How AI Is Set to Transform Small Business Operations and Growth

Hello, and welcome to the latest edition of the Techstrong AI Leadership Insight series. I'm your host, Mike Bazaar. Today we're with Alex Victoria, who's CTO for Zen Business, and we're talking about, well, how AI will be applied to small businesses and how it can maybe help them compete more effectively against the big guys.

Alex, welcome to show. Thanks for having me, Mike. Good to meet you.

So maybe level set us a little bit here first, but what is fundamentally different about using AI in the context of a small business versus what we might have seen in these larger enterprises where there's a lot more resources? What do they need to do to kind of turn AI on and actually, you know, leverage this? Yeah, there's, there's really two, um, two different ways I think about it.

For, for Zen Business, our goal is to really, uh, help, uh, very small businesses get up and running and get started from the beginning. And a lot of our customers are first time business beginners. Um, and that, that, that journey is a complex one where you, you run into many, many different complex questions throughout the beginning of the process, but even after you've got your business up and running, there's a constant kind of barrage of work that the, the, the entrepreneur never thought they would have to do.

You know, a lot of these folks thought they were gonna get to do a side gig with their dream, but instead they spend 70% of their time making sure their accounting books are correct. Uh, and so, you know, Zen Business has always, even before this AI boom, been focused on that problem. Uh, and so what we're really trying to do is leverage AI to get, to make all that easier for the small businesses to get started faster.

The second way I think about it is how are small businesses going to leverage AI to deliver their products to their customers? And that's not so much what we're focused on at the moment. Um, although I really do think that's where we head in the long run is that we start using something like Velo.

Our, our personal AI assistant for small businesses. We like to think of Velo as kind of like their co-founder, productivity partner. Um, I could see in the future where they're using velo to like, operate things inside of their business as well, um, or, or other AI platforms.

So it's really, it's really, um, either of those, our focus today is really on, on them operating the business. All right. Well you mentioned Velo.

So yeah, for the uninitiated, what is Velo exactly? So Velo is, you know, our vision with Velo is that it's the productivity partner, co-founder for the small business. And then it begins your, with your journey, uh, from the day you start interacting with Zen business.

So the majority of our customers are people who are either already working or already have a business that are starting a new one. And so they'll find us by way of Google or something like that. From that moment we have Velos sitting there with them, uh, helping them research.

So the other thing is a lot of customers will spend sometimes months researching what to do with their LLC, what state to put it in, whether they need a DBA or not, what accounting software they should use, what the website should be. There's a lot of these questions that they have to answer. And a lot of times they're using sources from all over the internet, jumping around and doing all this research maniacally.

'cause it's scary. It's a scary decision to get a business started, especially when you've never done it before. And so the idea from when you get the Zen Business Velo is there from the beginning, you can, you can use Velo for free and you can research all the different options of your business, uh, needs before you get started right on Zen business.

And then it follows them down the journey so they can create an account with Velo and use it for free just to do research. We'll, even one of the things Velo does is it'll even tell you who our competitors are and why you might wanna pick one or the other. The idea is to really just, uh, help the consumer make the right call.

And of course we want them to choose us 'cause we think we're the right call. Um, but we wanna be honest with them too. Uh, the next step from there is if they actually decide to start with Zen Business, then they end up with a full account in Velo.

But at each, each one of those steps throughout the way, Velo is learning more and more about them, their desires and what they wanna accomplish, so that by the time they're up and running, it's got all the context already. And that's, that's allowing us to do things like, say you need to file an EIN velo can fill the form for you because it knows so much about your business already. Uh, and, and that's just one example of what's coming in the next couple weeks.

But there will be a long line of these, like all this kind of operational filing stuff, your annual reports, all the things you have to do around kind of keeping your business compliant with the state. Uh, Vela is gonna be able to do most of that for you. Um, and today, today the customer has to go click like through a 30 page form filler and enter all these complex details to get these things done.

But by the time they've getting gotten to the point through us, we have collected all that context and can make things, these things happen faster for them. Um, so, you know, velo, like when a tax season's coming around, Velo reaches out and says, Hey, it's tax season time. I know that you own an Italian restaurant in Austin, Texas, and you've, your annual reports are due on this date.

