← All transcripts

The $2.7K/Day AI Business That Practically Runs Itself Transcript, AI Summary & Key Points

Chris Koerner on The Koerner Office Podcast · 3 days ago · Entertainment · 37:48 · EN

Watch on YouTube

Answer

A business selling AI customer support agents to small businesses — roughly $2,000 setup fees plus $500/month retainers, with the builds, demos and reports automated by AI — can generate around $2,700/day equivalent economics and nearly run itself; DocsBot itself passed $1 million ARR with a 90%+ gross margin and only three people involved.

AI Summary

DocsBot has grown past $1 million in annual recurring revenue in about three and a half years, and its founder spent 6 months launching eight failed or fading AI products before the ninth — AI customer support agents — stuck. The business opportunity for non-technical people: sell AI support agents to local businesses for roughly $2,000 upfront plus $500 a month in maintenance. The agent can be built with vibe coding tools and DocsBot, customers are found by scraping the web for companies with knowledge bases but no chatbot, or companies hiring support staff, and even the monthly client report is automated with AI. DocsBot itself has pivoted from selling directly to businesses toward courting agencies and implementers as a white-label sales arm, with gross margins above 90%, lifetime value a little over $2,000, and monthly revenue churn around 3%.

Key Points

  • DocsBot surpassed $1 million in ARR this year after about three and a half years, growing at roughly 3–4% per month, with gross margins greater than 90%.
  • The founder tried eight different AI products in 6 months while keeping a full-time job — a WordPress AI image plugin (viral but zero revenue), AI product photos, AI profile pictures (about $10k in two months before the fad died), and a custom AI children's book — before AI customer support agents stuck.
  • Pricing for the agency model: roughly $2,000 setup fee per client plus around $500/month for maintenance; DocsBot charges agencies $50–$200 per agent depending on white-labeling.
  • DocsBot's own plans run $50/month (personal), $150/month (most common, average customer), $500/month (business self-serve) and enterprise; ARPU is about $130, lifetime value a little over $2,000, revenue churn around 3%.
  • Cancellations due to 'I'll build it myself' rose from 1% to about 13% over the last year, but those are solo indie customers, not the good-fit companies with real support teams.
  • Good-fit customers are product manufacturers — lighting, construction panels, generators — with complex product manuals and spec questions; an AI agent can answer 70–80% of tickets, worth up to ~15,000 tickets/month, saving roughly $7,500/month versus two support hires.
  • The value rule of thumb: deliver 10x what you charge — if charging $500/month, show the client the math on how it saves $7,000/month.
  • Lead generation, ranked: in-person relationships (chamber of commerce, conferences) best, then phone calls, then cold texting, with cold email and LinkedIn DMs least effective and fading.

Business ideas

Sell AI-powered customer support agents (chat and voice) to small and mid-sized businesses: charge roughly a $2,000 setup fee and $500–$1,000/month retainer. The whole workflow — lead finding, agent building, demo generation, and monthly client reporting — can be automated with AI, so no coding is required.

For
Small and mid-sized businesses that field repeated customer questions — especially manufacturers with product manuals (lighting, construction panels, generators), local contractors, and any company with a support team or job listings for support agents
Solves
Small businesses answer the same questions repeatedly and staff support teams around the clock; AI agents can handle 50–80% of tickets, and owners lack the expertise or desire to set the technology up themselves
  • DocsBot (Aaron's company): AI agents for customer support, over $1M ARR in 3.5 years with 3 people, 90%+ gross margin, ~$130 ARPU, ~3% revenue churn, ~$2,000 lifetime value

Build small (~6-inch) shoulder-mounted robots — 3D-printed, AI-coded, connected to a voice agent — to wear at conferences as an unavoidable conversation starter and live product demo. Almost every attendee asked to buy one at ~$200, suggesting a standalone product business.

For
Anyone needing to stand out on a conference floor: B2B sellers, agencies, conference exhibitors; or end consumers who want to buy the robot
Solves
Getting attention in the age of AI where everyone has polished content; a $25,000 booth does not guarantee leads, while a robot on your shoulder pulls the whole room over
  • Aaron's robot at a conference: ~$50 build, thousands of dollars of resulting business, attendees offering to buy units at ~$200 each
🔒  Build steps and tools for 2 ideas. Unlock

Links mentioned

🔒 Full analysis locked

Unlock more videos and the full analysis

A credit unlocks one video's full analysis for good — the build steps, the tools and how each was used, the methods behind every use case. Pro opens the whole library instead, and raises how many videos you can analyse a day.

Unlock full analysis — free

Transcript

Searchable transcript of The $2.7K/Day AI Business That Practically Runs Itself — Chris Koerner on The Koerner Office Podcast (37:48). Search for a phrase, then click its timestamp to jump straight to that moment in the video.

Captions sourced from the original video on YouTube, published by Chris Koerner on The Koerner Office Podcast. The video, its captions and all related intellectual property remain the property of their respective owners; AINotes claims no ownership. Provided for research, accessibility and search — see the Transcript Notice and Copyright Policy.

00:00 hit more than a million dollars by this year. So,  super exciting. Yeah. Around a $2,000 setup fee, it's already built. You're just selling it to  them. Now, it's finally become possible for normal people who aren't technical. You can automate all  this process. You can and you should. And then the monthly retainer that you charge them on top of  that, you're not even doing anything.

