Project Type: Digital Transformation and Innovation Project — User-Centric Design
Author: Cátia Ribeiro | Published: Aug 29, 2024 | Read time: 11 min
Original article: https://medium.com/@catiaribeiro19/study-case-aiconnect-85c22c4fddb2
My role has been very versatile, from the initial problem discovery to the final design.
Stages:
- ✅ Initial Problem Discovery
- ✅ Analysis and Research
- ✅ Dashboard & APP Structure
- ✅ Design
I utilized data and information by conducting extensive research on current industry practices and trends in disruption management within the aviation sector. This included gathering data from industry reports, collecting feedback from passengers through surveys and social media, and analyzing existing solutions and competitors in the market to identify gaps and opportunities for differentiation.
- Many airlines use very outdated system structures, which limits the integration of new players of technology. Additionally, users are becoming increasingly demanding and expect more advanced and seamless interactions with apps and websites.
- Developing a product of this magnitude requires a significant engineering effort and resources.
- Implementing a backup system is crucial. Manual data entry is not a viable backup solution, necessitating the hiring of additional personnel, which increases payroll and training costs.
- Integrating data from multiple sources can hinder efficient management and the system must be able to adapt to increases in demand and complexity without losing performance.
- Generating relevant and precise AI/ML recommendations is vital to improve operational efficiency and user satisfaction.
Strategies to address limitations: Modernize infrastructures by collaborating with airlines and using middleware to facilitate integration without complete replacements. Design intuitive and advanced interfaces to meet the increasing demands of users and Airline Staff. Data integration use ETL (Extract, Transform, Load) solutions and flexible architectures like Lakehouse to efficiently manage multiple data sources. Develop and train precise models using varied data and continuous feedback systems to improve and adjust the models in real time.
The aviation industry frequently encounters disruptions such as overbooking, missed connections, and cancellations, leading to frustrated passengers, a tarnished airline reputation, and operational inefficiencies. These disruptions result in significant operational challenges, costing the industry approximately 8% of global airline revenues, estimated at $60 million annually.
Traditional manual methods for managing operations are slow and ineffective, relying on incomplete information and short-term considerations. This lack of accurate, real-time data prevents operational staff from anticipating the impact of their decisions, leading to high levels of stress and dissatisfaction among passengers.
- Disruptions have impacted over 50% of travelers, causing booking anxiety for nearly one-third.
- 69% of airlines stated that reaccommodating passengers during disruptions and cancellations is the most important benefit technology can provide.
- 72% of airlines indicated that improving data quality and analysis is their top priority for disruption management.
AIConnect emerges as an innovative solution with next-generation digital tools and advanced technology infrastructure to transform how airlines manage disruptions.
- Automating tasks associated with rebooking passengers and managing disruptions. #OperationalEfficiency
- Providing real-time information, offering solutions, and facilitating communication during disruptions. #PassengerExperience
- Creating a single source of truth to provide transparency and data access to the entire airline.
- Supporting decision-making to ensure the best outcomes for the entire organization.
- Demonstrating a proactive and data-driven approach to passenger care, fostering positive brand perception. #BuildingTrustAndBrandLoyalty
| Metric | Target |
|---|---|
| Reduction in average resolution time | ⬇️ 40% |
| Improvement in customer satisfaction | 📈 25% |
| Data Integration from all relevant sources | 🌐 90% |
- Traditional methods of handling flight disruptions are Slow and Inefficient
- Passengers experience high levels of frustration and stress during disruptions
- Airline need a robust Data-Driven approach to manage disruptions effectively
While AI plays a central role, creating a positive human experience remains crucial. AIConnect ensures that passengers receive friendly notifications from the airline through multiple channels, informing them of delays and suggesting alternative flight options instantly. Timely and accurate information significantly improves passenger satisfaction. Confirmations with updated details, including new boarding passes, are provided promptly. The AIConnect app is user-friendly and easy to navigate, especially during stressful situations like flight disruptions. It also supports multiple languages to cater to a diverse passenger base.
Customer Pain Points:
- Insufficient updates during disruptions.
- Extended waiting periods to reach customer support.
- Difficulty finding alternative flights.
- Conflicting information from various sources.
Ground Handling Pain Points:
- Insufficient access to real-time data for quick decisions.
- Poor coordination between operational teams.
- Inadequate communication between ground staff and passengers.
- Difficulty managing resources during peak demand or disruptions.
- Staff fatigue impacting decision-making during critical moments.
- Method 01: User Surveys
- Method 02: Complaint Analysis on Trustpilot
- Method 03: Analysis of Negative Mentions on Social Media via Brand 24
The graphic created shows the results of a sentiment analysis based on user complaints, collected through the Brand 24 platform from February 12 to March 13, the complaints focus on operational problems (Overbooking, Delays, Canceled Flights, Missed Connections).
User Research Summary: Employed a mixed-method research approach combining quantitative and qualitative data. Implemented a survey targeted at regular travelers to gather information on their experiences. Conducted an analysis of reviews and comments on public opinion platforms to identify trends and recurring themes.
