Matt Dixon

Fractional CTO helping organizations reach their business goals through technology.

[Matt Dixon @FrontRangeSys]

Most owners hand AI the wrong job first. They start where a mistake hurts.

Zone two of four is automate. Start where a mistake costs you nothing.

That's how your team learns to trust it.

[Matt Dixon @FrontRangeSys]

Zone one of four: the no go zone.

The things AI will never touch in your business. Your list won't match mine, and that's the point.

Write it before anybody automates anything.

[Matt Dixon @FrontRangeSys]

Every owner has a list of things AI will never touch. No two lists match.

A guy told me AI will never send an email for him. My AI sent him one after the call.

We're both right. Zone one of four: the no go zone. Write your list first.

[Matt Dixon @FrontRangeSys]

One question exposes a fake AI builder.

Red flags: apps wired together from a tutorial, the same setup they sold the last guy, a link when you ask how it works.

The question: show me where it can break.

[Matt Dixon @FrontRangeSys]

One person onboards every new hire. A 100 line Google Sheet, about 10 emails, about 10 systems. He needs two weeks' notice and gets two or three days.

Start smaller than a new tool. Automate one email, then the next.

How many things in your company run through one person?

[Matt Dixon @FrontRangeSys]

Your biggest expense line is hiding something.

An owner told me a dollar total doesn't tell him a story. He wanted gallons per truck, per week.

A total hides a person. Split your biggest line by who spent it.

[Matt Dixon @FrontRangeSys]

Every feature is a new room on the house. Keep adding them and the foundation starts to shift.

AI makes the rooms cheap. Somebody still has to live in the house.

From Ep. 19 of The Operating System Show with Mishka Kofman.

[Matt Dixon @FrontRangeSys]

Three levels of AI in a business.

Beginner: everyone has their own subscription.
Amateur: every team has its own setup.
Pro: one second brain with every procedure in it, and people approve the critical stuff.

Which level is your company at?

[Matt Dixon @FrontRangeSys]

There's a guitar in my office I can't really play. My son learns a new song on it in an hour.

Same guitar. Different outcome.

You bought ChatGPT. What job does it do for your team every day?

[Matt Dixon @FrontRangeSys]

Our AI dispatcher got fired. On a sales call.

The owner tested the demo playing a driver who just wrecked. The AI told him to hang up and call 911. He asked if he should call our competitor.

It should have sent the truck and stayed on the line.

[Matt Dixon @FrontRangeSys]

A company spent $200k on a hire that made the problem worse.

Product wasn't shipping, so they got a principal engineer. Nine months later, still debating architecture.

The engineer was good. Nobody had written down what they were building.

[Matt Dixon @FrontRangeSys]

15 years running a landscaping company. Never a vacation longer than two days.

150 people, 90 trucks, all of it built around him.

So he sold it and built the opposite. GreenPal. 300,000 transactions. No offices, no trucks.

90 trucks needed him. 300,000 transactions didn't.

[Matt Dixon @FrontRangeSys]

111 changes on a racing sim app. Zero lines of code.

Track map, tire wear, temperature, PSI, and it calls who is about to pass me.

My job was deciding what to build and checking the work.

Quoted a year for the tool your team needs? Get a second opinion.

[Matt Dixon @FrontRangeSys]

Every business will need an AI operating system within two years.

Quoting in a spreadsheet. Scheduling in an inbox. Pricing in one guy's head. Nothing talks to anything, so a person is the integration.

Count how many times today someone moved data between systems by hand.

[Matt Dixon @FrontRangeSys]

Anthropic shipped an AI operating system for one person.

Now picture 200 people at one company, each running their own. Same question, two answers, both acted on.

Your people just got faster at working alone.

[Matt Dixon @FrontRangeSys]

A public works contractor had someone reading bid sites for hours.

Now every night a bid finder scores the projects, pulls the plans, and starts the takeoff. In the morning there's a short list.

What's the finding work in your business?

[Matt Dixon @FrontRangeSys]

In-N-Out has a line around the building for three burgers.

I'm working with an owner who does something different for every client.

If what you sell can't fit on a menu board, your team can't make it the same way twice.

[Matt Dixon @FrontRangeSys]

In-N-Out has a line around the building for three burgers.

I'm working with an owner who does something different for every client.

If what you sell can't fit on a menu board, your team can't make it the same way twice.

[Matt Dixon @FrontRangeSys]

I'm about to lose a deal to a marketing department.

The dashboard is the easy part. The hard part is who can see whose commission.

Get that wrong once and everybody knows what everybody else makes.

[Matt Dixon @FrontRangeSys]

I told my AI to draft an email. A few seconds later it was sent.

I never wrote down what draft means. To the agent, a draft was unfinished work.

Everyone you hand work to has their own definitions. AI just acts on them faster.