And you have, because I have access to your accounting software, I can tell you what your next steps are on getting your taxes prepped and all that kind of stuff. So will it generate the annual report that I might not need to file for the small business based on what it knows about me? Or how far will it go?

Well, today, it's, it's not gonna get, you're, you're gonna have to tell it to finish. You know, it's not just gonna go do it, I don't think just yet. Um, but in a lot of cases, you know, the back end of our filing systems are automated as well.

So we, we do foresee a point where velo could just know its annual report season, prepare your annual report, send you a note and say come hit okay on it. Uh, and, and the thing is, annual reports are quite different in every state. Some, some do them in different, they have different purposes and at different times of the year.

Um, but it could certainly have it ready for you. And when you say, okay, then automation on our backend is also, uh, doing it quickly for you. And so that's the vision is that, um, in this space, we've, we've been challenged with combining technology and, and people and AI has become this, this thing that is allowing us to do it more effectively.

Uh, and it's, it's, it's just turning up the automation opportunities to the roof for us. And then what we're learning is how to inf help our customers. You know, the next step is like how to figure out how to get them to do it too.

I truly, personally believe that like the small business is probably the primary way for the average am average American or average person to get their piece of the AI boom. You know, I don't want to get like philosophical, but you know, a lot of wealth is accumulating in, in, in a small number of places. Um, I actually think this is a better time than any for people to start a small business because they're gonna get to use AI to make it more effective for their customers.

So it's exciting time to, to be where we are. We've been slogging at this for seven years and it's, it's all of a sudden, uh, I think we're at the point where we can actually achieve our dream, which is the, is to be the most effective as possible to help our customers make this decision to get started. 'cause it's a scary one.

And we believe in entrepreneurship and everybody having a chance to be their own boss. What's behind the platform exactly is because I don't think, did you build your own foundation model? Seems unlikely.

So what is No you guys working with here and how does it Yeah, what come together? What I tell it's funny, people ask that, and I tell 'em investors, I'm like, well, if you gimme a couple billion dollars, you know, I'll go, I'll go make a model if you really want us to differentiate. Um, but no, we don't need to do that.

In fact, velo velo itself is a, a fairly common public pattern that we've written about on our tech blog as well. Um, it's a, it's an, um, a it coalition of routers agent a, a coalition of agents within a routing agent that, that decides how to answer questions or do tasks for a customer. Each of those is dependent on, uh, prompts and a bunch of tools.

So what makes Velo different for us is that it has our knowledge base, which is really every detail of everything a small business would need to know in all 50 states. Plus, uh, it has access to our tools, which are all the different platform APIs we apply. So we have your website, bank account, domain names, uh, uh, accounting software, and then all the compliance and formations needs as well.

Uh, and so all those put together, uh, allows us to do this. Um, the model part is interesting. So one of the things we did early on was we decided to centralize LLM usage in our platform.

And so we have a central part of the Zen business platform that obfuscates the details of the different providers. So we use Google, we use OpenAI, we use philanthropic depending on the task, and, um, we can change those on the fly if needed. Um, and we've had to do that, like all of a sudden Gemini has a problem, we switch it all over to OpenAI as a back a backup thing that we wrote about that on our, our tech blog as well.

But another crafty thing we've done, 'cause people get concerned about the cost, is we use the, we use very old small models. So like we're using the Gemini two oh flash for like, most of what we do. And then what we do is when there's a, when there's a certain case that just needs the additional logic, we can upgrade it to like, say, go use OpenAI four oh or whatever.

But we, we, we avoid the, the biggest, most effective, most expensive models, kind of like the plague. And what we're finding is, you know, for most of what we need to do, we don't, we don't need 'em. So we like the ability of not being locked into one vendor.

We can switch back and forth and, and honestly, they all, they all are producing really amazing technology that it works. You know, it's, we, none of this would work. The real, the real, the real thing we learned was the tool calling the model has to be good at calling tools and, and we solve a lot of problems with tool calling.

