Before You Build Another AI Product, Follow the Traffic

Whenever I research a new product idea, I tend to start the same way.

I search for competitors.

Then I open ten or twenty tabs, compare landing pages, look at pricing, read feature lists, check reviews, and maybe build a spreadsheet.

After an hour, I usually know a lot about what people have built.

But I still don't know the thing I actually care about:

Which of these products are getting users?

That question changed the way I think about market research.

Competitor research usually focuses on the product.

What features do they have?

How much do they charge?

Who are they targeting?

All useful questions.

But if I'm deciding whether to enter a market, I also want to know how the market is rewarding those products.

Who has traffic?

Who is growing?

Where are those users coming from?

And perhaps most importantly:

What can their distribution teach me before I build anything?

A Competitor List Isn't Market Research

Finding competitors is easy.

Understanding them is harder.

Suppose you're thinking about building an AI video product.

You find thirty companies.

Almost all of them mention some combination of:

• AI generation
• automated editing
• avatars
• subtitles
• templates
• voice generation
• social exports

If you only compare features, the market starts to look incredibly crowded.

But now imagine adding traffic.

One company has 10 million visits a month.

Another has 700,000.

A third has 150,000.

That gives you more context, but it's still incomplete.

What if the third company had only 40,000 visits three months ago?

Suddenly, the smallest competitor may be the one worth studying.

This is one reason I increasingly look at fastest-growing AI tools, not just the largest ones.

Size tells you who is already big.

Growth can tell you where something interesting is happening.

Don't Only Study the Winners. Study the Movers.

The biggest company in a category is often the least useful competitor to copy.

Large products have advantages you probably don't have:

brand recognition, large teams, existing audiences, partnerships, backlinks, capital, and years of accumulated distribution.

If you're starting from zero, studying the number-one player can sometimes teach you the wrong lesson.

A much smaller product that has doubled its traffic in three months may be more relevant.

Something changed.

Maybe they launched a free tool.

Maybe a feature went viral.

Maybe they found a valuable SEO niche.

Maybe creators started talking about them.

Maybe they entered a market competitors ignored.

That is where market research becomes interesting.

Instead of asking:

What does this company have?

Ask:

What happened here?

A sudden change in traffic is often the beginning of a research trail.

Where the Traffic Comes From Matters Even More

Two products can each receive one million visits per month and have completely different businesses.

Imagine this:

Product A

60% search
20% direct
5% social

Product B

10% search
15% direct
50% social

Same traffic.

Very different distribution.

Product A tells me that search may be a major acquisition channel in this category.

That immediately creates new questions.

What are people searching for?

Are they finding product pages, free tools, templates, comparison pages, or educational content?

Product B tells a different story.

Maybe the product is inherently shareable.

Maybe short demos perform well on TikTok or X.

Maybe creators are driving adoption.

Maybe the founders have built a strong audience.

The important point is that traffic source is not just analytics.

It's evidence about how a market works.

Distribution Can Be Reverse Engineered

When I find a competitor with interesting traffic, I don't want to stop at:

“They're doing well.”

I want to understand why.

If search is unusually strong, I look at the search opportunity.

If social dominates, I look at what people are sharing.

And at what people say underneath it. The comments on a competitor's TikTok demos, YouTube walkthroughs and Reddit threads are where users describe what they hoped the product would do and where it let them down. ZocialComment↗ (opens in a new tab) exports those comments from Instagram, TikTok, YouTube, Facebook and Reddit to CSV, Excel or JSON, so you can read a few thousand of them in a spreadsheet instead of scrolling.

If referral traffic stands out, I start wondering which websites, directories, affiliates, partners, or integrations are sending users.

This is why I find looking at AI tools with strong search traffic useful when researching a category.

Search traffic in particular can reveal more than a marketing strategy.

It can reveal demand.

If multiple companies in the same category are attracting meaningful organic traffic, users are probably searching for that problem.

And the queries they use can tell you how they describe it.

I think founders sometimes separate product research and distribution research too much.

In reality, they overlap.

How people discover existing products can tell you what they want from the next one.

SEO Data Is Also Product Data

We usually treat SEO as something that happens after the product is built.

Build first.

Then find keywords.

Then publish content.

But competitor search traffic can be useful before you write a line of code.

Suppose you're researching AI presentation tools.

You might discover people aren't only searching for “AI presentation maker.”

They may be looking for:

“AI pitch deck generator”

“AI PowerPoint from PDF”

“turn document into slides”

“AI presentation for students”

Each query represents a slightly different job to be done.

That is product information.

A competitor's SEO strategy isn't just telling you how they acquire customers.

It can show you how users describe their pain.

Competitor SEO isn't just a marketing strategy. It's a map of what the market is asking for.

That can influence positioning, features, landing pages, and sometimes the entire product idea.

Traffic Doesn't Tell You Everything

There is an obvious limitation here.

Traffic isn't revenue.

A company can have millions of free users and a weak business.

A smaller B2B product can generate far more revenue from far fewer visitors.

Traffic also doesn't tell you whether customers are happy, whether retention is strong, or whether the company is profitable.

So I don't think traffic should replace traditional research.

I think it should sit next to it.

Look at the product.

Look at pricing.

Read customer conversations.

Understand the category.

But also look at where users appear to be going.

Market research becomes more useful when several signals point in the same direction.

Once the market signals are strong enough to justify building, founders still need to turn the validated opportunity into a reliable product. Working with experienced custom software development services↗ (opens in a new tab) can help translate requirements into architecture, integrations, development, testing, and a scalable first release.

This Is Why I Started Building AITrustList

I kept wanting to answer the same questions whenever I explored an AI category.

Who has the largest audience?

Who is growing fastest?

Who is winning through search?

Who is strong on social?

Which smaller products are suddenly moving?

Where are their users coming from?

I wanted a faster way to see those signals together.

That eventually became AITrustList.

The goal isn't to tell founders which market to enter or which competitor to copy.

It's to make the first stage of research easier.

A ranking or traffic chart doesn't give you the answer.

It gives you somewhere better to start asking questions.

And sometimes one unusual number is enough to send you down a useful rabbit hole.

Follow the Traffic, Then Ask Why

If I were researching a new AI product today, I wouldn't begin by making the longest possible competitor spreadsheet.

I'd find the category.

Identify the established players.

Then look for movement.

Which companies are growing?

Which acquisition channels seem to work?

Which products are unusually strong in search?

Which ones have found social distribution?

Then I'd ask why.

The landing page tells you what a competitor wants you to believe about its business.

Traffic can give you clues about what the market is actually doing.

Neither view is enough on its own.

But together, they're much more useful.

So before building another feature, copying another competitor, or entering another AI category, I'd spend some time following the traffic.

You might find that the most interesting part of the market isn't where everyone already is.

It's where users are starting to go.

Explore AI tool traffic rankings on AITrustList