%20(1).png)
About GTAA
Greater Toronto Airports Authority (GTAA) operates Toronto Pearson airport, one of the largest airports globally. GTAA serves 50 million passengers per year, welcoming those who enter Canada.
The issue
Extended immigration waits affected the airport's reputation, on-time performance and passenger experience. The consequences of chaotic queues were extensive:
- Passengers experienced delays, waiting inside planes for an average of 2 hours due to an overflowing immigration area.
- These prolonged waiting times resulted in missed connecting flights and further delays.
- Pax would have a negative first impression upon arriving to the airport.
An analysis of the sentiment from 1,500 Google reviews reveals that 76% of passengers most commonly associate stress with “'customs,' 'immigration,' and 'layovers.'”
The challenge
Accurate visibility on arrivals PAX flow and wait times were needed quickly before the summer peak. To maintain immigration queues under 20 minutes, it's crucial for airports to adapt staffing and operations using real-time passenger flow information.
This data helps accommodate varying passenger profiles based on flight timings and origins.
GTAA's objectives
- Enhanced passenger experience: Passengers needed transparency about their airport journey to alleviate anxiety.
- Operational efficiency for customs agencies: Accurate data was pivotal to ensure optimal staffing during high-traffic periods.
- Quick implementation for minimal disruption: With the summer travel surge on the horizon, swift deployment was non-negotiable.
- No extensive closures or business process alterations: The solution couldn't involve extensive closure of the airport areas or changes to the existing business process.
Technical requirements
- Privacy: It was imperative to process passenger data anonymously, preserving individual privacy.
- Re-use existing infrastructure: Existing security cameras were re-used to support swift deployment, reducing CAPEX.
- Accurate wait time per queue type: GTAA has multiple queue types tracking different categories of passengers. Balancing between passenger types can be complicated, so it’s important to have granular data on each.
- Flexibility: The airport's existing infrastructural challenges, like low ceiling heights, needed to be taken into account, alongside future modifications.
Why prior solutions fell short
The immigration process, with its myriad of entry options, amplified the complexity. Whether it was e-gates, kiosks, agent podiums, booths, or fast passes, the array of choices and constant merging and separating of queues made queue management a significant challenge.
Existing Wi-Fi solutions couldn't deliver the granular data essential for optimal queue management, notably:
- Offered a broad overview, not taking into account the variable waiting times across different queues that could have a big difference in expected wait time.
- Couldn't provide accurate process rate and wait times in a complex queue environment.
- Were not efficient in group settings, restricted zones, or changing layouts.
How Zensors AI made a difference
Zensors AI stepped up with a tailored solution, capitalizing on its advanced machine learning capabilities, specifically designed for intricate queue scenarios.
5 key achievements
- Quick implementation by utilizing existing cameras and integrating additional ones where necessary.
- An impressive 95.8% accuracy when matched against manual human validation.
- Real-time data visibility on 17 distinct queue lines with corresponding wait times, enhancing collaborative decision-making.
- Detailed insights into queue movements, service rates, and the passenger-to-staff ratio, ensuring optimal staffing.
- Seamless integration with third-party terminal performance monitoring systems.
Accuracy validation
Zensors validate the accuracy of AI captured data on passenger flow and predicted wait time employing a third party.
Our third-party validator randomly selects an hour long video that captures the queue area of interest, then performs:
- Flow assessments: Count the number of people entering and exiting queues on a per-minute basis, and compare this to Zensors’ platform data.
- Wait time assessments: Record when a person crosses the entrance and exit tripwires, and compare this to Zensors’ platform data.
- Reflect physical changes to locations: The client ensures that the locations for entry and exit tripwires and queue areas are correctly captured in the Zensors platform.
%20and%20Zensors%20AI%20observed%20wait%20(min).webp)
Highlights
Visibility in peak period performance - Ensuring that the terminals are optimally staffed.
- Ability to drill down into specific products and areas to identify and treat exact causes of long queues.
- Understanding of queue movement and service rate vs #PAX waiting in line to know that the staff to queue ratio is optimal.
- Data is fed into third party systems that the team is already using to view terminal performance data such as internal planning and BI systems.
The impact: 81% less wait time compared to 2022
In summer 2022, the client-observed average wait time was 30 minutes across all hours for the immigration process. However, by summer 2023, data from July 1 to August 28 shows that the average wait time had reduced to just 5.84 minutes, representing an 81% decrease from the previous year.
Summary
- Tracking 17 queues in real time using existing camera infrastructure.
- 81% less wait time at immigration terminal as compared to last year.
- No queue complaint tweets during the summer of 2023 (as of August 16) and a significant improvement in Google Location reviews regarding the immigration process.
For a more in-depth exploration on advanced queue management, dive into our detailed how-to queue blog, where we elucidate our approach and techniques.
