Verified evidence from recorded runs. Figures are snapshots from the cited runs, not live claims about today.
1. Compare one hotel with two Luxembourg competitors
The Local AI Presence Monitor run rd2OGxwfywNZObpME measured Yasha Hotel Mersch and its website, then compared it with Hotel Le Place d'Armes Luxembourg and Hotel Parc Belair Luxembourg.
Recorded output:
- Yasha Hotel received a baseline presence score of 75.
- Le Place d'Armes showed rating 4.4 from 866 reviews in that run.
- Parc Belair showed rating 4.1 from 802 reviews in that run.
- The run stored nine competitor comparisons.
- It used technical dry-run billing: real Google and website observations, zero charged Actor events, with platform usage billed separately.
Public task: https://apify.com/travelmonitorlab/local-ai-presence-monitor/examples/three-luxembourg-hotels-ai-presence-comparison
Detailed proof: https://adagasolutions.lu/apis/local-ai-presence-monitor.html
Why it matters: a hotel can establish a repeatable baseline instead of manually checking its reviews, website signals and two competitors. A baseline is not a guarantee of future change detection: the monitor must be run again with the same target before claiming a delta.
2. Measure three fixed-model answers without calling them universal visibility
The AI Brand Visibility Tracker run rTzuzBVg0h45cFz3L sent three fixed queries about Yasha Hotel to recursive/flint-v1.0:free through OpenRouter.
All three answers were measured and mentioned Yasha Hotel. None was marked as citing the official yashahotel.com domain. Two answers contained one model-provided citation each, illustrating why total citation URLs and official-domain citations must remain separate metrics.
Public task: https://apify.com/travelmonitorlab/ai-brand-visibility-tracker/examples/yasha-fixed-model-multi-query-brand-tracker
Detailed proof: https://adagasolutions.lu/apis/ai-brand-visibility-tracker.html
A separate competitor proof, run dk1Hi70ELwgSUp01I, compared Adaga solutions and Manus across prompts and models. It retained three real answers across two models and one provider error. The error was not counted as a successful answer.
Public dry-run configuration: https://apify.com/travelmonitorlab/ai-brand-visibility-tracker/examples/brand-p2-multi-model-competitors
Why it matters: the output is evidence for a named model, prompt and timestamp. It is not a ranking guarantee, a representative sample of all AI users or proof that every statement in the answer is true.
3. Preserve published hotel rates and their provenance
Hotel Rate Monitoring run 7ObEuoyCma8msaGhq, on build 0.5.4, fetched two Visit Luxembourg pages in 3.686 seconds and returned two hotel cards with zero source errors.
Hotel Martha, Am Kaesch, L-7593 Mersch:
- Single room: published minimum 91, maximum 136.
- Double room: published minimum 111, maximum 151.
- Triple room: published minimum 147, maximum 197.
Hostellerie Val Fleuri, 28 Rue Lohr, L-7545 Mersch:
- Double room: published minimum 108, maximum 128.
- Suite labels: minimum 201, maximum 195, preserved exactly as an inverted source anomaly rather than silently corrected.
The dataset contains each source URL, fetch timestamp, HTTP status and HTML SHA-256 hash. EUR is an input label because the source rate labels do not repeat a currency.
Public task: https://apify.com/travelmonitorlab/travel-monitor-launch/examples/mersch-luxembourg-published-hotel-rates
Real dataset: https://api.apify.com/v2/datasets/HzYByHb18ia1v9p0T/items
Detailed proof: https://adagasolutions.lu/apis/travel-monitor-launch.html
Why it matters: published min/max cards can be monitored for changes and anomalies with inspectable evidence. They are not live availability, date-specific booking quotes, OTA parity scores or proof of a pricing opportunity.
Reusable lesson
Use three different tools for three different questions:
- Local AI Presence Monitor for repeatable business and competitor observations.
- AI Brand Visibility Tracker for model-specific brand-answer evidence.
- Hotel Rate Monitoring for published source rate cards.
Combining them into one synthetic “visibility” or “parity” score would hide the limitations that make each dataset useful.