00:14 Yeah. For that 500 bucks a month. It's doing it for you.  What did that cost you all in? About 50 bucks. Wow. My guest today tried eight different AI  businesses in the span of 6 months while he still had a job. And the ninth one he tried makes over a  million dollars per year of profit with no office and three people helping. But the part that should  interest you is that people with zero coding skills are charging $2 to $3,000 to set this up  for local businesses and $500 to $1,000 a month to keep running it.

00:46 And a robot writes the monthly  report. And his 19-year-old son just signed a client that said his service will save him over a  hundred grand a year. So, you make over a million dollars a year selling AI services to small  businesses. Yeah. By the end of this episode, will people have enough knowledge to start to  do the same thing? Yes, definitely.

01:02 Selling AI agents is the biggest business opportunity I  think of our generation. Yeah, I agree. We make it easy for you and I'll show you how easy it is  to automate that full process and do it without having any technical knowledge needed at all.  Okay. Well, Aaron, why don't you start by telling me who you are and what you do? Yeah, I'm Aaron  Edwards.

01:22 I'm the founder of Docspot and we help build AI agents for businesses. So, primarily  they're used for customer support. That's like a huge use case for AI agents for business that  provides a lot of value to them. Okay. How long have you been doing this? Yeah, the business has  been going like three and a half years and before that I spent 6 months just trying to launch any  product I could cuz I got that entrepreneur itch listening to podcasts like yours and being like,  hey, I need to be doing something for myself.

01:49 So tried eight different things in 6 months and  this is the one that stuck like perfect timing, perfect place, figuring out the right niche and  been growing since then. Okay. What were the ones that you launched and failed? Yeah. So the first  it was like all AI image stuff. It was right when AI images started becoming like possible and right  when it became available I was like I need to jump on this wave you know this new AI wave and images  are just so visual and everything.

02:09 So first I made a WordPress plugin kind of went viral on Twitter  X. Oh that's cool. but didn't get any revenue from it at all. Everyone loved it. No one wanted  Yeah. No one wanted to pay. Exactly. And then we kind of pivoted. I'm like, maybe this could be  like for businesses, you know, if they want to do product photos or things like that. So tried that.

02:27 Couldn't figure out how to sell it to anyone. I think the technology was still very early, you  know. And then we found the AI profile picture, viral wave. I don't know if you remember that  where like on Twitter. Yeah. Like everyone's like creating that and so we're like, "All right,  let's jump on this bandwagon because it already does that." And so that was kind of the first  success.

02:45 We made like 10 grand in like two months. Wow. Just selling AI profile pictures, you know.  Yeah, that's crazy. Yeah. So that was fun. But the fad died off, you know. Of course. So you're  always kind of tinkering, but with a full-time job. Yeah. Yeah. Exactly. And didn't didn't quit  anything yet, but Yeah. Eventually found something that worked and that was AI agents.

03:06 What was the  like the spark of the idea that led to you trying this? Yeah. just I knew that AI was kind of the  new technology wave, you know, just like in the past like I grew up like in the.com era, you know,  with websites and the web and all this stuff and uh seeing kind of that wave and how that launched  a whole new wave of businesses and then even like mobile like mobile apps and things like that and  I was like this time AI is going to be that next wave and I need to be on that.

03:31 So I need to learn  everything I can to figure out how to do it. I had no idea how to like code that stuff. It was  just like try to hack something together and MVP and see if people would pay for it. Yeah. So,  what was the very first thing you did? Yeah. So, the very first thing, well, I talked about kind  of the plugin and things like that. And then after the profile pictures kind of died out, then my  wife was like, "Hey, I've always wanted to make a children's book.

03:53 What if we use that technology  to do like a custom children's book for kids?" So, we did this one like we called it like imagine AI  and the kids would upload or a parent would upload the photos. Oh, that's awesome. And it would like  transform the children like into the pages of the like a educational like inspirational book for  them. And I'm like, this is the billion-dollar idea, right?

04:13 I mean, you're merging like uh  virtual and physical, right? Yeah, exactly. And parents and they want to see their kids  and educate and this stuff. And so we sold like an initial wave or whatever, but we could not I  could not figure out how to do ads, then convert, you know, it was like around Christmas time. I  was like, it was the perfect holiday gift.

04:29 And it was just a struggle. I was like, it's I realized  how hard it is to sell to like consumers. They have a limited budget. And something like this, it  actually cost a bit. I mean, paying like 60 bucks, 40 to 60 bucks for a book is a lot. You know, a  stretch for some people. Yeah. Because they're custom printed each one. Yeah. But we did have  some success with it and even someone discovered it in our like local magazine and Southwest Dallas  and they kind of did a cover thing interviewing us and then uh Dallas

04:59 Morning Show. Basically, they  saw that and the producers called us up and like, "Hey, we want to feature you, you know, way  on Good Morning Texas." Yeah. And we're like, "Okay, that's scary." A big uh jump. So, we  showed up there, kind of did a live interview, showed off the book. It was like really fun. But  the funny thing is like not a single sale from any of that.

05:19 Not one. Not one. Holy cow.  Yeah. Yeah. And I think it's just I mean, if people are shopping for something like that,  it's going to be like a click on a website and then then make that kind of quick purchase, you  know, and if they watch it on TV or whatever, yeah, they think it's cool, but yeah. So after  that like the real pivot into what business works was docbot and that's AI agents for customer  support.

