Bonus: $65M Exit, Zero Employees: How Olauto Automates Everything Except Customer Service

We could use AI to make the business a little easier for us in terms of, like, the impact on our day to day.

But never sacrifice the humanity of the business by doing so.

So that's kind of where human touch comes in. People shop with small mid-sized businesses and performers because they want to feel that connection. So if you're optimizing for deflection rates and you're optimizing for automation, you're missing out on some moments that could really make an impact to your total customer life-time value and the relationship that those customers have with their brand.

Hello and welcome to the DPSI podcast. Tyler Handley sold inkboxed a big for $65 million and his new company sells a $33 air freshener. Launched last September, it's already profitable has zero employees, no CX team, no marketing department, no ops hires, two brothers, some contractors and AI running everything else behind the scenes. Except for one thing, when a customer emails a lot of a human answers every time on purpose.

And the guy who built the software who makes that possible is Mike Malacic, who's Tyler's known since high school in St. Katherine's and that lives in Cambridge, which is my hometown that I'm headed to next week, which is wild. It's really exciting to have you guys both on the podcast. I'm really excited to explore how far brands can push AI before customers stop trusting you. And you guys are actually kind of running this experiment in real time.

Welcome to the DPSI podcast. Thanks for having us. Thanks for having me. Tyler, why don't we start with you? Give me a bit of your hero's journey, maybe a little bit like about the ending of inkbox.

And kind of how you evolved into a lot of, and what you're really trying to build here.

“And why human touch has been so important for it.”

Yeah, so with inkbox, we sold temporary tattoos. This was pre-AI obviously. And we were really focused on the artists in the artistry. A lot of other temporary tattoo brands at the time. That kind of indirectly competed with us because we had novel technology that made the tattoos last

a lot longer, look a lot more realistic. They would just use, like, really like clip like created art, like not designed by like real tattoo artists.

And so what we always rested upon as a brand was the real artistry of the tattoos.

And, you know, in the age of AI now, like, if you're restarting a brand, it would be very easy to look at it and be like, Oh, you know, we're going to, we're just going to outsource this to AI and no, that's not possible. Anyways, a long way of saying that with inkbox, it was a very human-focused company. We had a lot of employees because we did a lot of R&D, we manufactured the product ourselves. It was my first time running a business and I was kind of learning along the way, right?

“So when you don't know what you're doing, you kind of just put in more hours because that's how you make up for it, right?”

You can't be quite as strategic with your time because you don't know what that means. You don't know how to optimize your time because you don't know what you don't know. And so when it came to starting a new business, being a lot of, we obviously seeing AI develop and what it could do as well. We wanted to find this perfect balance where we could operate a business and all the stuff that we, you know, we could oversee, but not necessarily do day-to-day, we could outsource to agents.

And then everything that was very like human about the business, obviously the brand, the creative, the customer service, the emails, like, you know, nothing's written with AI. You know, nothing that the consumer sees, you know, it might have passed through a little bit of AI at some point for inspiration or using it as like a sounding board for us, but at the end of the day, like it's all like very human. And so we wanted to find this balance where like the business didn't take over our lives.

So we could use AI to make the world a little easier to make the business a little easier for us, you know, in terms of like the impact on our day-to-day.

But never sacrifice the humanity of the business by doing so.

And so that's kind of where human touch comes in. It's super interesting because I, you know, I'm reporting on AI in this business since we, we first started since we, since AI, couldn't make hands on images.

“And I think CX is been one of the spaces that people have had eyes on that will be impacted most by AI.”

And Mike, you kind of went, you went a different direction with that. Talk to me about the origin of human touch and why you built it. Yeah, so just context for listeners. And I can add a little bit of color to Tyler's story, too, just because I was at eight bucks in the early days too. So all the things he said earlier, I'm like nostalgic about. And I can say with confidence that like eight bucks would have been a different business in the early days if there was AI.

