Why I Stopped Trying to Build a SaaS

I wanted to build a SaaS company.

For months, I was making plans to transition Ken AI from an agency to a software business. The whole thing was mapped out - product roadmap, pricing tiers, the works.

Then something clicked. SaaS and agencies are converging onto the same place in the AI revolution. And I was about to make a move that didn't account for this.

Here's what I mean.

The traditional wisdom is clear: agencies trade time for money and can't scale beyond headcount. SaaS products scale infinitely with high gross margins and command premium valuations. If you want to build something big, build software.

I bought into this completely.

But I kept looking at the math and realizing I could get everything I wanted from a SaaS - high profit margins, high skill leverage, high enterprise value - with an agency too. If I built it smartly. And with way lower risk.

SaaS right now is actually a high-risk business. Tons of competition. And there's a new problem: companies can just build your SaaS themselves with AI coding tools. That threat didn't exist two years ago.

Agencies are harder to run, sure. But they can scale. My bet is that I can scale an agency as easily as I can scale a SaaS by building internal products that multiply what our humans can do.

Here's the convergence I keep seeing:

Software is becoming more personalized. Not "Hey, {First Name}" personalized. I mean custom UI blocks for each user based on their data. Dynamic APIs that write themselves based on customer needs. Support that uses AI interviews to personalize the product automatically. Bugs that solve themselves from user feedback.

The path forward for SaaS is deep personalization. Building custom UI blocks similar to Notion based on user data. Adding inputs to all your data forms. Building dynamic APIs that craft endpoints specifically made for each user.

If you're a SaaS company, your competition isn't just other SaaS tools anymore. It's the ability for companies to build it in-house in a way that's customized to them. You need to win that battle too.

Agencies are becoming more scalable. They're increasingly able to automate and build internal products thanks to AI. Agencies usually don't have good tech talent, but if you get at least semi-technical people and build internal products with AI tools, you can scale like software.

We have 4 developers on our team of 16. That's unusual for an agency. We're solving operational problems with development, which is an unusual way to solve operational problems. But it works.

Give it six months and some software companies and agencies will be indistinguishable from each other.

If you're a SaaS, you build better support and scale it using AI. You create personalized experiences using AI. You get closer to an agency's level of service quality.

If you're an agency, you build internal products and scale your humans using AI. You get closer to SaaS margins and leverage.

At the end of the day, it's the same destination from different starting points.

And here's the thing about personalization - if you do it right, you get the IKEA effect. When customers feel like they built something themselves, or like the product was built specifically for them, they don't churn. Retention goes up naturally.

Code is cheap now. Coding agents are powerful enough for most customization you might need. The technical barriers that used to separate "software company" from "service company" are collapsing.

If you're running an agency, start thinking about what internal tools could multiply your team's output.

If you're running a SaaS, start thinking about what personalization could make your product feel custom-built for each user.

Within 3-6 months of implementing this stuff, you'll start seeing the impact.

I switched back from planning a SaaS to building what I'm now calling a "world-class agency" - one with the margins, leverage, and defensibility that used to belong only to software companies.