Use cases
Ask in plain language; the assistant chooses the tools. These prompts work well.
Example prompts
| Goal | Prompt | Tools used |
|---|---|---|
| Market screening | “Rank Gulf cities by revenue per listing and show their median asking rate for 2-bedroom apartments.” | list_markets, then get_market for the leaders |
| Investment checks | “What could a 3-bedroom villa at 24.774, 46.738 earn per year? Show the comparables.” | estimate_revenue |
| Seasonality | “Compare occupancy by weekday in Riyadh and Dubai over the last 90 days.” | get_metrics twice with past_days 90 |
| Pricing research | “Summarize asking rates in Al Khobar by bedroom count and district.” | list_listings, paging with the cursor |
| Gulf comparisons | “How do Dubai and Doha compare on supply and median nightly rate?” | list_markets or get_market |
| Neighbourhood check | “What is occupancy in the rectangle 24.70–24.85 N, 46.60–46.80 E over the last 60 nights?” | get_metrics with bounds |
| Reporting | “Draft a one-page market brief for Abha with supply, rates and tracked performance.” | get_market, get_metrics |
Getting better answers
- Name the city and the window: "last 90 days" maps to
past_days. - Ask the assistant to state
maturityand how many listings are tracked, so you know how firm a figure is. - For listing-level questions in large cities, ask for a sample or a specific district: paging through every listing uses many requests.
- Ask for the comparables behind an estimate and open their links.
What it cannot do
- Change prices, bookings or anything on Airbnb, Gathern or Agoda.
- See photos, amenities or review text.
- Return history from before JokerPrice started tracking a destination.