But we probably would have still met a lot of the same challenges, right? And I don't think we would have outsourced a lot of the creative that we did. We wouldn't really want it to maintain the human element. So yeah, as it relates to the human touch, you know, I was showing a prototype to Tyler in the early days and he loved it so much.

But it was sort of the synthesis of a couple things coming together.

I had been doing some consulting previously with a business called Fusant Labs and there were all of the transparency and the gradient formulation. So I had transparency on my brain. And then, too, just as a consumer myself, you know, Tyler and I have been in the Congress for a decade plus. I like to support small e-commerce brands when I can. I had some questions and I was on a website and it was talking to somebody on a chatbot platform.

“Um, as getting some weird responses and so I asked them, I said, you know, are you a bot?”

And they're like, no, and I'm human. It's like, right, what did you have for lunch today? They can answer me. I think the response was something like, I can only answer questions about the product. And I was like, that's, that's a little bit strange, like, where are you living?

What was the weather today that I could only answer questions about the product? So like, with transparency on the brain and then having that consumer experience and then sort of the business background that I had.

I did some investigating, you know, I've inspect the element.

I've got some technical shots, so I was looking at it. And then I went into the app marketplace and Shopify. You know, I searched like customer support. You just control F-A-I. Everything's A-I now.

So it just felt weird to have this consumer experience and the experience that I had running e-commerce brands. And then, you know, see everything that's on the market there.

“And it felt like there was a transparency layer that was missing, right?”

So that became the sort of ethos of what I wanted to build next. And like I said, you know, I got Tyler involved in early stages. Showed him the prototype and sort of reminisced about some of our paths forward together. And talked about the mission and the vision. And then very quickly, you him and his brother became sort of power users for us, which is great giving us sort of that early stage feedback.

But yeah, I found it really interesting. And you know, I get it as somebody who's operated e-commerce brands before. Selling out e-books, a solid majority, maybe a bit on the tech side too. That's mild, I know. And we used to say, that smile, like our biggest competitor is a merchant's time, right? Your e-commerce owner.

And you have so many battling priorities, your prioritizing all the time. Automation is useful for you, right? But then consumers on the other side are getting bad experiences. And there's consumers that are anti-AI now. So it feels like there's loud groups on both sides, and both are kind of wrong.

“And I think the merchants that are going to be successful tomorrow or the ones that sort of strike that fine balance.”

And it's something that I see, you know, to Tyler and his brothers credit. They're really, really well and then eventually they've started. Yeah, there's that there's a way to use AI without it becoming slop. Tech companies will promise this crazy future where no one works.

And I mean, they always have these utopias that when you actually think about it,

if you're like dystopias. So, you know, they're trying to like push this narrative. And specifically, young consumers, like are having absolutely none of it right now. Because they're like, we never got to participate. You know, our jobs are going to go away like, you know, we never got to participate.

Like what happens for us, you know? It's a good point. It feels like there's a consent transaction that's missing, right? There's a consent layer. Yeah, we don't want to. You know, it's like, we don't want to. Yeah, they're automatically all students.

Right. And so, you know, I think there's a way to use it. Like ethically is a big word. Because we all might have a slightly different like view on the ethics of it. But like, I think there's a way to use it. Ethically across culture, across music, across content creation, across business development.

And obviously, in a business, like there's things that, you know, having run a quite a large organization before that, like, yeah, I could just offload this onto an AI. And it's going to be no worse for where for anyone in the world. It's just going to make my life a bit easier, right?

It also just fits the D to C ethos. So much like people are buying from these, you know, owner operated brands to avoid maybe supporting BlackRock or one of these huge, huge corporations, right? And so if you're going to be using the same Slop techniques or Slop support, you can be cast with the same brush.

When you really are, you know, you're two brothers building this business. You want to have that connection with your customers for us, by the way. We're two of us by the way, we're two of our brothers. And then we have two other friends, and we're between Austin and Toronto. It's kind of like the makeup of the business.

