WEBVTT 00:00:00.160 --> 00:00:16.585 Normal people are making millions of dollars with AI, and they don't need to make faceless content or sell an app that they made for cheap. There's a new niche that's quietly printing millions of dollars that nobody is capitalizing on yet. I know about this niche because I'm one of its customers. 00:00:16.825 --> 00:00:18.825 I'm David. I run Filterbuy, 00:00:18.825 --> 00:00:21.305 an air filter company doing $23,000,000 00:00:21.305 --> 00:00:26.745 a month. The biggest AI opportunity in the next five years isn't building sexy startups. 00:00:26.825 --> 00:00:32.010 It's building boring software for boring industry. And I'll show you exactly why, 00:00:32.330 --> 00:00:43.610 exactly what to build, and exactly how to get started. Let's start with the biggest unlock. Every boring business you know, whether that be the HVAC company, the local manufacturer, 00:00:43.955 --> 00:00:50.915 or the freight broker in my examples, is running software that doesn't fit. That mismatch is your opportunity. 00:00:50.915 --> 00:01:03.440 The entire tech industry is quietly panicking because they know that AI is about to cannibalize them. I want to give you a concrete example from my own company. We used to pay a software vendor about $10,000 00:01:03.440 --> 00:01:12.975 a year for a connector, a tool whose only job was to move transaction data from Stripe over to QuickBooks. That's it. One very boring task. 00:01:13.135 --> 00:01:25.220 $10 a year. I actually built the replacement myself. I understand our data structure. I understand what we need. And with AI tools, I just did it on a weekend. One day, done. That's $10,000 00:01:25.220 --> 00:01:28.340 a year back in our pocket permanently. 00:01:28.340 --> 00:01:42.355 That price drop doesn't help big companies build better tools. It blows open an entirely new market. The market of niches that were always too small to justify custom software before. Out of a warehouse in Alabama, 00:01:42.515 --> 00:01:50.115 I'm running my own personal tech company. AI didn't replace my team. It made 12 developers do the work of a 120. 00:01:50.460 --> 00:02:01.660 Filterbuy is an air filter company. We're not a tech company, but we've always bet on technology as a competitive edge. Here's what AI is doing for us right now. The biggest one is data. 00:02:01.820 --> 00:02:08.715 We pull information from a lot of different places, things like Amazon or our Amazon ad platform, Google Ads, 00:02:08.875 --> 00:02:09.515 Meta, 00:02:09.675 --> 00:02:10.555 our website, 00:02:10.795 --> 00:02:13.835 all of our internal systems, our shipping platforms. 00:02:13.995 --> 00:02:14.635 Before, 00:02:14.875 --> 00:02:21.280 consolidating all of that into one place where you could actually analyze it was a major engineering project. 00:02:21.360 --> 00:02:40.165 You needed data engineers writing pipelines, designing schemas, cleaning everything up. That kind of build could take months. We rebuilt most of that stack with heavy AI assistance over the last six weeks. Now we have a real time data warehouse that shows us contribution margin, customer acquisition cost, ad efficiency, 00:02:40.245 --> 00:02:57.630 and profitability by product across all of our channels at once. What used to take weeks of engineering time can now be done in a day or two. We're also using AI in customer service, marketing efficiency, pricing decisions, and product development. Every part of the business that touches data 00:02:57.630 --> 00:03:14.685 has gotten faster. The way I think about it is this, AI shortens the feedback loop. You have an idea, you test it, you improve it, and then you move on way faster than before. That speed compounds and it becomes your competitive advantage. It's one of the biggest advantages a smaller, 00:03:14.765 --> 00:03:34.255 nimbler company can have right now. If you wanna understand where AI adds the most value in a boring business, start with the data. Every business generates more data than it can process. That's the pain point, and that's where you need to start. You might be thinking, these big tech companies are huge with billions in resources. 00:03:34.415 --> 00:03:37.135 Won't they just add AI and crush everyone? 