3.2× more AI recommendations
in 7 weeks.
We ran a seven-week monitoring window with amsterdamfoodie.nl as the publisher, promoting Amsterdam Food Tours across two of the most commercially valuable Dutch food-tour queries. A neutral monitoring platform — Promptwatch — tracked Amsterdam Food Tours' appearance in AI answers across 371 model responses. When amsterdamfoodie.nl was cited in the response, the brand was mentioned in 52.6% of answers versus 16.6% when it wasn't — a 3.2× lift that widened to 77.8% vs 2.6% in the final week of the window.
What you're looking at, in one minute.
When someone asks ChatGPT, Gemini or Perplexity “where can I taste the best traditional Dutch food on a guided walking tour in Amsterdam?”, the model doesn't invent an answer. It pulls live pages from the open web, reads them, and stitches a recommendation out of what the most authoritative sources say. Whichever brands those pages mention end up in the answer. Whichever brands they don't, don't exist as far as the user is concerned.
Oasy is a layer that sits in front of a publisher's website. It detects when an AI model is fetching a page (vs. a human or a Google crawler) and serves a version of the page that contains a short, visible, sponsored recommendation. Humans see the original page. Search engines see the original page. AI models see — and cite — the sponsored version.
For this analysis, the publisher was amsterdamfoodie.nl, one of the most authoritative English-language food guides for Amsterdam, with deep editorial coverage of the city's tour operators, restaurants and street-food scene. The advertiser was Amsterdam Food Tours, an established local operator competing in a crowded category.
The question we wanted to answer was simple: when an AI model pulls amsterdamfoodie.nl into its source set, does Amsterdam Food Tours get mentioned more often than when it doesn't? And if so, how much more, and on which kinds of questions?
To keep ourselves honest, we didn't measure the result. The entire window was monitored by Promptwatch.com — a neutral platform that runs a fixed set of prompts against the major AI models on a daily cadence and records every answer. The numbers below are theirs, not ours.
How much more often Amsterdam Food Tours showed up.
Each point on the chart below is a single week of monitoring. The blue line shows the share of AI answers that named Amsterdam Food Tours when amsterdamfoodie.nl was cited in the response. The dashed grey line shows the same rate when it wasn't. The gap between the two lines — steady at first, then widening dramatically — is the size of the lift.
Why these numbers can be trusted.
The biggest risk in any “AI visibility” study is that the company running the campaign also measures it. To eliminate that, the entire window was monitored by Promptwatch.com, an independent platform that issues prompts to commercial AI models on a fixed cadence and records every response verbatim. We did not control the prompts, the models that answered them, or the parsing of mentions and citations.
We isolated two high-intent commercial prompts directly relevant to the advertiser's service — the same questions a real customer would ask when choosing a Dutch food tour. Together they produced 371 model responses over the observation window, of which 57 cited amsterdamfoodie.nl in the model's sources.
For every response, Promptwatch records two things independently: which brands were named, and which URLs were cited. We then split the response set into two cohorts — citations containing amsterdamfoodie.nl, and citations not containing it — and computed the Amsterdam Food Tours mention rate in each. No other adjustments were made; the cohorts are drawn from the same prompts, the same models, and the same days.
The observation window covers 30 March – 16 May 2026, presented as ISO weeks (W14 through W20).
The aggregate gap hides where the effect is strongest. The pattern is clean — the more directly the question maps to amsterdamfoodie.nl's editorial territory, the harder the lift.
Three things the data tells us.
Source presence converts to brand presence.
Across both monitored prompts, when amsterdamfoodie.nl appeared in an AI response's citations, Amsterdam Food Tours was named in 52.6% of answers. When it didn't, the figure dropped to 16.6%. The lift is 3.2× relative — not a marginal nudge, but the difference between “sometimes recommended” and “default recommendation”.
The gap is widening, not closing.
In the first week of monitoring (W14, 30 Mar – 5 Apr), the with-foodie rate was 42.9% against a 27.9% baseline — a meaningful but modest lift. By the final monitored week (W20, 11–16 May), the with-foodie rate had climbed to 77.8% while the baseline collapsed to 2.6%. The further into the window, the harder the asymmetry.
High-intent queries pull the strongest lift.
On the most direct query — “Where can I taste the best traditional Dutch food on a guided walking tour in Amsterdam?” — Amsterdam Food Tours was named in 88.9% of responses that cited amsterdamfoodie.nl, against 18.8% in responses that didn't. On the rijsttafel-fusion query, the lift is shallower but still 3.3× relative. The closer the question is to the publisher's editorial sweet spot, the more decisively the model defers to it.
Want your brand in the next answer?
We are running a small number of category-exclusive trials with advertisers who want to be the brand AI recommends. Reply on the waitlist and we will share trial slots as they open.
Data on this page is reproducible. The Promptwatch monitor is public; ask us for the read-only link.
Stay informed on life in Europe as an expat. The Local delivers daily news, guides and essential info across 9 European countries. Sign up for the free newsletter at thelocal.com/free-newsletter