DEV Community

Cuteprob
Cuteprob

Posted on

What I Learned Building an AI Baby Dance Website Around One Core Query

A few months ago, I started building a small AI product around a very specific search intent: turning a baby photo into a dancing video.

The result is aibabydance.com.

At first, I thought the product itself would be the hard part.

It wasn’t.

The hard part was figuring out how to structure the site so that:

  • Google could understand what each page is for
  • users could get to the right workflow quickly
  • and the product didn’t collapse into one overloaded homepage

This post is a quick breakdown of what I learned while building the product and how I now think about homepage vs workflow pages for AI tools.


The original idea was simple

Take a photo, choose a dance template, generate a short dancing clip.

That’s the core homepage experience:

👉 AI Baby Dance homepage

At first, I tried to put everything on one page:

  • baby dance generation
  • more general AI dance video generation
  • image fixing / prep
  • model-specific workflows
  • templates
  • examples
  • pricing
  • FAQ

It looked comprehensive, but from both an SEO and conversion perspective, it was messy.

The homepage was trying to do too much.


What changed my thinking

I realized there are at least three different user intents hiding behind what looks like “the same product”.

1. The core intent: “I want an AI baby dance video”

This belongs on the homepage.

That page should be the clearest possible answer to the main query:

  • upload a baby photo
  • choose a dance template
  • generate a dancing video

2. The broader intent: “I want to make any photo dance”

This is not exactly the same as baby dance.

It deserves its own page:
👉 AI Dance Video Generator

This page is better for broader use cases:

  • adults
  • creators
  • full-body portraits
  • more general dance-style output

3. The input problem: “My photo isn’t ready for animation”

This turned out to be a separate workflow entirely.

Some users don’t actually have a good source image:

  • the subject is cropped
  • the pose is wrong
  • the framing is messy
  • the full body is missing

That’s why I split out:
👉 Dance Image Editor

This page is not just “another feature page”.
It solves a different problem:
getting the photo ready before generation.

4. The advanced / model-specific path

Some users want more control or a more specific workflow.

That’s where pages like this come in:
👉 Kling AI Baby Dance


Why I stopped treating everything as “one landing page”

A lot of indie makers make this mistake (I did too):

We think one product = one page.

But search intent doesn’t work like that.

Google is not asking:

  • “Do you have one product?”

It is asking:

  • “Does this URL clearly solve this exact query intent?”

That means different intents deserve different pages, even if the backend workflow is related.

For me, the structure now looks more like this:

  • Homepage = main baby dance query
  • AI Dance Video Generator = broader dance-video query
  • Dance Image Editor = input-fixing / prep workflow
  • Kling AI Baby Dance = advanced / model-specific workflow

This made the site easier to understand for both users and search engines.


One surprising lesson: workflow pages are not just SEO pages

At first, I thought workflow pages were just there to catch long-tail traffic.

But they ended up doing more than that.

They also:

  • reduced confusion on the homepage
  • made the product feel more structured
  • created clearer upgrade paths
  • made premium features easier to explain

For example:

  • if your image is not ready → go to the image editor
  • if you want a broader dance generator → go to the AI dance page
  • if you want a more advanced path → go to the Kling page

That’s not just SEO.

That’s product architecture.


The real challenge wasn’t traffic. It was conversion.

One thing I learned quickly:

Getting search traffic is only the first step.

The harder problem is:

  • what users can try
  • what belongs in paid plans
  • how to explain premium workflows
  • and how to avoid turning the product into a “free curiosity click” with no revenue

That’s part of why I started separating:

  • free exploration
  • official templates
  • premium editing workflows
  • creator-oriented upgrades

This is still something I’m actively iterating on.


If you’re building an AI tool, here’s my takeaway

Don’t just ask:

“What features do I have?”

Ask:

“What distinct workflow intentions exist here?”

Then build pages around those.

A good AI product site usually needs:

  • a clear primary page
  • a few workflow pages
  • better internal relationships between them
  • and copy that matches user intent, not just feature lists

That shift helped me think much more clearly about my own site.


Current product pages

If you want to see how I structured it, here are the main pages:


Final thought

I used to think the job was to build “the tool”.

Now I think the job is to build:

  • the tool
  • the search entry points
  • the workflow structure
  • and the upgrade logic

All at the same time.

That’s the real product.

If you’re also building AI tools around search, I’d love to hear how you think about homepage vs workflow pages.

Top comments (0)