But it's just for us doing, like, the whole business, essentially. So can you just pay me a picture of, like, what, you know, this sort of, like, hybrid human touch that is still has automation, because you still, it's not just going to be someone on the keys all the time. Talk to me about how it actually works to provide CX in this sort of, like, human touch way.

You know, a good example is, like, automated, I mean, the most straightforward simple example is automated responses. Like, there's, I think, a 65% of our customer service requests, that a lot of, and salines, I think, what it was at Ink Box, too, is probably shipping.

It's something that's usually beyond your control.

“But customers are just looking for some insight, right?”

And so you have your, like, automated fields, right? Like, forms and stuff like that. But, you know, oftentimes, there might be irregularities with a shipping, like, they, obviously, they could have a wrong shipping address, or it got delivered, but then stolen, like, that's a very, that's a good one, because that's a very human one, right?

Like, if you just had automated responses, like, that would just be, oh, like, oh, I'm sorry, like, it looks like your order was delivered. It's like, no, it was delivered and then maybe stolen, right? And so the ability to actually go in with human touch and, like, look at every form field before it goes out and, like, before it gets sent for, for an actual response to a customer,

I think is really important. That's just one edge case.

But I think what it allows for is this concept of, like, never getting caught on edge case.

You know, you can use AI to automate this kind of, like, your, your back end responses to, like, your most frequent requests, right? But you never get caught sending one out without having, like, being able to, a, a review it and, like, understand that it's something that could be more particular. Yeah.

You guys were the biggest source of feature requests for me, which was great. It was great. I mean, experience getting online live with you guys and seeing kind of your actual music. And that was one of the early things that came up. It's like, hey, like, 65% of our inquiries can be solved with automations.

To be clear, I'm not against automations. I'm not really even against AI. It's, it's a nuanced thing.

Again, it's, it's a bit of a striking balance.

It needs to be supervised at AI, right? You need to have it that, that transparency layer. You need to, like, customers know when they're interacting with AI. And it's also automation. Yeah.

Yeah. There's an automation that simply references an audio based article and it can be delivered very quickly, almost instantaneously via a, a chat widget.

“I think customers are generally okay with that, right?”

It's all of these other things that sort of push beyond that. And one thing that I think merchants might be making a mistake with now is sort of optimizing towards the wrong things. They optimize it. I was listening to merchant talk the other day about sort of optimizing towards deflection rate. I don't even know about their high deflection rates.

It's like, well, what's, what's kind of missing in that number, right? Like, yes, what's a deflection rate for the audience? Things that miss the inbox entirely, right? So it's, it's solved by the AI. It's solved by an automation.

It's solved by just, you know, somebody pressing it by and getting the answer that they need. And again, I think there's a little bit of nuance in that and saying, like, well, was there a missed opportunity for a human connection where you might have an opportunity? Let's be honest, these are some of your best loyalty moments when people feel connected to your brand. And again, so we're going back to my experiences as someone who's been in the congress for a while.

And I don't think we both might experience as a consumer too. Like, I bought different consumer brands because I want to support the mission. I want to feel a part of the brand community. I want to feel a part of whatever that human is doing. We live in a day and age where we can just press it button.

We get a drone to drop something off at our door from these big suppliers that are out there. And then if there's something wrong with it, talk to a bot, who will send you shipping label and they send it back, you can, you can do that all that at a touch and button, but people shot with small mid-size businesses and the commerce because they want to feel that connection. So if you're optimizing for deflection rates, and you're optimizing for automation,

you're missing out on some moments that can really make an impact to your total customer lifetime value and the relationship that those customers have with your brand. And again, to you and Raiden's credit, like, Raiden, even on the market, I took some convincing for me to have him do. Some of the automations and off-hours for the widget, and he even still has eyes on it, to have working all the transaction records and looking at everything he has everything from high to by his eyes or on it.

