Blockbuster didn't lose because it ignored Netflix. It lost while copying Netflix. It launched its own online rental service in 2004, and in late 2006 it rolled out >Total Access, which let mail subscribers return discs at their local store. Smart idea. But the business underneath was still built around stores and shelves. I think a lot of small businesses are doing the same thing right now by adding AI to existing software and calling it a strategy.
Every app on your desk just got an AI button
Open your CRM. There's a copilot. Your email offers to write replies. Your accounting software promises smart insights. Your scheduling tool has an assistant.
Some of that is genuinely useful, and you should use it. You're already paying for it. If the AI summary in your CRM saves your salesperson ten minutes after a call, great.
But look at what it's actually doing. It speeds up one step inside one app. The person still copies the lead from the web form into the CRM. The office manager still builds the Monday report by hand. Someone still has to remember to follow up. The process is the same process, with a slightly faster button in the middle.
Why AI buttons alone aren't an AI strategy
Two problems.
First, everyone gets the same button. Your competitor down the road uses the same CRM with the same copilot. If that's the whole plan, nobody gets an edge. You've just kept pace.
Second, the button is stuck inside the app. Most of the wasted time in a small business isn't inside any one tool. It's in the gaps between them: the handoffs, the re-typing, the "did anyone call this person back?" A feature living in one app can't see those gaps, so it can't close them.
I'm a process guy. Most businesses I meet have good people and a product that works. What slows them down is the process, and a new button doesn't change the process.
Don't throw out your software, either
Here's where I disagree with the louder version of this argument. Blockbuster's mistake wasn't keeping its stores. It was keeping the old way of working and attaching the new thing to the side.
The fix for a small business is not ripping out your CRM and starting over. Software is changing too fast to justify that, and the tools you have are usually fine. The fix is redesigning how the work flows, then using the tools you already own to run the new version. Custom only comes in where nothing you have fits, and only if it pays for itself.
A 5-step test for redesigning one process
Try this on a single process this week. Quote requests, new client onboarding or invoicing all work well.
- Write down every step. From the moment the work starts to the moment it's done. Include who touches it and which app it lives in.
- Label each step. Is it a decision (a person judging something), data moving (copying from one place to another), drafting (writing an email, a quote, a summary) or waiting (sitting in someone's inbox)?
- Ask the blank-page question. If you set this process up today, knowing AI can draft and software can move data on its own, which steps would still exist? Usually the decisions stay. Data moving and drafting shrink. Waiting gets replaced by automatic nudges.
- Check what you already own. Before buying anything, see whether your current tools can connect or already have the feature. Often the pieces are there and just aren't wired together.
- Price the gap. If something is truly missing, estimate the hours it would save each month. Build or buy only if those hours are worth more than the cost.
What that looks like (a hypothetical)
Picture a 12-person HVAC company. The bolt-on version: the office manager uses the CRM's AI to write quote emails a little faster. Everything else stays manual.
The redesigned version: a web form or call note lands, a workflow creates the contact in the existing CRM, pulls the job details into a draft quote, and puts it in front of the owner to approve. If the customer doesn't respond in three days, a follow-up goes out on its own. Same CRM. Same people. Very different process.
How I'd approach this
I'd start with the bottleneck, not the technology. One process, the one costing you the most time or money, mapped step by step with the people who actually do it. Then I'd connect the tools you already pay for, usually with n8n, and add AI (Claude, ChatGPT or Gemini, depending on the job) where it drafts or sorts something a person used to do by hand. A human stays on every decision that matters.
If your data is scattered, I'll get it into one place you control, often with Supabase, so the AI is working from your real customers and jobs. You own whatever gets built. You can read more about my process for the first 30 days, see the kinds of work I automate, or check my FAQ. And if the software bill itself is the bigger worry, start with the subscription graveyard funeral math.
Frequently asked questions
Are the AI features in my existing software worth using?
Usually, yes. If your CRM can summarize a call or your email can draft a reply, use it. You're already paying for it. Just don't confuse it with a strategy. Built-in AI speeds up a step inside one app. The bigger savings come from changing how work moves between apps and people.
Do I need new software to use AI well?
Rarely. I start with the tools a business already pays for, connect them, and add AI where it saves real time. Software is changing too fast to justify ripping everything out. A custom build only makes sense where nothing you own fits the redesigned process, and when I build one, you own it.
What does redesigning a process around AI mean in practice?
It means mapping how a job actually flows today, then asking which steps would exist if you set it up from scratch with AI available. Usually the steps that disappear are data moving between apps, drafting from a blank page, and someone remembering to follow up. A person still makes the decisions.
How do I know which process to redesign first?
Pick the one that costs the most time or loses the most money when it slips. For many small businesses that's lead follow-up, quoting, invoicing or weekly reporting. Fix that one completely, measure the difference, then move to the next. One finished fix beats five half-started ones.
Don't wait for the next button
Your software vendors will keep shipping AI features, and your competitors will get every one of them too. The edge is in how your work flows. If you want to figure out where that starts in your business, let's talk for 30 minutes about the one process slowing your team down most. No pitch. I'm in Overland, about 20 minutes from downtown St. Louis. Book a time here.