Do you think as we go along that, um, small businesses typically interact with other small businesses, and if we're all gonna have our own AI tools and AI agents at some point, you know, does my AI call your AI to complete a task as a small business? Or how does that play out in your mind? I mean, I think we'll inch towards that.

Uh, it's funny, I think, uh, years even even when this all started, like a year or two, a couple years ago, we were like, pretty soon recruiting is just gonna be like, you know, the, the hiring person's AI talks to the candidate's ai and they just, they make a decision. I don't think we're there yet. Um, I certainly think AI can facilitate the communication between the two.

That's like a no-brainer use case today. And that's actually where most of the real value I've seen is coming from these AI products is like customer service. You know, that, that it's things that are heavily, uh, layered of upon people talking to each other can be facilitated by ai.

So I could see AI going, you're about to do a deal with so and so, and we're gonna redline the contract and back and forth it between two ais and present it to both of you at the end and say, this is, this matches both of your needs. Like hit hit okay to go. Mm-hmm.

Um, we, I, you know, that's, that's, that's, that's here we could be able to do that today if, if someone wanted to. Do you think small businesses are prepared for this scenario though? 'cause at some point their customers will also have AI agents and those AI agents will be optimized to maybe buy something at the lowest cost possible and your AI agent is optimized to sell it as the most profitable way possible, and the two of them will battle it out somewhere in the ether.

Yeah, I mean, I guess supply and demand, that's like the, the purest form of supply and demand curve is mm-hmm. Two algorithms go until they find the happy point and go. Um, yeah, I, I think certainly it could be facilitated by that.

Um, it, it's a, you know, we've found that like a lot of our customers, the ones who aren't in technology businesses, uh, managing all this stuff is quite complex. I mean, e even even for like, you know, entering your expenses and having the time to even go log into something every day and make sure you uploaded your receipts for your, uh, for your expenses, that, that, that's challenging for a lot of the folks. And so I do think there's gonna be some work to figure out how to make them able to use this technology because it's, it's pretty complex today.

I mean, we we're lucky to have a, a bunch of really smart, uh, product development people at Zen Business, um, and, and you know, Google and Anthropic and those guys have everybody else. So I, I don't know how the, you know, right now the average small business that doesn't know how to wield this stuff is, it's gonna be hard to figure out. So that's where I think there's an opportunity for us in the future and somebody else to help.

I also think, you know, all those engineers that people say are having trouble getting jobs, I think they, they have a lot of value if they know how to wield this technology. Um, I was telling my nephew who just gradua is graduating from college this year, as with a CS degree, is like, you know, if you learn how to use this stuff, you're gonna be really valuable. You know, so, because I think it's pretty hard to be getting a job right now and Well to that point in the societal impact, right?

So we're seeing it's harder for, uh, new people to get hired into a space than ever. It doesn't matter what the vertical is these days. And we're also seeing more folks get laid off because companies are thinking they're gonna automate certain things that they used to hire people to do.

But does that create a larger pool from which we will see more small businesses start to emerge and people will be able to engage in that activity because the complexity of starting a business is dropping? Yeah. And that I, I agree with that completely.

Um, I think, well, one thing I don't, I don't necessarily think they're all getting laid off because of ai. I think maybe they overhired and there's just a reckoning happening. Um, but there, I do think that, uh, that's gonna be a way for people to benefit from all this and it's gonna be easier than ever.

And, and honestly, when we started Zen Business, one of, one of the ideas was that, uh, the, the barrier of entry for starting a business is lowering because of technology. And particularly at that time it was kind of SaaS technology. You don't need the firm anymore to get a business started.

You don't need to hire 10 people. You could get online and start taking money and put a product together and sell it. And you could do that all yourself with technology, AI is like the next evolution of that where, I mean, you could build your, your product with cloud code if you, if you knew how to wield that thing a little bit.

And that's why I think, I think, you know, I'm very passionate about those ag agentic development tools. They work and it, uh, people should be committed to using them. We use them a lot.

Um, they still require a technology person to run 'em though, you know? Mm-hmm. And I, and I think that's gonna be the case.