And I think that's the wrong approach. Deflection is rarely a good thing unless you're like in combat. And then you're trying to, or in like a, you know, verbal fight. You're trying to deflect things. Otherwise, you want to connect.

I'm curious, like, what are some moments that maybe a lot of merchants out there are deflecting

“that they should redirect towards human connection?”

One that comes to mind often for me is, is anything that's product related? Some people think they can just post a knowledge base about the product, the block and reference it, and AI can say this is exactly what you need. If they're searching for answers in product, they're pretty low funnel for you, right? They probably have a cart, they're probably ready to check out and just having a little bit of that white,

love support getting them across the edge. That's going to lead to higher customer lifetime value customer. It, in my opinion, and so it was interesting. I actually ran into this problem when I was building in, when I was building human touch, and I was building some of the automation features to reference knowledge base.

And some of the product questions I was testing it with, because I wrote a knowledge base on just the Shopify default

They load up a whole bunch of snowboards, right?

Because Shopify, people know Shopify lore on this, but they originally started as a snowboard shop. And so they load up a bunch of snowboards, and it's like, what happens when I ask the AI, create a bunch of knowledge base articles, ask the AI about the snowboards, and it was giving me the worst answers.

“It was like, you should not take that response, and then go out and snowboard that.”

It was like giving me snowboard tips too, and stuff was like, I don't know. That's terrible, right? And so, if somebody's asking you a question about the product, they're definitely in work. They're definitely low funnel.

You have an opportunity to really cure and experience for them.

And go the extra mile too, right? So not just answer something directly about the product, ask something deeper. How are you going to use it? How does it fit your lifestyle, like get into it with them? And make sure that they have a good experience.

And if you even want to go even further, follow up with them. Hey, a week later, like, how's it, how's it doing for you? You're not going to get AI to do that. I mean, you could, but it, it's going to be challenging.

“But also say, like, does the AI, what counts as a deflection?”

What defines the customer need being solved? The customer problem being solved, right? Like, if it's a missing product and the deflection is just, hey, it looks like it was delivered. Like, what did that solve? I keep going back to this, because it's like, it is one of the most common things you get in any commerce

is just shipping issues. And there's many different ways you can. You could probably deflect them. But like, it's really nice having a human voice there, too, for those situations. I'm so, obviously, like, most people pretty sick of, I don't give a shit when I get an email that just says, like,

I'm sorry to hear that Tyler. Like, you know, it's like, you can really tell when there's no care taken. And like, I get it when it's a Amazon or something, because Amazon service is insane. Like, I feel like, for them, like, whatever. If I'm talking to an AI, like, you give me refunds when things are clearly stolen.

Like, it's, they really favor the consumer, right? But for us, like, small, smaller brands, like, that's not the case, right? And so, I would expect from us, like, more of a, just a human element in conversations. One of the things that keeps coming up on the podcast is just how important the voice of customer is. And so, when you're, you know, when you're building your marketing angles, when you're building your landing pages,

your product, you know, all the content about your products, the voice of customer being able to echo that back. And so, I imagine having human touch with your support being able to probe into different situations is going to reveal more valuable aspects of that voice of customer. And then you still have the AI to quantify it all and be able to give you a report on kind of what people are saying, but just having the human able to probe into those situations is going to reveal more valuable insights in that voice of customer.

Yeah, and again, you can put in your region instructions when you've been kind of built into the platform.

“And you can say, here's my grand voice in tone, here's all the key topic in points, but just having human supervision, right?”

In our, to AI transparency layer, it's either, you know, it's AI derived through, you know, an automation is the only way that you can do that. It's fully home human or it's AI assisted, right? And it's fine, again, to have AI instructions that really capture that voice of customer, but make sure that you're getting it right. And add a little bit of, like, so to speak, the human touch in the follow-up communication and, you know, the conversation leading up, make it conversational. If it's purely AI driven and it's purely based off of just an instruction set, it can get predictable for customers that are loyal and returning.