05:39 So right after chat GBT came out, I was thinking like what is the perfect use case for  this that a business like would find value for and pay for like right away. And customer support was  one of those cuz the company I worked for before, I saw how we had so many thousands of tickets  coming in, you know, and they're answering the same questions over and over again and we're  hiring these huge teams that maybe have to man the live chat 24/7 and it's like AI can do like  most of that, you know, answer a good percent,

06:08 you know, maybe even 70 80% of the questions,  you know. What was the company you worked for where you saw this? Yeah, it was called WPMU Dev.  Okay. Yeah. And they were my first customer. Oh, cool. Okay. Yeah. How were you still working  for them when you got them as a customer? Yeah, pretty much. Yeah. It was more like a proof of  concept thing through them and then I was like, "Hey, I'm not going to like give this to them.

06:26 This is going to be my thing, you know." So, so there definitely was a transition period like  building this up and getting more customers and things like that before I finally quit and  just went full-time for myself. I mean, that's an interesting framework is if you have a  full-time job, but you want to do your own thing, what problem can you solve for your employer?

06:43 Yeah. That might make that your first time customer, right? Yeah. What did you charge them  and how did you come up with what to charge them? Yeah. Starting out, I believe we had like a entry  level plan for like 20 bucks a month and then like 50 bucks a month and 100 I think was the max or  something like that. Is that like per user or this was just like for the whole product?

07:05 So they  deploy it like in their company to like start like basically create a chatbot. That's how it started.  And you put that on your website or whatever and then it can answer your customers questions and  then if it gets stuck or it can't answer then it escalates to creating a ticket or to a live chat.  Yeah. Hey, if you're liking this episode, please smash the subscribe button.

07:24 It would mean a ton.  Knowing what you know now, if you could go back in time, what would you have charged them or like  how much could you have charged them back then? Yeah. knowing how much value you had adding and so  I think I was lucky and that I had kind of like a lot of influence from like entrepreneur Twitter  and things like that like telling me hey charge more than you think you should.

07:46 So we always had  like some bigger plans like 500 bucks a month and always had that thing like contact us if you want  enterprise and that worked out really well cuz like even one of our early customers was Sony over  in Japan. I mean, that was a huge customer. Like, what do I quote them? I have no idea, right?  Yeah. And so, I thought I shot high. Now, looking back at the numbers and what I've learned  about enterprise sales, it's like a tiny amount compared to the contracts that they're used to  signing, you know?

08:08 Yeah. But you learn. Yeah. Yeah. I spun that out and kind of shared it with  people that I knew and in that kind of space, you know, and say, "Hey, is this something you'd  be interested for your business?" And I put like a weight list sign up form on there and just  like sign up if that's what you want. And even had a survey like what were what the core features  you would need, you know?

08:24 So that kind of gave me like I got like two or 300 people on that weight  list that were like business owners and stuff. Yeah. And they told me exactly what features they  needed. So I'm like, "All right, this is my MVP, you know." Yeah. I just need to like do these  few check marks just to get something going. And I just worked, you know, 247 for like two  weeks and got it out there.

08:42 Love it. How many Twitter followers did you have then? Uh then it  was small, like a few hundred, you know, and I'm not huge now. Yeah. I don't want people to hear  this and think, "Oh, great. He's this big tech, right? That I had it close. Yeah. Like anyone  could get 300 followers, right? And that that's not to downplay 300, but it's like we all know 300  people.

09:02 We all have 500 Facebook followers. So, yeah. Exactly. And actually, what really kicked  off the product was going viral on Twitter with 300 followers. Yeah. With 300 followers. So, I  just posted like videos, demos of me building it and and stuff like that. And some tech influencer  in Japan, I have no idea how he found me, he reposted it. He like created like a bot in  Japanese and he's like this is like great AI for Japanese.

09:29 And so this is like a Sunday and I wake  up in the morning and I have like two 300 support tickets all in Japanese. Wow. And and the whole  servers, right? Yeah. And the servers are crashed and everything like that. I'm like what is going  on, you know? And turns out like it went viral. Like that tweet had like three million views  or something in Japan.

09:46 Yeah. Yeah. And luckily with AI like you can serve customers anywhere in  the world, you know. So you say support tickets, these are people that had signed up as users. Yes.  And then had support tickets about the experience and paid and then something went wrong or it  crashed or whatever, you know. I mean this was a a very rough MVP. We didn't even have like really  good support for other languages yet, you know?

10:05 I mean the AI could answer questions in that, but  the dashboard was still in English. All this stuff. Yeah. You built something and then you were  willing to post about that. Exactly. Willing to share my story. Yeah. The phrase build in public  has been played out and a lot of people like, "Oh, there's no benefit to building in public." I've  heard that a lot recently.

10:24 Like, it's just you invite competitors. And I could not disagree more.  I don't care if no one if I get three views on this, I'm going to post it regardless because all  it takes is one dude in Japan. Exactly. You never know. Yeah. You never know. And he he used it and  tried it and posted about it because he liked it. What a great story. Okay. So, that was like month  one.

