The Fourth Option: What AI Can Actually Do for Your Business

Most people I talk to think about AI the same way. They use it to do things. Write an email. Summarize an article. Answer a question. Clean up some text. Useful stuff, and there's nothing wrong with it. But that's the small portion of what AI can do. And because it's the only version most people have seen, they've quietly capped how useful they think it can be.

Here's the broader idea: AI can build. Not just do a task, but build the tool that does the task for you, forever. Take the data you already have and shape it into custom software built around exactly how you work. No migrating into someone else's platform. No monthly subscription. A thing built for you, out of your own stuff, that you own outright.

That shift — from "AI does tasks" to "AI builds tools" — is the whole thing. Once you see it, you can't unsee it. So let me show you what it actually looks like.

The problem with your other options

Think about how a small business usually solves a process problem. Say your system is a mess and you know it. Your options have historically been:

• Keep plugging away with what you've got.

• Find some off-the-shelf software and migrate everything into it.

• Pay a developer thousands of dollars for something custom.

Option one means living with the mess.

Option two means renting a platform that's loaded with features you don't need, missing a couple you do, charging you every month forever, and demanding hours of data migration to get started.

Option three has always been out of reach for most small businesses.

Building your own custom tool with AI is the fourth option. And for a lot of problems, it's the best one.

I built the first ones for myself

Before I built any of this for clients, I built it for me, because I had the exact problems I now help other people solve.

I do a lot of hourly consulting and freelance work, which means tracking time and building invoices. My old system was the usual duct-tape stack: a spreadsheet, a doc, a calculator, and a lot of copying numbers around. It worked, technically. But it took a couple of tedious hours a week, and because it was tedious, I fell behind constantly. Which is a great way to lose billable hours and work for free without meaning to.

So I built a time-and-invoice dashboard. I hit a button to start a timer, tell it what I'm working on, and it tracks my hours against my rate automatically. When it's time to bill, it does the math, I add a payment link, and it builds the invoice as a PDF. Two tedious hours a week became a few minutes, and I stopped falling behind, because there was nothing left to fall behind on. (You can

actually try this one yourself — it's live on my tools page.)

Before I built any of this for clients, I built it for me, because I had the exact problems I now help other people solve.

Then I built a second one for the money side — an income dashboard that tracks what's coming in and helps me route the right portions off to separate accounts for taxes, savings, and reinvestment. Same result. A stale, stressful, always-behind picture became a clear one that updates as I go.

These are the tools I run my business on. Every day. Built from my own data, owned entirely by me, costing me nothing every month. That's the reason I can build them for clients with a straight face.

The spreadsheet that ate two years

My wife does medical sales from home. Over about two years she'd built a system that worked — technically. A spreadsheet with over a thousand rows and close to fifty columns, plus broken formulas and repeated data and all the junk that piles up when a thing grows one patch at a time. On top of that, a Chrome window with a hundred-plus tabs to track everything the spreadsheet couldn't.

A spreadsheet with over a thousand rows and close to fifty columns, plus broken formulas and repeated data and all the junk that piles up when a thing grows one patch at a time. On top of that, a Chrome window with a hundred-plus tabs to track everything the spreadsheet couldn't.

None of that happened because she did anything wrong. Every decision made sense in the moment. That's just what happens when a system grows faster than you can redesign it.

Here's the interesting part. We started out trying to get AI to do her work — take over pieces of the process. And we kept hitting walls: proprietary information she couldn't put into an AI, token limits, hallucinations gumming things up. Then I asked what she actually needed to be able to do, and she said: track orders, know where they are in the process, and get reminders to reach out about reorders.

That's when it clicked. We didn't need AI to do the work. We needed it to build a tool.

So we did. A dashboard that shows where each client sits in the ordering process, flags issues before they become problems, and reminds her when it's time to reach out. Plus a product menu that builds an order in a few clicks instead of hunting around a spreadsheet with a calculator, wired to the dashboard so the information just flows. A thousand cells and a hundred tabs became one clean screen with exactly what she needs and nothing she doesn't. (You can see the before-and-after of this exact build on my tools page.)

When two people can't see the same picture

Another version of this: a realtor client who works alongside her husband, also a realtor and a contractor. Between them, there are a lot of moving parts. They worked together on most of it, but they had no system that let them actually work together. He'd know where one thing stood, she'd know where another did, her notes were scattered, and a lot of "wait, where are we on this?" was happening in person or not at all.

Two capable people, quietly losing time in the gap between them.

We mapped out their pipelines and built one dashboard they can both see — check progress, leave notes, sign off on steps, track time-sensitive to-dos. One source of truth. Plus a stripped-down mobile version of the client notes, so either of them can skim it in the car before a meeting. The fix wasn't working harder or communicating more. It was building the one place the work could live for both of them.

Renting versus owning

A lot of these come down to a single distinction: renting a solution versus owning one.

A local restaurant I work with wanted easier ordering but was clear they did not want the big ordering apps — no third party taking a cut of every order forever. So I built them a custom interactive menu that runs off a Google spreadsheet they control. Change a price in the sheet, and it updates on the customer side instantly. I also built a simple backend so any staff member can edit it without touching the raw spreadsheet and breaking something. They own it outright

and never pay rent on it.

A lot of these come down to a single distinction: renting a solution versus owning one.

A private chef I've worked with had the opposite problem — a menu that wasn't working hard enough. His was a static PDF, a wall of text that overwhelmed potential clients into closing the tab. For a caterer, the menu is the start of a sales conversation, and a PDF can't hold a conversation.

So we built an interactive menu where clients filter, browse, and build their own custom menu tied to their email. Now the chef walks into meetings already knowing what they want, and every menu built hands over an email address on the way. A dead document became a tool that sells and captures leads while it sits there. (Both of these are live to click through on my tools page.)

It's not always about saving time

One more, because it's the one that reminds me why this work matters.

A local substance abuse recovery organization had a genuinely great resource list — a big categorized spreadsheet pointing people toward help with jobs, clothes, childcare, detox, and more. The information was gold. But it was trapped in a spreadsheet, and the people who needed it most were pinching and scrolling on a phone in a hard moment. Every bit of friction there is a chance someone gives up.

So I turned it into an app. Tap a category, find an organization, tap to call or get directions. Same information, none of the friction. Plus a backend so the staff can keep it current without a developer. Sometimes the tool isn't about saving a business a few hours. It's about removing the exact bit of friction between a person and the help they need. (You can actually try this one yourself — it's live on my tools page.)

What all of these have in common

None of these are magic. None of them required a developer's salary or a platform migration or a monthly subscription. Every one started with data or a process that already existed, and turned it into a custom tool the owner controls.

That's the fourth option. Not AI doing your tasks. AI building the tool that does them — shaped around how you actually work, owned by you, out of stuff you already have.

If you've got a spreadsheet quietly eating your time, a process held together with duct tape, or something on your site that should be working harder than it is, there's a good chance it's more fixable than it feels.

If any of this sounds familiar, book a call. Let's talk about what your fourth option looks like.