Why Most People Will Never Use AI (And Why That's Good for Business)

The Replit Agent 3 released last week. Another step that blows my mind. Another tool that makes everything I do feel like it's already obsolete, or soon to be.

For about 30 seconds, I felt that familiar feeling. Shit. Is this it? Am I done? Should I try something else?

Then I thought about the world outside of my computer screen. I remembered: My wife still uses ChatGPT in a browser window. No account. Just types health questions and closes the tab (and she NEVER forgets to close the tab).

Here's what most people (me included) who are chronically online in nerdy arenas forget about everyone else.

Most humans aren't scrolling Twitter each morning hunting for the latest AI agent release. They're not signing up for beta access or watching 28-minute YouTube tutorials on how to automate an inventory system for a local pet shop (something I'm trying to figure out now).

They have businesses to run, kids to feed, problems to solve, and no time to figure out how to automate or "optimize"

And those problems? They're not going anywhere. The business will be run the same way until someone shows them a better way so they can spend more time feeding those kids.

The Excel Wizard

The bridge between cutting-edge tools and actual results will always need builders.

I think about how Excel changed so much in the working/business world. It made calculations 100x faster than pen and paper.

I just looked it up and it came out 40 years ago. Only 12 years ago I remember getting to my internship as a 21 year old and I was looked at like some tech genius for knowing how to do a vlookup.

That internship was at one of the biggest companies in the world. Basic excel knowledge made me a 'spreadsheet wizard', 30 years after Excel had been released.

It was adopted because it had to be. It made companies 100x most efficient, but most still didn't have to use it.

Why's that? Because you still had to:

• Buy the software

• Transfer your data

• Learn the formulas

• Debug your spreadsheets

• Train your team

Most people paid someone else to figure it out, smart money. Then the employees just ran what they were told to run and only those who needed to get hired needed to learn (me).

Today's AI Feels Identical

That's what ai automation feels like to me in today's world.

Sure, there will be agentic solutions coming that obsolete some automation, but back then there was custom software built that obsoleted the use for excel in some capacity.

But companies cannot constantly change solutions. It takes time and money to learn or train your employees to learn. Even with fully agentic solutions that will build full applications something still needs to:

• Understand what it can do

• Learn the interface

• Iterate on outputs

• Deploy and debug

• Actually solve your specific problem

That's mental processing power. That's time. That's barriers that will never disappear.

Where The Real Money Lives

The opportunity isn't only in building the tools themselves it's in the transition.

If you are early there will be plenty of people to get on board. Why do I know they'll get on board? Because they're time and money is valuable to them and just like excel, this will make their business 100x more efficient.

So if you're thinking about this field just know your target market isn't the person who'll spend three weekends using Replit Agent 3. It's the business owner who wants their manual process automated but doesn't have someone internal to figure it out.

People just want point A to point B. Done for them without the learning curve.

I spent months diving deep into every new automation tool, getting lost in the possibilities. I do think this is important, to a degree, but ai overload is definitely a thing. I experience it.

Lately I've snapped out of it after talking to 4 business owners in the real world. It gave me clear opportunities to work on. The opportunities are there and I see them now, selling the system is the next, arguably more difficult, part.

My System For Staying Focused

So I am operating with this system going forward for staying current without losing focus:

1. Set up clients first. Solve their immediate problems with proven tools.

2. Then iterate. Use stable time to test new tools on existing setups.

3. Evaluate quickly. Does this new tool make client results 10x better? If not, not worth it yet.

4. Implement strategically. Only add tools that dramatically improve outcomes.

The gap between "possible" and "practical" is where a business fails.

Every new AI tool makes this gap wider, not narrower. There are more possibilities, more complexity, more decisions. I've realized my job isn't to use every tool. It's to be the person who knows which tool solves which problem and can implement it without the client thinking about it.

A business owner I spoke with last week spelled this out perfectly. He runs a landscaping company and a dog grooming shop, dealing with all the typical small business headaches. When we talked about what he'd actually pay for, he said:

"Your sweet spot could be, you take the place of a $45,000 year position for inventory management, that these people are never gonna hire."

He nailed it. Small businesses know they need better systems. They know there are people who could fix their problems. But they can't justify hiring a full-time inventory manager or procurement specialist.

They will, however, pay someone to come in and solve the specific problem without requiring them to become experts themselves. The market demand is there it's just packaged wrong for most small businesses.

Your Market Is Waiting

The world that's "not online" is massive.

They know ChatGPT exists. Maybe they've tried it once. They don't know about agents, automations, or APIs. They just know their current process is slow, expensive, tedious, or broken.

Take the same business owner's invoicing nightmare. He's sending 200+ invoices every two weeks using Jobber, a platform built specifically for landscaping companies. But listen to his frustration:

"A lot of them are repetitive, but there's just like a touch more...like Jimmy Smith doesn't want his lawn cut this week so you can't invoice Jimmy Smith for two cuts every single week."

The tool is 90% there. It handles most of the work. But that 10% of exceptions - the weeks when Jimmy Smith skips his lawn service - requires human judgment every single time.

So he hired someone in the Philippines just to handle invoicing. Not because Jobber is bad, but because even good tools can't eliminate that final layer of decision-making. The gap between "mostly automated" and "fully automated" is where businesses still pay humans to think.

Remember, their most valuable asset, not technical knowledge. With that being said, it has to be practical. Point A to point B is my focus going forward.

So, while everyone else is arguing about which AI agent will rule the world, there are thousands of businesses right now manually doing things that could be automated in an afternoon.

That's the market. That's the opportunity.

Time to build the bridge, stay focused, and let the tools serve the outcome, not the other way around.

2 automation opportunities from my conversations this week:

Thanks for reading,

James