RightTech

We put our money where our mouth is.

Damien and I went through our own tech stack and asked a simple question...

What are we actually using?

At the time, we were paying about $2,000 a month in software licenses.

Probably more now if we had kept them all.

This is not an attack on SaaS tools.

A SaaS tool is just software you pay to use each month.

A lot of them are excellent.

The problem was simpler than that.

We were often paying $97 a month... $197 a month... sometimes more...

And only using 10% of what the tool could do.

So we started replacing small parts of those tools with smaller internal apps built with AI.

Not giant rebuilds.

Not trying to outdo companies that had spent millions building polished platforms.

Just taking the one workflow we actually needed...

And building that.

That shift taught me something useful.

Expensive software is not always expensive because your business is complex.

Sometimes it is expensive because you bought a whole department store when you only needed one shelf.

That is why I keep saying software and the cost of software should not be the limiting factor to growing your business.

We replaced tools for content, scheduling, and outreach.

In many cases, the running cost dropped to cents.

A post might cost 15 cents.

An image around 3 cents.

A video maybe a dollar.

Literally anyone can now build software.

The bottleneck is no longer skill.

It really is imagination.

Have you looked at your software stack lately and asked what you actually use?

The quality of an AI build is usually decided before the build starts.

Not in the code.

In the brief.

I’ve learned that if I spend 10 minutes turning a messy idea into a simple MVP brief, I get a better first version and far fewer fixes later.

MVP just means the simplest usable version.

This matters now because literally anyone can build software.

We’re entering an age where it’s getting even easier for people.

The bottleneck is no longer skill.

It really is imagination at this point.

Here’s the process I use before I let Claude Code or Codex touch anything.

1. Start with the brain dump

I write the idea exactly as it comes out of my head.

Usually messy.

Something like... “I want a habit tracker that reminds people, shows streaks, and feels simple.”

2. Force the use case

I ask, who is this for and what do they need to do first?

Not every possible feature.

Just the core job.

3. Ask follow-up questions

What should happen when they log in?

What should they see first?

What needs to be saved?

What can wait?

This is the part that saves me.

4. Turn it into a build brief

I write a short spec with the pages, actions, and basic rules.

Nothing fancy.

Just enough that the AI has clear instructions.

5. Then build

I really want people to see how easy this is now.

You don’t necessarily need the overhead of other tools.

If the brief is clear, AI can do a lot.

If it’s vague, you’ll get chaos dressed up as progress.

Do you brief your ideas first... or prompt on the fly?