And uh, uh, one of the ways engineers can get a lot of value in their careers too, is learning how you still need to know how to build software ultimately to build something complex, you know? So do you think small business owners will develop something that feels like a relationship with these AI agents and maybe even getting the name? Or are they gonna be more, you know, seen as disposable kind of software widgets and we're not gonna get that emotionally attached to them?

That, that's an interesting idea. I'm trying to think of, the way I use, the way I use Chad GBT is I have, I have like dialed that thing in to, to be like, really, um, personalized to the way I like to communicate when I'm using the voice mode. And I, I, you know, I, I don't want any fluffy language.

I've come up with all these prompts that make it say as little as humanly possible and it just treat, I want, I'm like, I want you to treat me like a robot. Just like say things you know, don't. And uh, and then I even picked the voice and one day I was using it and my wife was like, did you ever notice the voice you chose?

Sounds like me. It's like, I don't know. Mm-hmm.

I, maybe I did. I don't, I didn't do that on purpose, but Right. Yeah.

I think people are gonna wanna personalize their experience if, especially if they're com communicating with voice, which is right around the corner for us. Uh, we're, we're gonna enable that sometime soon. And I think a lot more and more people are talking to these things.

I think you're gonna want it to kind of know who you are and know how you like to communicate and probably give it a name. I mean, I would, I'd, I'd, I'd love to high if it was running my business for me, I'd be high fiving it. Like if I could high five cloud code, I would, yeah, CLO code's amazing.

Well, maybe, you know, you're gonna do what the agent says you should do because it sounds like somebody you're already listening to, right? Well, yeah, there's, there's a lot of like futurist thinkers saying agentic CEOs are around the corner. So I could see a world where you've created your own agent to just kind of manage you day to day and you're just doing what it tells you.

But of course you created it. There you go. If you don't like it, you can veto it.

Yeah. Um, so as you look into 2026, you know, it's a new year, what are you excited about? What do you think we're gonna see next?

Well, I mean, uh, for Zen business, you know, like I said, we've always been in this spot of just trying to make it easy to get your business up and running and then run it, focus on what you love to do and make some money and be your own boss. Uh, I think for the first time velo, you know, having the technology in place to deliver velo throughout that experience is the most exciting thing I've, uh, I've, I have going for me. Um, you know, and then also when we started Velo, we really made the choice to, to kind of learn in, in public and write about it on our tech blogs.

So, uh, I'm, I'm excited about, you know, showing off the talent we have in the team, but also giving back to the folks that, you know, put some of these things in place to get us up and running. And maybe, you know, some of our customers learn from it too, uh, and I think, uh, we're gonna, we're gonna really accelerate what we can do. The other, the other learning that we've had is, like our velo team has only been around for a couple months and when you know how to leverage this technology in the right way, they've, they've delivered at a rate that is way higher than traditional engineering teams.

And so I think we're gonna see an extremely fast pace like feature roadmap, you know, Eve's gonna do more and more and more and more every day as we continue to focus on it. And, and you know, that's probably true of anybody who's building an age agent system like this. So in terms of like the macro view, we're gonna get a lot of software.

You know, we built, we have this internal system, we built a very complex management UI for it with, with cloud code and uh, you know, with automating engineering pipelines. And of course I had to like, take care of it, but the thing it built was so complicated, it actually made me think like, do we really need to build this? And I think that's the world we're getting into.

Like you can literally build anything you can think of and if you can, and you can do it with a low number of resources, except you gotta pay anthropic, OpenAI or Google, uh, and, and your maybe open source models get to a point where you can just use 'em to do it too. But if you can build anything you want, I think that's pretty exciting world. You know, like, yeah.

So I, I think it's gonna produce some cool companies and cool businesses small and large. Alright, well folks you're hearing in here, hey, if you're outta work or you're just playing fed up with your existing job, go start your own company. 'cause it's gonna be a lot easier to do than ever.

Hey Alex, thanks for being on the show. Yeah, thanks for having me. All right.

And thank you all for watching the latest episode of the Techstrong AI Leadership Insight series. You can find this episode and others on our website and invite you to check them all out. And until then, we'll see you next time.