00:03:37.295 --> 00:03:46.690 I actually don't think HubSpot and Salesforce are going away. Let me be clear about that. But here's the crack in the wall. Look at the financials of any big SaaS company, 00:03:46.850 --> 00:04:05.625 whether it be Salesforce, HubSpot, others like them. Lots of examples. About 50% of their revenue goes to sales and marketing, just getting customers to buy. That's how expensive it is to sell one size fits all software to everybody. That business model made sense when it was the only game in town. When the alternative was spending $300,000 00:04:05.625 --> 00:04:09.625 to build something custom, paying a $100,000 a year for Salesforce 00:04:09.625 --> 00:04:20.310 was an easy yes, but the alternative just got 90% cheaper. Now a smaller player can come in and offer an HVAC company something better, software built specifically 00:04:20.310 --> 00:04:33.945 for how HVAC companies actually work at a fraction of the cost. And they don't need a massive sales team to do it because they're only selling to one type of customer. The big players will feel margin pressure. They'll add AI features to stay relevant, 00:04:34.185 --> 00:05:16.480 but they'll struggle to go deep on any single vertical because their whole model is built on going wide. Vertical software for niche industries is the play that's hardest for them to copy. Someone is going to build the sales force of HVAC companies, the sales force of freight brokers, the sales force of pool fence franchises. That person doesn't need to be a genius, and they definitely don't need outside money. The question to ask yourself is, what industry do I already understand? Your industry knowledge is the moat. AI is just the tool. I know a guy who built custom software for his franchise business that five years ago would have been completely out of reach. It changed everything about how his company operates. 00:05:16.720 --> 00:05:35.875 Once he discovered AI, he started making millions. On my podcast, Boring Money, I talked to a guy named Eric Lupton who runs a pool fence safety business with a number of franchisees across the country. Pool fence installation is about as unglamorous as it gets, but running a franchise network is operationally complex. 00:05:36.330 --> 00:05:38.730 You've got franchisees in different states, 00:05:38.890 --> 00:05:40.970 different jobs, different pricing, 00:05:41.050 --> 00:05:44.330 different customer histories. Eric built custom software 00:05:44.330 --> 00:05:45.290 specifically 00:05:45.370 --> 00:05:46.810 for his business, 00:05:47.130 --> 00:05:57.695 software that was designed from the ground up around how pool fence franchise operations actually work. Things like quoting or job tracking, franchisee dashboards, customer records. 00:05:57.935 --> 00:06:02.015 Five years ago, that project would have cost him somewhere between a 150,000 00:06:02.015 --> 00:06:03.855 and $500,000, 00:06:04.015 --> 00:06:10.140 and after that, he would have needed ongoing developers to maintain it. Because of AI assisted development, 00:06:10.300 --> 00:06:15.260 the same project today is a fraction of that cost and can be built in weeks, 00:06:15.340 --> 00:06:24.115 not months. And here's what I want you to notice. There is almost certainly no venture backed startup that's building software for pool fence franchises. 00:06:24.435 --> 00:06:40.010 The market is too niche. It's too boring, and that's exactly why it's an opportunity. If you wanna watch Eric's full episode, subscribe to watch boring money on this YouTube channel. If you have a similar business story and want to come on the show, click in the link in the description. 00:06:40.250 --> 00:06:42.250 Now let's get back to the actionable. 00:06:42.490 --> 00:06:52.175 Here's my exact playbook to capitalize on this. The best skills, the biggest mistakes to avoid, and three product ideas that you can launch this week. 00:06:52.255 --> 00:06:56.735 Here's the thing that most people get wrong about building software for boring businesses. 00:06:57.055 --> 00:07:04.900 You probably don't need to be as technical as you think. The most important skill for building boring AI software isn't coding. 00:07:05.140 --> 00:07:09.700 Now I wanna be honest with you about something. I'm not a professional developer, 00:07:09.860 --> 00:07:14.340 but I personally rebuilt our Stripe to QuickBooks connector myself. 00:07:14.660 --> 00:07:24.635 One day, done. The cost of entry into software is coming down fast. You don't need a computer science degree. What you need is to understand a problem 00:07:24.795 --> 00:07:27.595 deeply enough to describe it clearly, 00:07:27.675 --> 00:07:36.440 and then you use AI tools to help you build toward a solution. That being said, I'm not gonna tell you that coding knowledge doesn't matter. It does. 00:07:36.760 --> 00:07:41.480 Someone who understands systems can work with a developer, can review what gets built, 00:07:42.085 --> 00:07:51.045 and is gonna move faster than someone who doesn't. But the most valuable skill in this space right now isn't coding. It's industry knowledge. 