It can be, you know, people are really wise nowadays to certain keywords that tell you that it's that it's AI, it can be telling that it's AI and it can feel off put it off put by that. There's some risks alongside, and again, so that that full automation path. So again, I advocate for some human element interacting in that. And I think it only strengthens your ability to use the voice customer. And I think it's easy behind this as well. We're talking about this as being very beneficial for your brand, but the EU has just sort of passed some laws that make this transparency piece like required.

Are you selling into the EU Tyler and are you, are you, are you kind of aware of what's going on there?

Aware of it. Yeah, I think the EU is always ahead on privacy policies.

You know, any policy consumer facing that you is always far ahead, especially if you've ever worked in like the beauty sector or the food sector. It's surprising that with digital with AI and and privacy and transparency like the EU would be forward facing a lot of occur and they don't. We don't, I mean, we technically ship there if like you buy from there, but we're not marketing there yet. We just started this year. So the US is a huge market and Canada and the US alone is you can sustain quite a large business before you start.

They're just complicates supply chain, right? And if you're trying to build a similar simplified business, you know, that that that eats into that simplification.

And we had begun international expansion at that time begun shipping to the EU and actively marketing there.

And the reality is it it touched a lot of different parts of our business with a lot of different stakeholders.

You know, our email consent. We had to do double locked in. We had to do review cookie policies. We had to review pot privacy policy. We had to have legal involved CX involved my email automation team involved. There are a lot of parties involved to make sure that we stay compliant. But once we were compliant, you know, it was sort of the set standard movement forward.

“And I think some people are going to be playing catch up a little bit with the EU AI Act.”

And article 50 literally just like 10 days ago now, August 2nd just became live law.

So if you are selling into the EU, it is something that needs to be compliant with. And the EU AI Act is really interesting in that it requires merchant selling into the EU to be transparent, right? So there's sort of varying degrees of severity of requirements. So if you're selling anything financial and anything related to somebody's health, you know, your level of risk and your level of compliance is quite high. But for the average merchant, the average outcome is a requirement for transparency and for compliance and for record keeping of when AI has been utilized throughout customer touch points.

“And I think a lot of the big VC backed tech providers that are up there right now are sort of back pedaling a little bit and trying to now position as, you know, AI power, but human supervised, this is all human supervised.”

For us, like, you know, we built it as part of the ethos from the ground up, right? The AI transparency layer is first and foremost touches everything we do. And we're even building and launching your compliance center today, where you can just like one click and go in and download your full audit logs of everything that AI is touched. And if the EU comes knocking on your door, you can just kind of show them the records and show them the transparency profile that you've got set up.

So yeah, it's interesting. GPR, EU, AI Act, I always feel like they're a step ahead in terms of their privacy laws and I commend them for it. I think it's something is great.

Tyler with Inkbox, how big of a CX team did you have? Oh, man. It was quite substantial because we did custom tattoos as well. And that was like about a quarter of our business and like people would submit all types of stuff right custom uploads. So there's a lot of oversight. I actually don't recall, but I imagine it was probably something like 13 to 20 people on the CX team rotating.

“And he'd give him time. Yeah. And then with this business, how many people are on the CX team?”

No, I mean, well, we have, we have like my brother oversees the CX on the team. It's you guys. Yeah. And then we, I mean, well, Mike, do you want to jump into how human touch could help us expand our CX team as we scale? Yeah. I mean, so one thing that we, and we discussed this in the early days because you're like, you know, eventually we are going to have to hire, can't just be me and my brother. And by the way, if anybody wants to talk to Tyler, Brandon, just send an email support tip in over to a lot of, and you'll get that in your connection.

But if you have sort of the, the record of like that progressive record of CX connections over time and even introduce a new agent to your team, they can be assisted by AI. Again, not anti AI, but having that level of supervision and that level of that fine balance of integration can get them up to speed faster, right? And we talked about this. You're like, when I do do make that first time, I'm going to pull the trigger on that, put that bum in the seat, so to speak, I don't want to, I want to limit the time that that meeting my brother have to sit with that person and coach with that person, right?

