I removed 95% of my product's features. Here's what happened.

When we started building our AI visibility product, we thought more features would make it more valuable.

The product could track a brand's citations across popular AI models, with multiple models from each AI provider. We also had website analysis, brand sentiment analysis, competitor insights, action plans, and more.

We even built it as a self-hosted product. Customers could deploy it on Railway, use their own AI API keys, own their data, and have control over their monthly AI costs.

Technically, it was a pretty powerful product.

But there was a problem.

The people we wanted to sell to didn't want all of this.

Most of our customers ended up being from technical backgrounds. We had initially expected agencies, brand owners, and solo founders to love the idea of owning the product and controlling their costs.

Instead, the product felt too complicated and overwhelming.

After talking to more potential customers, we started hearing a much simpler need:

"I just want to know where brands in my category are getting cited by AI models, especially ChatGPT."

They didn't need another dashboard.

They didn't need dozens of features.

They needed to know where AI models were getting their information from.

We also considered a "done-for-you" approach, but there was another problem. Automating things like publishing content across industry websites, blogs, and review platforms can easily turn into spam and potentially hurt a brand's reputation.

After spending years studying AI citations, my conclusion was that one of the most effective approaches is to identify the sources that AI models already use in your category and then build your brand's presence on those sources manually.

So we made a pretty radical decision:

We removed 95% of the product.

Instead of a complicated dashboard, we now give brands the sources of AI citations in their category through a simple Telegram experience.

No sign-up.

No subscription.

Just send /start, tell us your brand name and category, and get the sources.

We're still early in this new direction, but the product feels much closer to what customers actually wanted.

It was a good reminder that building more isn't always progress. Sometimes the best product decision is deleting what you built.

If you want to see what we ended up building:
mayin.app