10:42 Yeah. Okay. Mhm. And so you had hundreds of users at this point or thousands, I mean hundreds  of support tickets. So yeah, probably like uh hundreds of customers. I think we got like 3  months we hit like 10k MR and that was where I was like this might actually be a thing. And you still  had a job. Yeah. Yeah. I might actually be able to quit, you know.

11:04 And so like I talked to my boss  and he was real sweet about it. is like, "Hey, like I'll you can like do like part-time and just  help like with the transition, like train up like your replacement over the next year, you know,  and just act as like more of a consultant kind of role." So, that was awesome. So, I had the time  to kind of focus on the business more and still have like a good chunk of my old salary just in  case, you know.

11:22 Yeah. Yeah. So, all right. Could you have like today, could you build the software  that you have now with like a vibe coding app? Is it possible? You definitely can to some extent and  that's 80 90% of the way there. Yeah. And that's one of the hard things that we've had like we've  had some I was looking at my numbers of people who cancel, you know, cuz we're software as a service.

11:45 It's a monthly subscription kind of deal. Mhm. And what we're seeing like over the last year is it's  increased from like 1% like I'll build my own to like about 13% of people who cancel. Okay. That's  the reason why. But what I found the people who cancel like that, they really weren't a good fit  customer anyway. Sure. Like they're just maybe a single person.

12:05 Yeah. It's not worth the cost  to them. An indie developer. Exactly. Right. A soloreneur. Exactly. Yeah. Those are great. Those  are awesome. But for your business, they weren't good customers. Yeah. And the good customer is  like some a company that's big enough that they have like a support team. They have employees.  Yeah. People need a dashboard to log into where they can see the stats and improve it over time.

12:24 All the stuff that like vibe coding doesn't like really work for. Like you can get that proof of  concept or something janky out there, but you need someone that can maintain it constantly, keep it  going, all that kind of stuff. I'm smiling because it's the customers of yours that have the most  resources, the most talent, the most people, the most money, and the most money to save by building  it themselves are the ones that are not doing it.

12:49 Exactly. They're just writing a check every  month, which is great for you. Yeah. And it's the hustlers out there that you said it went from  1% of your cancellations initially said I'm just going to build this myself to now 13%. Yeah. Yeah.  But surely that's a much much smaller percentage of your lost revenue because they're small  customers. Yeah.

13:07 Do you have a free plan? We do. We always put kind of a free plan just so we could  say like no credit card required. And one of the things I love that I did at the first is I didn't  have like a free like trial or whatever. It was just like you pay us and show us you're serious.  We'll give you a refund if it doesn't work for you. Yeah. But like show us you're serious.

13:24 And  that as like a single founder that probably saved us I can't even imagine how much time you know.  Yeah. And avoiding hiring people and stuff cuz only like serious people uh would actually  like sign up and try it you know. What are your pricing plans today? So we have kind of like  a personal geared one for like that someone just getting started kind of thing or a tiny business  like at 50 bucks a month.

13:43 And then our main plan, and this is kind of our average customer is on 150  a month. Okay. Um and then we have like a business self-s serve that's 500 and then enterprise if  they need some crazy stuff above that. Okay. So 150 is your most common plan. What is your average  monthly price? Yeah. Our ARPU a revenue per p per user is I think like about 130 something like  that.

14:07 Yeah. So it's very close to that standard plan. And what is your churn and lifetime value?  Yeah, lifetime value, last I checked, it's u it's a little over $2,000. Wow. Yeah. And then churn is  like our revenue churn is around 3%. So Wow. Churn has always been hard. We've done a lot this year  to like improve that. Um but there's some things you just can't fix.

14:31 You know, if there's the wrong  customer, it's the wrong customer. Yeah. Well, if they stick around for a month or two, they're  probably going to stick around for years, right? Yeah. That's the idea. Okay. Do you provide  any services? Like do you build custom agents for people or just you offer your software? Yeah.  So here's kind of the shift that I've seen in the last year is before like we got our customers  like classic like software like SEO you know word of mouth things like that you know and we've  seen that

14:55 drop so much and so this year we've really had to push into like more of like a  service like actual talk to people you know and make it happen for them. Now we can use AI on  the back end to automate 95% of that but and this is the opportunity for everyone watching here  is that like building AI agents for like small businesses like most of them they don't have the  expertise they don't want to learn even like how to use a new software no matter how simple I make  it.

15:21 Yeah. They're installing water heaters. Yeah. Exactly. So they need that person that can  take like okay here's the problem they have and here's how I can use AI to solve it with the  solution. And that's really where the opportunity is. Yeah. Are you doing any of that? Yeah. For the  company, we've kind of pivoted a lot to that. So, right now, like most of our effort goes into like  courting agencies and and implementers instead of like going directly after customers because the  customers who are kind of technically

15:46 savvy enough to like sign up and and set it up themselves,  no matter how easy we make it, you know, that's kind of dried up and there's so many other  competitors out there. So now we have kind of a service offering like where you have like a white  glove setup and then like a maintenance for the first 90 days or whatever you know to to show them  that value and really get them onboarded and then all our agency partners who are spinning up agents  and showing them to small businesses for us you know and making a lot

16:14 of money in the process.  So are they white labeling that stuff to you? Yeah. Okay. So you're doing the work on the back  end and they're they're your sales arm basically your referral partners. Yeah. So there's different  levels. Sometimes they're like white labeling it. So they're just having signing up with their own  account from us. Mhm. Reselling bots, you know, cuz our platform supports teams and multiple bots  and things like that.