And so that's a layer that we sort of intentionally built human with, so human is our AI assistant, and so you can ask human questions as you're responding to, to tickets and as you're responding to chats in real time, and it will give you suggestions and it will base its responses on its agents instructions on your product catalogs on its data sources. And on past tickets that, you know, you unbrain and have responded to, so it matches the brand voice matches the tone of limits, the amount of effort that you guys have to exert in terms of getting that agent up to speed it.

So that's one thing that we can really thoughtful and that's a little bit rooted in the experience we had at Inkbox too, because let's be honest, it's a, it's, it's a high turnover position, it can be high stress, I remember the rows and rows of CX people sitting behind me, right?

I think you, even you and I were talking about the other day, Tyler, the, when we first launched the custom tattoos, right?

“And some of the custom tattoos that people make, and some of the, the customer service tickets, the customer service agreed we would have to respond to when those come in.”

So that sort of a learning experience that has to transfer from agent to agent has quite a bit of nuance within it that, that that would require like the sensitive good response there, I'll let you tell the story on that one though, because that was. It's positive.

But what I've never fared well with purely AI automated CX, Bob team.

Did you get some raw T tattoos out there?

“Yeah, I mean, like, listen, there's, there's these nuances of humanity that AI doesn't quite, I think, understand and, like, calls of judgment.”

For example, we would have, when you submit a custom tattoo on the website, there's obviously things we wouldn't have printed racist symbol sexist things, anything insane like that, swear words fine, whatever. You know, we're not prudes. And I would imagine people have nudity on their tattoos. Yeah, nudity, yeah, of course, but, like, obviously if it was like a graphic, maybe it was a minor or two graphic, yeah, like. Yeah, so we had, like, we had actually built some automation to filter out, like common things.

So if there was pattern recognition for common symbols of hate, for example, that we would filter out immediately in cancel the order. But there were a lot of things that, like, really sat on this, in this gray area where it, in specifically regarding, like, for some reason, kink really falls into degradation fast into segregation. And there's these crossovers between things that we find horrific in society and things that people find very sexually attractive. So that, like, that crossover was very fascinating. And, you know, often, we also didn't want to like deprive people of their ability to, you know, express themselves and regarding their sexuality.

So we had to basically have, like, a team internally when these tattoos would come in, that would, we'd discuss, like, do, is this something we, we should produce.

“And oftentimes, we would actually have to reach out to the customer and be like, hey, like, I'm not going to say what one was, but like, you know, you said this in this design, like, why?”

Like, what's the reason, like, why are you getting this? And it almost, like, 90% of the time it was like, my partner and I, my partner and I, this is something we're into. And then I was like, okay, like, then we'll print it. Most of the time, it wasn't something super bad because I think human nature is genuinely quite good. But, you know, there were a few occasions where we caught some pretty gnarly shit that we couldn't print that we had to. Yeah, fine line, it really does need that human eye. And like you say at conversation, sometimes that's interesting. You've got to take it that step further, Tyler, how big you got into the subt really solid eight figures or higher. I don't even know with ink bucks. How big can you grow a lot of with for team members.

I mean, that's a really good question. I think eventually, like, we'll have more contractors on board. I mean, it's not like we don't work with some already, but I think what I'm seeing with a lot of my friends companies that they're starting in specifically in the consumer space. People are, I don't know if there's a term for it yet, but like, friend foundering like where you're building these businesses with your friends or a couple people like that you can kind of sit on top of the whole business. And so like, you know, for example, a lot of I oversee like not totally because there is crossover, but like for the most part I oversee brand finances, the planning and all the product development.

