If anyone can build software with AI, what's the real competitive advantage?

There are so many people building amazing AI tools in their basements. A great tool doesn't protect you anymore. An audience does.

"Build the audience first. And then you basically are building optionality for how you want to leverage it... especially because of AI, the real moat is audience."

What happened to the old moats?

They used to be skills. You build great websites, and that's really hard. You build software, and that's really hard. Neither is hard anymore.

For people who aren't tech geniuses or billion-dollar companies, the only reasonable moat left is an audience you can sell your own things to, or other people's things. When 100 times as many websites and 100 times as many software products exist, the most needed thing is attention. You can build the greatest software on earth, and if nobody knows it exists, you have a zero.

Why does audience matter more now?

Because it's so easy to build, more people are building, and it's crowded trying to get in front of buyers. I think competition for eyeballs goes up at least 100 times in the next couple of years, because there will be 100 times more software and 100 times more companies. Layoffs add to it: people who lose jobs start side businesses.

Follow that to the second-order effect. If ten times as many businesses are trying to reach people, most of them will do it by buying ads, so ad costs probably go up. Ads and SEO used to be enough, which is why plenty of businesses never bothered building an audience. Those channels won't disappear, but I think they get 10 times harder.

That makes an owned audience, like a newsletter list, worth more than it is today, because audiences get harder to build when everybody's trying. The one caveat I always add: that assumes email itself doesn't get screwed.

Big companies know what that's worth. A newsletter that started as a side hustle was bought by HubSpot, a multi-billion-dollar company. Dylan Redekop told me on my podcast that HubSpot reportedly paid upwards of $10 million for Starter Story and about $10 million for My First Million.

Should I build the product or the audience first?

The audience, when they compete for the same week, unless the product work is unblocking customers who already pay you. A plan that builds an audience before there's a revenue model is fine with me. The return is optionality. And I discount any strategy whose only advantage is that the tool is better.

An audience also turns AI progress into good news. If things get better and easier and you have an audience, that only helps you. Regular business owners still won't figure this stuff out in the near future, even when it gets easy. They'll pay someone to do it for them.

When I asked Claude to pressure-test my own strategy (what's my moat, how am I going to win), every single answer came back to the media brand. We could probably succeed with cold outreach alone. The media brand is what brings people to us.

What if someone copies my AI tool?

Assume they will. Somebody sees a good idea and builds a copycat. That's a risk for any business, and especially for SaaS and AI. I have three answers to it:

1. Own your media. Content about your brand and what you do, so people already know you.
2. Be ready to niche down. If a big player builds something similar and powerful, pivot to a narrower version: the local media tool, or the newsletter tool for HOAs. Each of those is a seven-figure opportunity.
3. Keep a service on top. The average customer is never going to know which software got built. A done-for-you agency still works even if you have to run it on someone else's tool.

Can the product itself be a moat?

Only if it's more than a basic feature. If greetform just made a prettier form with an offer wall at the end, anyone smart enough to build things wouldn't need it. Most people aren't, so a simple version is still viable in the short term. The real defense is the complexity I've added: media kits that pull a publisher's data from their email platform and get sent to a business owner based on how a subscriber answered, data enrichment, and SMS capability (built, not live yet, because there are legal issues to sort out). Nobody's going to build that for themselves.

Breadth works the same way. If one product did 12 different things for a local media company, they're not going to cancel it to go build their own version of everything. That buyer is somebody launching a six-figure business or already running a multi-six-figure one, not an individual entrepreneur. Shane Brady made a related point on my podcast: a self-built micro tool doesn't come with distribution, and someone still has to go sign up advertisers across hundreds of newsletters.

What's the edge when AI can write the newsletter?

A lot of people now know they could do newsletters with AI. They don't know how to get them in front of people or how to make money from them. When I break down what makes my model unique, it's the Meta ad funnel. My guess is maybe 100 people really understand it.

The model as a whole is hard to copy too. I consult with $10 million businesses that can't figure it out. You can't just listen to my podcast and replicate it. It takes a lot of time and effort to truly grasp, which is exactly why it's a blue ocean. If it were obviously valuable, everyone would already be doing it.

How do I test whether my business is defensible?

Ask whether AI can just do what you're doing within one to two years. Run that test on every repeatable content or service task. If the answer is plausibly yes, the value has to live in the brand and the relationship instead. That's why my team invests in design, branding, social media and the podcast, not only the newsletter.

What does audience-first look like in practice?

I'd been doing newsletters for two years, knew the problems, built tools to solve them, and already had an audience of newsletter operators to sell to. The audience came first. The software came second.

The reverse is the common trap. You could build the coolest thing ever, perfect and ready to go. How are you going to sell it? How are you going to get in front of people?

Related: will the AI labs just build my product, why AI makes trust more valuable and building an audience before you sell.