16:32 And so they'll create like an agent, do a pilot for a business, sell it to  them, upcharge like a big amount. They can have a monthly like fee on top of that for like service,  maintenance, reports, all that kind of thing. So the business doesn't have to worry about anything.  Okay. What are they charging for it on average? Yeah. So, the a good price point we found for  like a good fit customer is roughly like around a $2,000 setup fee.

16:56 That's what your agencies  are charging their business customers. Yeah. $2,000 setup fee. Mhm. Okay. As much as 500  bucks a month like like maintenance. So, and it's going to depend on the size of the business.  Obviously, you have to think about like what value you're providing them too. If you think maybe the  average like support agent that a company hires, maybe their salary is like 40 grand a year, you  know, that's probably like base like 20 bucks an hour and only goes up from there.

17:20 And if you  think like each support ticket like save or it takes about 5 minutes of employees time, right,  to answer you can kind of like math that out, but most companies I mean with the plan we provide  you can do as much as without paying any extra as much as like 15,000 tickets a month. Mh. So I mean  you're talking about like $10,000 exactly dollars a month that they're saving.

17:42 They'd be paying an  agency 500 bucks a month and theoretically saving 7,500 bucks a month. Yeah. Exactly. Two full-time  employees. Exactly. Yeah. Okay. Did you know that beef prices have never been higher? It is a great  time to get in the beef business, but you need people to find you. So whether you're selling  to businesses or consumers, you're going to need people to find you.

18:01 And getting the perfect  domain name is a great start. So, don't let your.com obsession kill your momentum. And I've  been there. I get it. You think of the best name like grass-fedbeef.com. Check if it's available.  It's not. So, you're going to start adding words grass-fed beef north Texas or throwing in hyphens,  adding fluff that will ultimately prevent you from getting more clicks, and suddenly your clean,  memorable brand name turns into something that nobody can even remember or spell.

18:27 Meanwhile,  that exact name that you wanted, grass-fed beef, it's probably available right now atonline. And  here's the part that matters for getting found. The word online shows up in keywords searched  over 500 million times every single month. 500 million. That's a little boost of SEO working in  your favor just from your domain name. Online is the second largest new domain extension with over  three and a half million businesses using it.

18:51 So, stop overthinking it. My listeners can use code  kerner koe r ne to get theironline domain for only 99 cents or you can just scan the QR  code in the description. Grab your name, launch the thing, and let's keep moving. And  then for that 500 a month or whatever, what are you charging the agency? Is it just your normal  software fee like 50 bucks or 150 bucks?

19:12 Yeah, like between like 50 if if they're white labeling  it, then between 50 and $20. Yeah. Per agent. So what's changed with you is you're still a software  company. You're just trying to better sell your software to service companies. Yeah. Whereas  before you were a software company selling your software to businesses directly. Is that right?  Exactly.

19:33 Yeah. Okay. And then there's people that are just starting from scratch like my son. He's  only 19, you know, fresh out of high school and he started shilling AI agents to companies and he's  got his first big contract that he's working on right now like setting everything up for them and  he doesn't know how to code. But I mean he's vibe coding the whole thing and using Docspot for some  things and stuff like that.

19:52 What type of business? This one is a contractor. So just local. Yeah,  local contractor family friend or someone I I knew and he's like hey I need AI so bad. He's like  here's my son. Like he'll he'll help you out. You know what's he going to be able to charge for  this? Gosh, I'm not sure what it's going to be, but like the value that he mentioned to me,  like this this contractor, he's like, I could see this saving me like $100,000 or more a year.

20:17 I'm like, your son's like, "Yes, your son's like, I'll charge him 500 a month." And you're like,  "No." Yeah. I was like, "Let's think about that." Like, talk to me before you pitch him. A good rule  of thumb, I've heard, and this is for software, like it may be different for other businesses,  but you want to provide 10 times the value that you charge.

20:36 Yeah. So, if you're charging 500 bucks  a month, make sure that they're getting $5 a month and walk them through the math. Right. Exactly.  All right. So, you said five minutes per ticket, right? Right. 2,000 tickets per month. Right.  Right. I say conservatively, we're going to knock out 50% of those. Probably 70, but let's  be conservative. We're going to save you $7,000 a month.

20:53 We're only going to charge you  500. Exactly. That makes it kind of a no-brainer offer, right? Mhm. What are your numbers today?  And how have those grown over time? Yeah. So, it's been three and a half years. We hit more than  a million dollars uh in ARR. Holy crap. This year. So good for you. Super exciting. Yeah. How much  of that is Well, I was going to say how much is service versus software, but it's all software.

21:12 Exactly. You're selling to service businesses, but you're still software. Yep. Okay. Yeah. What  would you have done differently from the outset if you could go back in time? So, I kind of had that  kind of bootstrapper mindset like, you know, make the profit like take the profit, you know, sock it  away. Like I would have like as soon as I was able to like financially, I would have like reinvested  a lot more like in ads and like faster growth and things like that.

21:35 I never wanted to go down like  that investment path like the kind of VC thing, right? Even though in the beginning like of course  they're calling me because they're like want to get in anything AI, you know, at that point in  time. Um maybe that was a bad decision. I don't know. But I love the lifestyle and that's what's  important to me. I prefer the lifestyle business.

