platform and actually owning your digital assets is such an important message. But over to you. Are you building a business on technology you control?

[00:27:01] Or have you become more dependent on those platforms that you don't own and don't control? Let me know. TechTalksNetwork.com We'll keep this conversation going. I'd love to hear from you. But I'm afraid we're out of time today. I'll be back again real soon with another guest. But let me know your thoughts. And we'll speak again soon. Bye for now.