How TreeHouse Grew LLM Traffic and Conversions With Promptwatch

Promptwatch showed us exactly how our brands were performing in LLMs and, more importantly, what we needed to do about it. Over the last few months, we've already seen traffic and especially conversions take off significantly.

TreeHouse operates four rental platforms across the Netherlands and Europe, including Huurwoningen.nl and Pararius. Its platforms connect people looking for rental properties with landlords and real estate agents.

With multiple brands, large amounts of content and a significant organic search presence, SEO had long been an important acquisition channel for the marketing team. But as people increasingly started using AI to search for information, the team saw a new growth opportunity.

AI Search was already on the agenda, as we knew our customers were asking questions to AI, but we still had no idea what AI was saying about our brands, and if we were actually showing up in those answers.

TreeHouse already had a mature SEO operation. But traditional SEO platforms weren't built to answer the questions the team now had about AI Search. Tools like Semrush and Ahrefs could show traditional search rankings, but they couldn't tell TreeHouse:

• Which brands were being recommended by LLMs
• Which competitors were appearing alongside them
• Which sources AI was citing
• Which prompts were driving visibility
• What information AI was using to form its answers

Initially, the team tried to solve this themselves. They would ask ChatGPT and Perplexity a small set of prompts hundreds of times a day and manually record the results in spreadsheets. It provided an early signal, but it wasn't scalable across four platforms.

Before, we'd mainly measure AI Search visibility by asking ChatGPT or Perplexity three prompts hundreds of times a day, and reporting those results in a spreadsheet. It was a very manual process.

TreeHouse needed a way to measure AI Search at scale, identify what was influencing visibility, and turn those insights into actions.

TreeHouse started using Promptwatch to monitor how its brands appeared across AI Search and understand what was influencing those results.

The team started with branded prompts to establish a baseline: what does AI actually say about each TreeHouse brand? From there, they expanded their prompt monitoring to organic and competitor prompts, with the help of the Promptwatch Google Search Console integration. Monitoring those prompts showed there were technical and content foundations to fix on their websites.

That insight helped us shape the strategy for AI Search: proprietary data was what LLMs valued the most.

We kept banging our heads against the wall: it's not a priority because the business case is hard to make. That was finally turned around with Promptwatch.

TreeHouse didn't just increase its visibility in AI Search. It started seeing a measurable impact on the bottom line.

By tracking AI referrals through to paid subscription trials, the team discovered that AI Search now drives approximately 25% of its organic conversions, with 99% of AI-referred conversions coming from ChatGPT. That means roughly 1 in 4 organic search conversions now comes from AI Search, and the channel is still growing.

Over the last few months, we've seen traffic and especially conversions take off significantly in LLMs.

As AI Search became a bigger acquisition channel, TreeHouse also streamlined how the team analyzes it, using the Promptwatch MCP with Claude. They can now investigate conversion spikes, analyze what changed, and generate a weekly report covering what went up, what went down, and what to focus on next.

This saves the team hours of analysis and reporting every week, turning AI Search data into actionable insights without adding more manual work. As Promptwatch continues to ship new features, TreeHouse plans to adopt them to identify new opportunities, act faster, and keep improving its AI Search strategy.

When asked what advice he would give to companies starting to invest in AI Search, Dave kept it simple:

Do the boring work first. Look at what's already on your website, which blogs from 2014 are still live, where the data is inconsistent, and what's no longer true. Fix those foundations before you start optimizing and scaling. And, of course, choose a tool that helps you get the right information.