21:53 Yeah. Yeah. So for me now it's like I can work  as much as I want, you know, usually like 20 to 40 hours a week, but like this month I probably  took like six weeks of vacation and this summer, you know, like interspersed between because  so much is automated now with AI and I just love that. And now we've grown. I hired like my  best friend like from high school going back, you know, best man.

22:14 And he's kind of joined the  company. So it's great not being alone anymore, just having him like helping out. And my older son  is working for us too, fresh out of college and he's like learning marketing and sales and and how  to like get by in this world and it's great having that opportunity and you can't do that maybe if  you're like a VC back hyperrowth kind of company, right?

22:34 What are your margins? Yeah. So the net  margin like it's greater than 90% is our gross margin. Yeah. So do you have employees? Yeah, just  that one partner who I've hired, my friend and my uh oldest son. That's awesome. And but how long  did it take you to get employees? A long time. you were you were making really good money without  any employees.

22:50 Exactly. Yeah. For a long time and I didn't necessarily have to hire them, you  know. It was more like a choice like I don't want to be alone anymore, you know? Yeah. All right.  So, how much are some of these agencies making um doing these services for small businesses? What  have you heard from your customers? Yeah. I mean, I've heard all kinds of numbers, you know.

23:10 I  mean, like that mid price point I've heard is like two to three grand for setup and then $500 a  month, you know? Mh. But you think about that that recurring revenue it stacks you know especially  like when you add that service part to it and what works best that I found is like usually like it's  an implement agency or if you want to start this for yourself like figure out what niche you want  to target and what's cool about this AI agents is it works for almost any business and even customer  support like any

23:33 business that gets any kind of like customer support whether it's tickets live  chat phone calls whatever they need this right and so it applies to almost anyone but we have  everything from like a good fit would be like manufacturing like product or manufacturers. So,  we'll have like companies like a like a lighting company or one that sells like construction like  panels or one that sells like generators, right?

23:58 And they have all these product manuals, you  know, that have all these crazy tables and specs in them, you know, and they have customers asking  questions all the time about the specs for this or this, you know, will this work for this use case,  etc. And that's a perfect case for AI. Do you know how your these service businesses are finding  them specifically?

24:12 Are they going to their family, friends, Facebook ads, like cold outreach on  LinkedIn? What are they doing? Yeah, I'll tell you some good techniques that I learned. So, like in  order in person, you can't beat that. You know, if you have a relationship, you meet someone at the  chamber of commerce, at a conference, whatever, like that's like the best lead you're going to get  because they get to know you face to face.

24:31 So, we do a lot of that ourselves, you know, and that's  what's works works for those agencies as well. Uh, next I would say is like a phone call. Like  pick up the phone, call a business, you know, like you want to save money with AI, you know, I  can help you like and you don't have to be like an AI expert. You just have to know a little bit more  than them, right?

24:49 Yeah. Yeah. Yeah. And then text, you can do cold texting, you know, that's  probably like next best and email is possible, but it's down there. And that's a that's a very  hard uh thing to do. Yeah. So, what are you doing at conferences? What what is this guy right here?  Yeah, this guy is fun. I don't know how to build things in real life. I just sent this to my AI  agent, coding agent, and I said, "Hey, create like the 3D models for me to print and like tell  me how to plug everything together and what to buy on

25:17 Amazon." And the idea was like we were going  to these conferences like in person, making leads, connecting with agencies, whatever. I was like,  I don't want to pay 25 grand for a booth. I mean, that's expensive. And I don't know how I could  ever pay that off no matter how many leads I get, right? So, how can we like grab attention on the  conference floor?

25:32 And so I had AI help us build these little guys, these little robots and they  just sit on your shoulder. There's actually like unbag shoulder. Okay. That changes everything.  Yeah. Cuz if you have a big table and it's sitting there, you might not even notice it. Exactly. Cuz  I mean we're like 6 in tall. Yeah. Okay. So you put it on your shoulder. Yeah.

25:52 So it mounts to our  shoulder like with magnets. So we're just walking around. Everyone in the whole place magnet under  your shirt. Yes. Exactly. Yeah. Everyone in the whole place, they came up and talked to us. You  know, you can't miss the the the two guys walking around with a robot on their shoulder, right?  Cool. And it's a perfect kind of like conversation starter.

26:12 And it also we actually had it connect  to our voice agents uh with Docspot. So, you can actually like talk to it and like have a  conversation. It's kind of like a live demo of our product. Yeah. So, what does it do? Does it talk?  Yeah, it can ask me a question, tell me a joke, or tell me about your business. So you can interact  with it like Yeah, you can talk to it and interact with it and things like that.

26:32 Yeah. What model?  Super cool. What model are you using for that? Uh internally it's powered by Docspot and we're using  the latest like open AI like voice real time live models you know. Okay. What did that cost you all  in? About 50 bucks. Wow. I mean a lot of time the first one surely with the help of AI. Yeah. How  many do you have? I made three for now but I mean we can make that could be a whole business.

26:54 I know that could be a whole business. You sell those for 200 bucks. Almost everyone on the  floor was like, "I want one of those." I'm like, "Maybe we're in the wrong business. We should be  selling. We should just have a second business." Right. Exactly. I know you're you're probably not  exactly, but if you were to back of the napkin attribute x number of sales or monthly recurring  revenue to that little guy, Yeah.

