How AI Finds AI: The Traffic Loop We Didn’t Plan For

The site is called SmallBizAI.au. It’s about AI tools for small businesses. So it makes a certain kind of sense that AI would find it.

But that’s not how I expected it to work.

In mid-March 2026, Bing AI was citing SmallBizAI.au around 16 times a day. By April 28, that number was 485. In a single day. That’s not a gradual climb. That’s a step change.

I didn’t build for this. It happened because of what we published.

What’s actually happening

Bing Webmaster Tools tracks how often Copilot references your site when answering queries. It’s separate from regular search clicks. It shows when the AI model decided your content was worth grounding its response in.

The queries driving those citations aren’t brand searches. They’re evaluation queries: “Rippling vs BambooHR Australia,” “best HR software for small business,” “how much does ChatGPT cost in Australia.” Someone asks Copilot a question. Copilot reaches for SmallBizAI.au as a source.

By April, that was 4,582 citations in the month. 140 different pages cited. The top page, a Flare HR company profile, had 563 citations in six weeks.

The loop

Here’s where it gets strange.

SmallBizAI.au is a site built using AI, about AI tools, read by people trying to use AI. And it’s being found primarily by AI systems.

The traffic from Bing AI citations isn’t large in absolute terms. But it’s additive to Google organic and works on completely different signals. It doesn’t care about backlinks. It doesn’t care about domain authority. It cares whether your page answers a specific question. Specifically, whether the page is structured around a decision: which tool, which service, which approach.

That matches what we publish. Comparison posts. Company profiles. How-to guides for specific Australian scenarios.

The AI citation channel rewards the same content that Bing organic does, and gets there faster. New posts start appearing in Bing AI citations within days of going live. Some comparison posts were cited 200+ times before they ranked anywhere in regular search.

Why it works

Copilot isn’t looking for the most authoritative site. It’s looking for the most useful answer to a specific question.

Eight of the top ten cited pages on SmallBizAI.au are comparison posts or detailed company profiles. Not overviews. Not “what is AI” explainers. Pages that exist to answer one question with enough detail to actually help.

The queries confirm it. “Calendar scheduling automation CRM vs standalone apps Australian SMBs.” “Workday Bamboo Australia SMB evaluation.” “Enterprise AI assistants SharePoint Australian data residency requirements.”

These aren’t people browsing. They’re people trying to make a decision. Copilot is surfacing SmallBizAI.au to help them make it.

What nobody planned

The site grew fast: 707 posts in about eight weeks. Most of it is structured the same way. Here’s the tool, here’s who it’s for, here’s how it compares. That structure wasn’t designed to game AI search. It was designed to be useful.

The traffic loop is a side effect of that.

AI tools are primary research tools for a lot of people now. Before they buy software, they ask Copilot. Before they hire a contractor, they ask ChatGPT. If your content is specific, comparative, and grounded in a real local context, you’ll get found.

We’re at 5,536 Bing AI citations as of early May. Up from zero in March.

The site is about AI. The traffic comes from AI. The loop closes on itself.

This is part of the Behind the Build series — posts about how SmallBizAI.au works, what we’ve learned, and where it’s heading.

📊 Compare AI tools side by side | 💼 Free resources & AI prompt packs