And someone else oversees the supply chain logistics inventory, things like that, and then you have someone else like my brother overseeing all the CS and all the influencer. We have like an influencer who is part of the team who helps with the influencer stuff who works with me on brand who does some photo shoots and stuff as well. And then we have one guy who does all of our advertising run all the ads. So, you know, as you can think of building the business out, you start to expand contractors underneath each of us. And maybe some permanent team members here in there as well as the business scales.

So today's your question, I don't know. I mean, what you're trying to really do is use AI as much as you can to offset the stuff that would be very meanial. So forecasting and supply chain planning is something we built all of our back end data analysis supply chain inventory planning of entire ERP system essentially for ourselves, which negates the needs for like financial planning and a lot of I think we know enough of it as well that like we when you've run a larger organization, you can kind of understand where you really need someone.

Someone and where you can just kind of oversee an agent doing something. Yeah, I don't know how big we can get. I'm not super interested. I think one of the things my team is running on, you know, we're running on cloud we're building agents, but we don't have a team cloud account yet.

But it's interesting that you guys are running this fully with a bunch of AI tools, but you're still kind of running it separately.

“And he thoughts about when you might want to look at centralizing your AI?”

I think you could just tie them together. Like I think we've built a few things with cloud and so, you know, we had to have our own systems we've built and I don't see why they can't just speak to each other. I guess so.

And you say you built your ERP, did you build that, did you vibe code that?

Yeah.

“So our experience with vibe coding is it's really easy.”

It's fairly easy to have an API hook to it has no, it's still like it does to to our sales volume and things like that and to like Shopify and what have you, but doesn't have API hooks into our 3PL because if you think about Shopify is a platform like what it's really good as integrations right and so all of our supply chain stuff actually throughout through Shopify already so all we have to do is hook into Shopify and we pretty much get everything we need. It's not like as complex as a full ERP system. We've built it as streamlined as we need it to be, but that's what's awesome about like being able to build these tools now for your businesses is SaaS companies historically would be building something, especially especially as they scaled and this is a good segue into human touches doing.

When you start scaling a SaaS business that that sells to companies doing D to see you get bigger you start to expand the market you start adding more features is a more feature requests at a certain point. The new customer trying to use your fast platform is overwhelmed because there's so much to learn what's awesome about what we've been able to do is we can just build exactly what we need nothing more.

“So it keeps it streamlined and it keeps it simple and then we can just add to it specifically to our needs and so I think this customization is amazing and that's kind of where he is much more vertical slices tailored the spoke to your needs.”

We identified a market gap where a lot of the traditional CX platforms for ECOM are incumbent now and they've been around for enough years that and a lot of them are venture back that they they've reached this scale where there's this complexity that not everyone needs and you guys are really kind of building that more from the ground up where you're always going to try to keep it as simple and streamlined as possible and then you can route off to the AI for the more complex tasks essentially. And I can speak to the the need you're referencing to and I can I actually and again not anti AI I used AI to help me prototype and I brought the prototype to Tyler brought the prototype to team of engineers that I partnered with for on the team now and they've been working without AI for the last six plus months.

The context window and our code base has has grown and grown and grown and I actually just yesterday gave them a cloud code subscription because they're like there's there's a few tasks in the code base that we think we can just tackle with cloud code and it'll go way quicker and then we can move on to the next year. I said yeah no problem right here's a cloud code subscription going and do that again not anti AI it'll it'll help us accelerate our team and our learnings and there was a little bit of friction in the handoff from when I had sort of.

And not the best coder myself I can handle anything front and UI UX the prototype was was very scrappy in terms of what I brought to Tyler and what I brought to the engineering team and then they just buttoned it up and took it to the next level for us there just absolutely incredible. And I think one of the big limitations I saw in that handover and where we're causing a little friction is just sort of the context window right so over time as I was prototyping you building the context of why I built something a certain way in why I positioned the UI and the UX a certain way.

What sort of lost in the code base and it was in my mind and it might have been somewhere in a reference and agent instructions but the the context isn't necessarily there.

