Disneyland Paris: AI FastPass & Mobile Order.

CASE STUDY · HOSPITALITY & LEISURE

Real-time AI video analysis to enhance the visitor experience.

FastPass slot optimisation and personalised ad hoc offers across 15 million annual visitors: GDPR by design.

Client

Disneyland Paris (15M+ visitors/year)

Expertise

Computer Vision · Native mobile apps · Dedicated agile squad

Tech Stack

YOLOv8 · Faiss · Native iOS · Android · Azure · Edge inference

Engagement

12 experts · 30 months

CONTEXT

Project Context

Disneyland Paris wanted to improve the visitor experience while optimising its operations. The challenges: managing large volumes of video data in real time across 142 4K cameras, strictly respecting the privacy of the 15 million annual visitors, and integrating the system with the complex's existing infrastructure.

The legacy FastPass, based on statistical estimates, no longer reflected the reality of flows during attendance peaks. Visitors built up frustration and waiting time, and the potential for additional revenue per visitor (food & beverages, shops, premium experiences) remained largely untapped.

The challenge: create a smoother, more magical experience that generates more revenue per visitor, without ever compromising privacy or the park's “magical” perception. Everything had to stay invisible to the visitor, yet decisive for their experience.

Strategic Objectives

OBJECTIVES

(01)

AI FastPass optimisation

Analyse visitor density in real time across the whole complex (2 parks, 7 hotels, an entertainment centre), dynamically adjust FastPass slots and provide waiting-time estimates accurate to the minute.

(02)

Mobile Order Food & Beverages

Enable the 15 million annual visitors to order Food & Beverages from the Disneyland Paris app, with integrated payment, real-time order tracking and coordination with the complex's 30+ catering outlets.

(03)

Journey personalisation

Generate personalised ad hoc offers and journey recommendations based on visitor profiles (family, group of friends, fans of a given franchise), without ever identifying people by name, in strict compliance with GDPR.

Solutions Delivered by Adservio

SOLUTIONS

Adservio deployed a solution combining edge AI video analysis and native mobile apps, with a multidisciplinary agile squad (Vision Engineers, Mobile Engineers, Data Scientists, DPO Lead).

(01)

Image acquisition & edge vision

142 4K cameras deployed at strategic points in the park, with local edge inference for real-time processing of video streams. No storage of biometric data, immediate aggregation of counts.

(02)

Detection & vectorisation

YOLOv8 + Faiss pipeline to identify visitor densities, extract anonymous features (timestamps, aggregated bounding boxes) and convert them into vectors the orchestration engine can interpret.

(03)

kNN search & classification

Using the Faiss index to find the 5 nearest neighbours in terms of historical attendance patterns, selecting the most frequent class and applying the appropriate orchestration strategy.

(04)

Mobile Order Squad

Technology advisory, development and maintenance of the Mobile Order feature integrated into the Disneyland Paris app (native iOS/Android), with a premium UX suited to one-handed use in queues.

(05)

GDPR by design

Privacy-first architecture: no facial recognition, no individual identification, aggregation only, immutable audit trail and DPO certification. Compatible with the strictest requirements of the park and the parent company.

142 cameras, 48,720 visitors:
orchestration in under 200 ms.

The YOLOv8 + Faiss engine analyses 4K streams with edge inference, aggregates anonymous features and dynamically redistributes FastPass slots. No individual identification, aggregation-only, GDPR by design.

fastpass-engine.dsl · v3● live

# AI FastPass : orchestration engine

# Author: Vision Lead Disneyland Paris · GDPR

flow_engine "fastpass-orchestrator" {

cameras: 142 · 4K · edge inference

backbone: YOLOv8 + Faiss kNN(k=5)

privacy: no_face_id · aggregate_only

on("density.spike") {

redistribute(fastpass.slots)

push_offer(category: "food&bev")

notify("ops-control")

}

slo {

latency.p99 < 200ms

offer_relevance > 0.72

}

✓ Deployed · 15M visitors/year · revenue +16%

park-flow.live

142 cameras · 4K

Attraction density · real time

Saturday · 14:32

Big Thunder Mountain12 min

32%

Phantom Manor8 min

21%

Pirates of the Caribbean15 min

38%

Star Tours24 min

60%

Avengers Campus · Quinjet52 min

92%

It's a Small World6 min

18%

Crush's Coaster38 min

78%

Buzz Lightyear18 min

44%

Visitors on site

48,720

Push offers/h

1,240

Additional revenue

+€18K/h

[PEAK] Avengers Campus · density 92% · wait 52 min

redistribute → fastpass.slots · push_offer → food&bev · GDPR aggregate-only

Revenue

+16%

Attraction wait

−58%

Satisfaction

+26%

Edge AI cameras

142

Results

MEASURED FIGURES

+16%

revenue

Increase through FastPass optimisation and personalised offers.

−58%

waiting time

Reduction in waiting times at attractions through flow management.

+26%

satisfaction

Measured improvement in the visitor experience (NPS score).

< 200ms

p99 latency

Edge to orchestration to push-offer pipeline in real time.

142

4K cameras

Cameras deployed with edge inference, no biometric data.

15M+

visitors/year

Annual visitors served across the whole complex.

Impact

WHAT IT CHANGES

+16% in revenue

Direct increase through FastPass optimisation and real-time personalised offers: impact measured over 18 consecutive months across the whole complex.

−58% in waiting time

Massive reduction in waiting times at attractions thanks to intelligent flow management and the dynamic redistribution of FastPass slots.

+26% in satisfaction

Measured improvement in the visitor experience (NPS score), a smoother and more personalised journey, with the magical perception preserved and no visible technological friction.