Start from the retail campus model (no sign-up needed; running a simulation is Pro). Before a show, the box office and the concessions counter queue, and some miss the start of the film. In this article, in the retail campus model’s cinema, your AI simulates the half hour before a 19:00 show. It ran the same tools an AI calls over MCP. Every figure and number below comes from that run.
1. The cinema and the half hour before the show
- The cinema’s 2nd floor has the box office (a 6 m counter) and the concessions counter (7 m) in its lobby, and four screens to the south.
check_spacefinds nothing.
The AI wrote this flow as a scenario (put_process_scenario).
- For the 19:00 show on screen 3, 120 come in from the bridge between 18:30 and 18:55 (most often at 18:45).
- Those who bought their tickets online (50 %, say) go straight on; the others queue at the box office and buy at a free window (30 seconds to 1.5 minutes, most often 40 seconds).
- 60 % queue at the concessions counter and buy at a free till (40 seconds to 2 minutes, most often 1, 1 minute 13 seconds on average). Then they go in to screen 3 and take a seat.
- Every number and time is an assumption. Put in your own cinema’s records.
2. The box office windows, and the concessions tills
The AI changed the box office windows open (2, 3) and the concessions tills (2 to 4) (sweep_process_parameters), 20 runs each. With the same scenario stopped at 19:00, it also counted those not yet in their seat then. The waits are a person’s mean; those queuing, those not seated and the last seated are means over the 20 runs.
| Box office windows | Concessions tills | Wait for tickets (mean) | Wait for snacks (mean) | Queuing for snacks (most at once) | Not seated at 19:00 | Last seated |
|---|---|---|---|---|---|---|
| 2 | 2 | 3.9 min | 8.2 min | 25.6 | 32.3 | 19:18 |
| 2 | 3 | 3.9 min | 2.5 min | 14.5 | 16.0 | 19:06 |
| 2 | 4 | 3.9 min | 0.6 min | 7.9 | 9.3 | 19:04 |
| 3 | 2 | 1.0 min | 9.7 min | 33.4 | 29.3 | 19:18 |
| 3 | 3 | 1.0 min | 3.9 min | 20.6 | 12.5 | 19:05 |
| 3 | 4 | 1.0 min | 1.4 min | 11.8 | 1.0 | 18:59 |
- With two windows and two tills, 32.3 are not in their seat at 19:00, and the last takes their seat at 19:18. The wait for snacks is 8.2 minutes on average, and at the most 25.6 queue for them at once.
- The 69.0 buying snacks take about 84 minutes of serving between them: 42 minutes even at two tills, more than the 25 minutes they come in over. With three tills, 16.0 are not seated at 19:00; with four, 9.3.
- A third window with two tills still leaves 29.3: tickets are bought sooner, so the wait for snacks grows from 8.2 to 9.7 minutes. Three windows and four tills leave 1.0.
3. The share who buy online
With two windows, the AI changed the share who buy their tickets online (30, 50, 70 %) and the tills (2, 3).
| Buy online | Concessions tills | Wait for tickets (mean) | Wait for snacks (mean) | Not seated at 19:00 | Last seated |
|---|---|---|---|---|---|
| 30 % | 2 | 8.2 min | 5.1 min | 43.8 | 19:19 |
| 30 % | 3 | 8.2 min | 0.7 min | 31.8 | 19:14 |
| 50 % | 2 | 3.9 min | 8.2 min | 32.3 | 19:18 |
| 50 % | 3 | 3.9 min | 2.5 min | 16.0 | 19:06 |
| 70 % | 2 | 0.8 min | 10.1 min | 29.8 | 19:18 |
| 70 % | 3 | 0.8 min | 4.2 min | 13.0 | 19:05 |
- Even with 70 % buying online, two tills leave 29.8 not seated at 19:00, because the wait for snacks grows to 10.1 minutes. The slowest part of the way in is the concessions counter: speed up the tickets, and as many more queue for snacks.
4. How to use this check
- Draw your own cinema (from a PDF of the drawings or photos too), and tell your AI each show’s audience and when they come, the share who buy online, the time for tickets and snacks, and the windows and tills. Your AI also runs a moved concessions counter or a mobile-order pick-up on the drawing.
- Not checked: the other screens’ audiences at the same time, ticket machines, mobile orders, the ticket check at the door, the toilets, and the length of the trailers.