27:09 what would you say it is? Well, so far we've taken them just to  one conference. So, we probably got like 10 to 20 leads from there. Okay. Um at least two customers  so far. And we're we're still talking to all those leads, you know, and each customer is worth  2,000 bucks. Yeah. And a lot of these there were like agencies and stuff. So, we're not necessarily  selling to them.

27:30 We're trying to help them sell to their customers, right? Yeah. Well, that's quite  the uh the ROI. Yeah. I mean, it's thousands of dollars from a $50 investment. And I love that  you just have to think different. Like that's what marketing it is. It's like being unique. It's  not being like everyone else. And even in the age of AI where you can make like amazing videos,  amazing content, amazing websites, whatever, you got to figure out how to stand out.

27:54 So you got  to do something different. I mean, people could do that for any business. It's just a conversation  starter. Exactly. That's a what a great takeaway. Yeah. Are you or your customers building any  other types of AI agents or is it still like customer support? Yeah, so our platform is kind of  flexible like it can be used for like sales teams, collecting leads.

28:13 We have voice agents too. So  the same like bot that has that can shout on your website. You can call it with a phone number um or  connect it to your own systems or app or whatever and and do the full like voice thing as well, you  know? But that's not really our target is like like the classic like blue collar service business  that needs like a AI receptionist like you can do that probably for cheaper like would go high level  or something like that.

28:32 It's more like if you want that complex like you got product questions,  things like that that are complicated um policies and things that you need to make sure that AI  follows all that kinds of stuff. We're really good at helping with that. Okay. But like the 8020 of  it is customer support. Yeah. From day one until today to most people. Yeah. We've really had to  niche down on that because before it's like, oh, you can use it for your internal team.

28:54 And that  is a good use case. Actually, some of our bigger customers, they don't even deploy it to their  customers publicly. It's used internally like for their team. Maybe there's a HR bot that people can  ask benefits questions within Slack or whatever, you know, for the team or different use cases like  that. So there is definitely more niche than just that customer support.

29:15 Please, I need you. I need  you to send me your stories. If you or someone you know made thousands of dollars in a short  period of time with a business, a side hustle, a growth hack, I don't care what it is. I want  you on this podcast to give you free publicity and to help grow my podcast. So if that's you,  go to tko yes.com, fill out the form. If it's a friend of yours, send them tk.com.

29:36 Have them  enter your name and I'll give you a,000 bucks for the referral. That's tko yes.com. What models  do you recommend for customer support bots? So, it's really just comes down to like a cost benefit  ratio. Um, and they're changing all the time. So, we're always like integrating the latest stuff and  figuring out, okay, what's the best like balance of cost and quality for this use case and we  do all the testing internally, so you don't really have to worry about that.

30:02 But, I love like  some of the newer OpenAI models like uh GPD6 Luna and stuff. They're so cheap and efficient and  super smart compared to like when we launched. Yeah. What do you think like the killer offer is  people can give to small businesses? Yeah. So the kind of steps that that we do for this and that  that people who sell this is you need to figure out like first make sure that like they have  knowledge.

30:25 So you can use AI agent to like you know scrape websites, find leads or things like  that and it's going to find those businesses. Maybe they have a knowledge base. Maybe they  have like a old style chatbot on their website, you know, and you can use like BuiltWith or  things like that that actually tell you, okay, here's all the websites that are using this Help  Scout widget or whatever.

30:41 And there's your niche right there. You can target all of them. Say,  hey, uh, Help Scout, you have customers asking questions. Help Scout just points them to the docs  or whatever. We can give you AI agent that's going to solve like 70, maybe even 80% of your tickets.  So, that frees up your team. You have to hire less people. But you can prove that value, you know.

31:00 Yeah. What were your sales over time? So, it's been 3 years. You're at over a million in annual  occurring revenue now. How did that grow? Yeah. So, it was kind of almost like a a straight line  the whole time of like 3 to 4% per month growth. Okay. Which is great. It's not like a hyperrowth  exponential thing, but that was great. And even like big companies like OpenAI, Anthropic, like  they've hired these big agencies to implement AI.

31:25 And we can build the best agent or software or  model or whatever, but it's still not going to get into the hands of these small medium-sized  businesses unless someone actually goes there, talks to them, helps implement it for them. And so  they're actually investing billions of dollars in that as well. They know the deal. And that's kind  of the opportunity, you know, is is shilling AI agents to business.

31:44 And I think it's still going  to be an opportunity in 10 years. I mean, there's 30 million businesses in America. Yeah. Yeah.  So, $2,000 upfront. Yeah. $500 ongoing is could be more, could be less. What all is included with  that? And how much time do people spend on that client every week or every month? Yeah, exactly.  That's that's the thing. And with AI, it's like become more and more automated.

32:05 So, essentially  what we do is what we've built it is we made it easy for our implementers. We have like a plugin  for Chad GBT, Claude, Grockbot, whatever you use. link. That's the best $20 subscription you can  buy. By the way, you know, chat GV2, your cloud automates so much with that 20 bucks. But you  connect it to our platform and you can just say, "Hey, here's this business website."