And if you're fully relying on AI so say I didn't have the engineering team and I needed to keep building and go from prototype to the one where I am today the context window for AI the AI instruction set which is rows substantially over time my token usage would just be absolutely through the through the roof. And it would be more expensive I think it's sometimes realizing that now too it's now more expensive to spend on on the tokens to get the coding tasks and the coding work done then it is to have an engineer with an actual like human brain that has a substantial depth to its context window that goes in and progressively is writing PRs and pushing PRs every day so.

That those were some learnings from us and again there's there's a balance right there's a balance to be struck if it's like just these we need to go in the code base which is large now we know exactly what we need to do and these little snippets of instructions can help us expedite it really really quickly.

So yeah it's it's totally a balance for us. It's a cool opportunity for listeners to the D to C podcast to work with a SaaS platform that is in its layer of founding merchants so go to human support.io/founding merchants if you want to work with human touch human support.io and I think you know Tyler you you mentioned you know your power user of the tool you probably and you've sort of like you have this great human connection with your friend Mike sir able to kind of request. Give me an example of something that you've requested for a lot of that that Mike's gone and built at human touch using it for wholesale requests for example we had a lot of requests to sell our product and retail stores to be able to manage all that through the platform great tickets which is something we asked Mike to help us with and.

“Express your mind honestly it's yeah yeah I'm glad you brought that one up because that was something that I didn't expect and then was a bit of a light bulb moment for us because I remember actually one of our.”

founders they gave it on the team again originally and my prototype he's like this can almost be like a CRM of sorts for people and it's like David we need to stay focused like we were not a CRM and then you guys come forward to us and you're like we want to use it as like also CRM is like okay well. You're right so we built a couple of features around that so you can use it a little bit more broadly and then you know we've got we've got some of the core features like agent detection so you're not colliding with different agents.

But assigning it to other people who are in there who are not necessarily CX that was something that was sort of net new for us and I thought that was that was really really interesting.

“And something that like didn't take too much of a product adjustment to get there and add that level of utilities so it's it's great to see you guys are using that.”

And you're probably not going to get that level of customizability with some of the larger incumbents at this point which I think is a really and I think yeah D to see is one of those businesses that the more you can customize it to your exact needs and your exact customer type rather than kind of painting my numbers from what someone else thinks you should do. The better results are going to have so I think it's as I say a cool opportunity to get in on the ground floor so you got to go to human support dot IO slash founding merchants reach out to michael make sure you're your contact info's in the in the show notes here reach out to Tyler if you want to get a little bit more background on what it's like to build a four founder no employee AI enabled yet human touch business.

I got to get one of these car for I love I'm really into sense these days, but I'm also so super paranoid about phallates and these other bad things that enter your body when you have.

“You know, so I'm I'm curious to see see what a what an a lot of smells like what's the best scent.”

Most popular is the green one on my way so you could lip to Japanese see her it's quite bright very fresh a little sweet very nice second most popular is root 66 which is like very sandalwood very creamy very warm that's my go to very relaxing I mean all of all of our sensor design ought to be like super like in your face though they will start out that way for like the first day maybe but. You're very calming they're meant to just catch whiffs of while you drive the last long time and they're failed and paraben free obviously no cancer causing.

I can speak to it lasting a long time what's the blue one called again Tyler that's truck.

That's I've got that's truck and that's minivan because I've got three kids and car seats in a minivan and it's lasted like six plus months it's been amazing and my kids love to keep that that fans stinky so it's it's done it's it's been a word for us. Of course for us. Nice well thank you guys both for coming on the D to C podcast today Mike I'll maybe grab a coffee when I'm in Cambridge. Love to great meaning you guys and good luck on everything we'll make sure all your details are in the show notes and go from there.

Thanks a lot. You're so close. Thanks so much for listening to today's episode if you're not a subscriber to our newsletter you can do that right now. At direct to consumer all one word.co. I'm Eric Dick and this has been the D to C podcast we'll see you next time.