Insight: "Although disruptions are somewhat frequent, they can cause serious inconveniences for passengers, as delayed, canceled, diverted, or overbooked flights lead to significant financial and non-financial harm for passengers."
When a disruption occurs, AIConnect anticipates every step the passenger will need to take to get back on track. Utilizing powerful automation, AIConnect sources and delivers all rebooking options and customer care resources seamlessly from a single interface.
The architecture of AIConnect is designed to integrate advanced AI and ML capabilities into an airline's operating systems, using a Lakehouse architecture to efficiently manage and analyze large volumes of structured and unstructured data.
- Data Collection: Collect data and store it in Azure Data Lake.
- Data Cleaning and Processing: Ensure data quality and prepare it for analysis.
- Model Training with TensorFlow and PyTorch: Train AI/ML models using the processed data.
- Model Storage in Azure: Store trained models in Azure for use in predictions.
- Real-time Prediction: Use stored models to make real-time predictions.
- Azure Functions: Handle real-time data processing and predictions.
- Passenger Notifications: Passengers receive notifications and relocation recommendations.
- Operational Dashboards: Display real-time updates and key metrics.
- Data Protection and Regulatory Compliance: Ensure data protection and compliance through encryption (Azure Security Center).
- Monitoring
- Enhanced Operational Management: AIConnect improves efficiency by automating routine tasks and providing real-time data analysis, enabling quicker decision-making and reducing delays.
- Resource Optimization: It optimizes resource utilization through predictive analytics, ensuring effective allocation of aircraft, crew, and ground services.
- Enhanced Passenger Experience: The system provides timely updates and proactive communication, keeping passengers informed about their flight status and rebooking options.
- Process Automation: AIConnect automates processes like flight rebooking and notifications, minimizing human error and freeing up staff to focus on critical tasks.
- Microsoft Azure (Data Lake, ML, Functions, API Management, Security Center, Monitor)
- TensorFlow and PyTorch for ML models
- APIs for real-time flight, weather, and reservation data
- Encryption and security tools
AIConnect handles scalability during unexpected demand spikes by utilizing a flexible, cloud-based architecture through Microsoft Azure. This infrastructure allows the system to automatically scale according to demand, ensuring it can manage large volumes of data and a high number of users simultaneously without degrading performance.
- Backup systems and Redundancies
- Disaster Recovery Plans (regular data backups and predefined procedures)
- Customer Communication during incidents
- Multi-Cloud Strategies
- Clarity: The Dashboard presents information clearly and concisely, allowing for quick decision-making.
- Comprehensive: Covers essential operations aspects from Flight status to passenger management.
- Usability: User-friendly interfaces with visual indicators and metrics summaries.
Starts with an overview of the operational performance of an airline in terms of regulations, passenger routes, and key indicators. Designed to provide an overview of flight operations, specifically focusing on managing disruptions such as delays, canceled flights, and overbookings, facilitating real-time monitoring.
Real-Time Monitoring dashboard designed to be used in real-time, allowing for the quick identification of flights with connection issues. The use of colors in the connection lines facilitates the immediate identification of the status of each connection, prioritizing attention on those that are at risk or have already failed.
An autonomous system powered by artificial intelligence that monitors and manages flight operations in real time. The system automatically makes decisions to resolve incidents, such as relocating passengers or notifying them of changes. Color-coded indicators and percentages allow for quick visualization of the status and progress in problem resolution.
Designed to enhance the passenger experience in flight cancellations, missed connections, or overbooking situations. When any of these issues occur, passengers receive an immediate notification through the AIConnect Operations system. The app automatically offers a rebooking alternative that passengers can accept with a single click, instantly receiving their new boarding pass.
AIConnect has been designed to offer a fully inclusive and accessible passenger experience:
- Screen reader compatibility
- Customizable interface (text size and contrast)
- Multilingual support
Assuming that the model training had been successful, significant results can be expected:
- An AI/ML model could expedite the detection of and response to operational issues, allowing for almost instantaneous corrective actions.
- The model's ability to offer real-time solutions to passengers affected by cancellations or missed connections can translate into an improved travel experience.
- By minimizing cancellations and optimizing flight alternatives, the airline can preserve more revenue by avoiding costly compensations.
AIConnect is a modular and independent Lakehouse system that can be developed and updated individually without affecting the entire system. This flexible and scalable architecture allows seamless integration with an airline's existing systems, facilitating the progressive training of AI/ML models without interfering with daily operations.
AIConnect is capable of collecting data from multiple channels, which not only improves operational management but also provides valuable insights for other areas such as marketing and sales. All data is stored in-house, ensuring the security and privacy of information.
The adoption of AI demonstrates the potential of the airline industry when data is fully leveraged. The story of AIConnect serves as a valuable lesson for other companies, demonstrating the fundamental role that AI, ML, and advanced analytics play in driving growth, efficiency, and a better passenger experience.
Digital Transformation AI Machine Learning Service Design Innovation Customer Experience Aviation UX Design Lakehouse Azure