Agent Place

What it is

Agent Place builds AI agents for companies. I own its go-to-market.

What I did

The keywords in this niche are expensive. Bidding on them alone was unlikely to be profitable, and a generic search results page is a poor place to compete for attention.

So I did not try to win those terms head on. I built a two-layer intent system instead. The first layer is search: keywords grouped into themes and ad groups, each with a landing page of its own. The second layer is YouTube: the same ad groups reused as audiences, with AI video whose on-screen text matches each one.

How it works

1. Keyword research, grouped into themes and ad groups.

I researched the searches in the niche, grouped them into themes, and then split each theme into ad groups.

2. A landing page for every theme and every ad group.

Each keyword theme got its own landing page, and so did each ad group. The page a searcher lands on matches the query they typed. This is the relevance layer.

3. The same ad groups became YouTube audiences.

I reused those ad groups as custom-intent audiences on YouTube, so each audience is built from the same searches as an ad group.

4. AI video, made with two models.

The videos were generated with Seedance 2.5 and Google Omni.

5. On-screen text matched to each segment.

For each video, I swapped the on-screen text to match its custom-intent segment. The ad a person sees is written for the same query as the landing page they arrive on.

Why it works

The video text, the custom-intent segment and the landing page form one continuous intent chain. Each step speaks to the same query. That means expensive search terms are no longer the only route to the customer.

Highlights

• Built a two-layer intent system for go-to-market, so the business no longer depended on keywords too expensive to bid on profitably on their own.
• Grouped keyword research into themes and ad groups, and gave every theme and every ad group its own landing page, matched to the searcher's exact query.
• Reused the same ad groups as custom-intent audiences on YouTube, with AI video made in Seedance 2.5 and Google Omni.
• Swapped each video's on-screen text to match its custom-intent segment, so the ad a person saw and the landing page they reached spoke to the same query.