00:07:51.205 --> 00:08:01.630 Ask yourself what industry you already know from the inside, and that becomes your starting point. The skill you need to add is enough technical knowledge to build a solution 00:08:01.790 --> 00:08:19.845 or clarity to direct someone who can. Before you run off and start building, I want to tell you about the mistake that I see over and over because it's the fastest way to waste a year of your life. The biggest mistake people make is starting with the technology. They see AI tools, they get excited, 00:08:20.005 --> 00:08:36.010 and then they start trying to find a problem to attach them to. That almost always leads to building something generic that nobody really needs. I saw a post on x a few weeks ago from someone who had just been to a medical conference, a room full of doctors running independent private practice. 00:08:36.505 --> 00:09:01.640 Not one of them had seriously thought about how AI could make their practice more efficient. Nobody in that room even knew what some of the basic development tools were. That's how early we still are. The opportunity isn't in building another AI chatbot or another generic workflow tool. The opportunity isn't being the person who walks into that room of doctors, understands their specific problems deeply, 00:09:01.800 --> 00:09:08.445 and builds something that fits exactly that. So what's the wrong play? It's building horizontally, 00:09:08.525 --> 00:09:21.940 trying to build AI for marketing or AI for operations or AI for small business. Those markets get crowded fast because the problem definition is so vague. Big, well funded companies will eat you alive. 00:09:22.020 --> 00:09:24.420 The right play is building vertically. 00:09:24.580 --> 00:09:36.125 Pick HVAC. Pick freight brokerage. Pick dental practices. Pick Amazon brands. Go so deep into one niche that you understand it better than anyone building software for it. That depth 00:09:36.125 --> 00:09:52.690 is your moat. The big players can't replicate it because they're trying to serve everybody. Three questions to ask before you build anything. Number one, would I have paid for this when I worked in this industry? Two, can I explain the problem and solution to a novice in a few sentences? 00:09:52.850 --> 00:09:53.490 Three, 00:09:53.650 --> 00:10:01.650 is this boring enough that a venture capitalist wouldn't fund it? If the answer to all three is yes, you're in the right place. I said I'd be specific. 00:10:02.065 --> 00:10:11.025 Here are three boring AI businesses I would actually build if I were starting today. Number one, HVAC dispatch and sales intelligence. 00:10:11.105 --> 00:10:19.790 Most HVAC companies with five to 50 trucks are still scheduling jobs on whiteboards or in basic software. Technicians drive inefficient routes. 00:10:20.030 --> 00:10:34.885 Upsell opportunities get missed on every call. You could build an AI layer that sits on top of their existing system and does three things, smart dispatching based on geography and skills, automatic transcription and classification of inbound calls, 00:10:35.045 --> 00:10:47.570 and sales prompts for technicians based on what they find at each job. The customer is any HVAC company doing 5 to $50,000,000 in revenue. You could charge $500 to $2,000 per month per location. 00:10:47.730 --> 00:10:55.570 Improving technician utilization by even a few percent more than covers the cost. Number two, freight broker copilot. 00:10:55.570 --> 00:11:00.530 Freight brokerage is a massive industry that still runs largely on phone calls and spreadsheets. 00:11:00.855 --> 00:11:04.935 A freight broker's job is to match loads from shippers with available carriers, 00:11:05.015 --> 00:11:06.455 price the load correctly, 00:11:06.455 --> 00:11:22.890 and then close the deal fast. AI can automate most of the mechanical parts of that, monitoring incoming loads, matching carriers, predicting price, drafting outreach. You're not replacing the broker. Charge a 100 to $300 per broker per month. A brokerage with 50 brokers is a $5,000 00:11:22.890 --> 00:11:30.465 to $15,000 per month customer. There are thousands of freight brokerages in The US. Number three, Amazon brand advertising 00:11:30.465 --> 00:11:31.505 autopilot. 00:11:31.505 --> 00:12:08.995 There are tens of thousands of brands doing a million dollars to $50,000,000 on Amazon, and almost none of them have sophisticated advertising operations. They're running campaigns manually, guessing at bids, and leaving money on the table every single day. Build opinionated software specifically for Amazon operators. Not a generic marketing tool, but something that automatically adjusts bids based on margin data, identifies wasted spend, and optimizes pricing against competitors. Charge $500 to $3,000 per month. The ROI is obvious because you can show them exactly what they're wasting. Notice what all three of these have in common. They're not trying to replace humans. 00:12:09.155 --> 00:12:15.870 They're trying to make one person as productive as a team, and they're built for one very specific customer, 00:12:15.950 --> 00:12:21.310 not everybody. If you wanna understand exactly how I would make my first million dollars, click here.