Why Most "AI for Your Business" Advice is Bullshit

That "AI and automation" stuff you’re fascinated with is probably going to make things worse.

Everyone selling you on AI or automation is skipping the most important question: What CONSTRAINT does this actually remove?

Your biggest problem is managing a low-cost labor force?

A fancy new $500/month follow-up automation software isn't going to fix that.

This is the original sin of venture-backed roll-ups who says their thesis is to “add technology to local service businesses.”

I knew I hated this strategy but now I finally have a language to describe what I hate about it.

The constraint in these businesses isn't CRM or ordering supplies more efficiently or even answering the phone.

The constarint is managing a large, diverse, low-cost labor force providing weekly cleaning services in affluent people's homes. That's what makes the business hard—not sending text message follow-ups.

Unless you can specifically say how AI and automation solves your actual constraint, it's a completely misguided strategy.

Becoming a franchisor in this space solves this problem. You delegate the low-level labor issues to your franchisee. That makes sense.

“AI” does not.

Technology only helps if it removes a real limitation.

The Four Steps Everyone Skips

My man crush Dr. Eli Goldratt laid out a framework for evaluating any new technology.

Almost nobody follows it.

Step 1: Name the actual limitation this technology removes.

Not "better customer relationships" or whatever vague garbage is on the landing page. The real thing.

Example: A CRM removes the limitation of losing track of customer conversations, forgetting to follow up, and having all your sales knowledge locked inside one person's head.

If you can't state the limitation in one clear sentence, stop. Don't buy anything.

Step 2: Find the old rules you built to survive.

Before the limitation was removed, you created workarounds. Habits. Rules. Ways of coping with the fact that this part of the business was painful.

Before CRMs existed, you probably had rules like: only your A-players handle big accounts because they won't drop the ball. Weekly meetings where everyone reports their deals because otherwise it's chaos. Customers only talk to "their" rep because nobody else has context. New hires shadow veterans for six months because the knowledge lives in people's heads.

These rules made sense. They were rational responses to a real constraint.

Step 3: Define the new rules.

Now that the limitation is gone, what should actually change?

With a CRM that captures everything: anyone should be able to handle any account because the context is right there. Meetings should be about strategy, not status updates, because the system has the status. Follow-ups happen on triggers, not someone's memory. New hires are productive in weeks, not months.

Defining new rules means admitting the old rules are now stupid. Most people would rather keep doing familiar things that no longer make sense than do unfamiliar things that do.

Step 4: Actually change the rules.

You deploy the technology and change the rules at the same time.

What actually happens instead: You buy the CRM. You still assign dedicated reps to accounts. You still run those status meetings. You don't enforce data entry because "it takes too much time"….accommodating a limitation that no longer exists.

If you install new systems but keep operating the old way, you get maybe 10% of the benefit. Probably less.

The Swimming Pool

Before you learned to swim, you had a rule: never go deeper than waist-high. Good rule. Kept you from drowning.

You take swimming lessons. The limitation is gone—you won't sink.

But you keep the waist-deep rule.

Those lessons were worthless. You paid money to learn something you refuse to use.

The Bottom Line

Technology is a LEVERAGE play. You're looking for the one change that creates a massive performance jump.

Before you add AI, automation, or whatever shiny object someone on Twitter is waving at you: identify the specific constraint it removes, find every rule you created to work around that limitation, figure out what the new rules should be, and change the rules when you deploy the technology.

Trust me, this will help.

Yallah Habibi,

Jon

Join Sagan and start working on your constraints