32:26 Or actually,  you could say, "Go find businesses, manufacturers that are doing this and that have a knowledge base  on their website, but no chatbot. It goes out, finds those clients for you." What model would you  use for this? Sorry. So, Cloud or OpenAI, Fable 5.1, Opus 5.5. Yeah, I mean the models change  every day. Right now, Opus 5.5, that's amazing.

32:49 And then the GPT uh like six, Luna and Opus, those  are great. Yeah. And you get tons of usage from that for your 20 bucks a month. Yeah. So 20 bucks  a month, ask it to go find these customers for you. Then you're going to ask for uh like point of  contact, I imagine. Yeah. So you do need to figure out like how you contact these people. So, one tip  that I found is I used to like pay like for HRES, pay for like Apollo to get leads, you know, all  these different services.

33:13 There's one that I found is called TEG and it's essentially a router for  all these APIs. How do you spell that? T R E G T. Okay. Yeah. So, Google Zapier kind of. Uh, yeah.  Or open router if you're familiar with that. So, basically, you just put like deposit 20 bucks of  credits in there. It already has they're reselling API access to Apollo to all the SEO providers.

33:36 You're going directly to the source. Yeah. All the things where you collect get information about  leads and stuff like that. And so you just attach that to your chatbt or claude and then suddenly it  has like a hundred different tools where you don't have to pay for all these subscriptions. That's  amazing. Yeah. And you only pay whatever it's like one cent per lead that it needs to scrape  or whatever, you know.

33:54 That's like an MCP. Yeah. Right. Or a connector. Yeah. Right. Yeah. You're  you're getting an account on Tre. You're putting in the URL, connecting it to ChatGpt or or Claude.  Yeah. And then you're saying like, "All right, what is the position at this company? Who is the  person? What is their contact in?" It'll find that for you. Another another good technique is look  for people who are hiring customer support agents and that can do that too.

34:22 is it can scrape, you  know, like LinkedIn and different job websites and find that cuz you know they're about to invest a  bunch of money in a new agent when you could sell them instead an AI for like much much cheaper  that's going to allow them to scale infinitely without hiring more people. So they already got  the budget, they're already ready, you know, they have the the ticket load.

34:40 Um they're a perfect  opportunity to target. If I were prompting this, I would go a step further and say, "All right,  LinkedIn's great, email's great. They're already getting blown up there, I'm sure. Mhm. I want  to know their Strava link. I want to know their PayPal. I want to know their Venmo. I'm going  to send them a dollar on Venmo. Get deep.

34:54 Yeah, absolutely. Because you got to stand out. You  got to be unique. If I get their home address, I'm going to mail them a box of chocolates or  something, a box of donuts. I'm going to door dash them some tacos. Like $2,000 per customer.  It's worth investing a little bit because your your close rate isn't going to go from 0.1%  to 2% if you do these unique expensive things.

35:19 It's going to go from.1 to like 12%. Right. Like  it really makes that big of a difference. Wow. If it's if it's cheap outreach, then it's just that  it's cheap. Yeah. Cold email is free. It's cheap. DMs on LinkedIn. So, it's like not effective at  all anymore and getting less. So, yeah. Yeah. And the great thing with AI is you automate that  whole thing.

35:38 You know, the outreach things, even creating things in real life to mail to them,  you can automate all of that. Oh, yeah. the tools and APIs and things like that. And then if you're  using like Docspot, I'm sure other providers, we have competitors too that you can use different  price points, but we have like an MCB plugin as well. So you can say, "Hey, find those customers  now.

35:56 I want you to create a bot based on their help desk or they've scraped from the website.  It'll automatically create that for them. Create a demo link that going to automatically send to  them. So they can just click and see it already working." Yeah. On their content, on their  docs. So you don't even have to do anything, you know? Amazing. And then you're just selling  it.

36:13 It's already built essentially. You're just selling it to them. Now it's finally become like  possible for normal people who aren't technical to to use these kind of agent employees. So there's  no excuse anymore. No, you can automate all this process. You can and you should. Yeah. And then  like the monthly retainer that you charge them on top of that like the value you're providing there  is you're showing them every month like how much value you're providing, you know?

36:37 So you're  telling your you just say, "Okay, Chad GBT or Claude, I want you every month this is my client.  I want you to go through, look at all their stats, their logs, analyze it, create a nice PDF report  with my logo on it, you know, and email it to this customer. So, you're not even doing anything.  Yeah. For that 500 bucks a month, it's doing it for you.

36:55 It's identifying like, okay, here's  questions I couldn't answer. So, now you're going to ask that client say, "Hey, you're missing  a support doc for this topic." You know, uh, you should make that. It'll make your bot better.  You know, things like that. You're providing that value and proving it every month, and you  don't even have to touch anything.

37:07 Yeah. Aaron, this has been amazing. Thank you for your time.  Yeah, thank you. Where can people find you or your businesses? Yeah. Well, if you want to follow me  on X, I'm uglyroot. Just one word, ugly robot. And I'm always building a Pelug talking about what I'm  building and AI. And if you're interested in like creating AI support bots and selling those to any  kind of business at all, uh we'll help you out.

37:32 We got a partner program. We can help you with  everything. And that's at docbot. That's docsb.ai. All right. Thank you, Aaron. Yeah. Thank you. If  you like this content, please share with a friend, hit a like or a comment below, and  we'll see you next time on the Kerner