Unlocking the Future of Airline Customer Service Using AI: Insights From Our Recent Endava and Microsoft AI Hackathon

The airline industry is no stranger to challenges, especially in the realms of customer service and social media engagement. Recognising these pain points, we recently organised a two-day hackathon in collaboration with Microsoft , aimed at revolutionising customer interaction with artificial intelligence (AI).  

This event brought together four teams formed of business representatives from the travel industry along with AI specialists and mentors from Microsoft to tackle two primary challenges: enhancing social media customer reviews solutions and rethinking customer service for disruptions. 

Enhancing social media customer reviews solutions 

Airlines know the importance of timely and meaningful engagement with customers on social media. However, statistics show the average response times can vary widely from minutes to hours, with many comments going unanswered, while responses that are provided could often be improved.  

The team decided to introduce an agentic AI approach to solving this. An AI agent is a software entity that uses artificial intelligence to autonomously execute tasks and make decisions without the continued support or need for full-time human oversight.  

At Endava, we have developed solutions based on AI agents, further information on Agentic AI can be found on our website. During the hackathon, the team successfully created three AI agents using ChatCompletion API and Semantic Kernel functions. 

The winning team’s development focused on four main areas:  

1. Categorising customer issues to streamline responses 

2. Analysing the frequency and nature of complaints 

3. Seamlessly integrating AI to social media forums  

4. Creating a chatbot for enhancing automated customer service capabilities. 

The outcomes of this solution could potentially include:  

1. Reduced response time: The AI-driven solution slashed response times in the hack event from 30 minutes to mere seconds, ensuring that customer queries and complaints are addressed almost instantaneously. 

2. Higher quality responses: By incorporating the airline’s brand voice and policies, the AI agent provided more relevant and valuable responses, enhancing customer satisfaction. 

3. Increased efficiency: Automating this process could lead to significant cost savings, reducing the need for extensive human intervention. 

4. Enhanced employee wellbeing: With AI handling repetitive and high-volume tasks, airline employees could focus on more complex, higher-value customer issues, improving their work-life balance. 

5. Dual response streams: The solution created two distinct response streams – public and private – allowing for more nuanced and appropriate interactions depending on the nature of the customer’s issue. 

Winning team leader Radu Orghidan , our VP of Cognitive Computing, said: "We were thrilled to work on this important challenge. We feel fortunate to have swiftly gotten started by setting the stage with brief personal introductions and outlining the tools we were familiar with – prompt engineering, machine learning, Semantic Kernel and Autogen. 

“Analysing the business problem with the support of the two assigned subject matter experts was crucial. Development began in the afternoon on the first day and continued well into the evening, and even throughout the night when our thoughts were clearer and the atmosphere was calmer.” 

Rethinking customer service for disruptions  

Disruptions such as lost baggage, delays and cancellations are critical moments that test customer loyalty. Traditionally, managing these situations has been a labour-intensive process prone to delays and inconsistencies.  

During the hackathon, our teams reimagined this aspect of customer service by integrating generative AI with chatbots and instant messaging systems. 

Our innovative application of this technology resulted in: 

1. Enhanced efficiency: The integration of AI with instant messaging solutions allowed for quicker resolution of issues, significantly reducing wait times and improving overall efficiency. 

2. Empathetic interactions: Designed to consider cultural contexts and demographic factors, the chatbot offered fast and empathetic responses appropriate to the customer’s situation. 

3. Automation: By automating routine inquiries and responses, the use of AI freed up human agents to tackle more complex and nuanced problems. 

4. Consistency: The use of generative AI ensured that all interactions were consistent in quality and tone, maintaining the airline’s brand integrity across all customer touchpoints. 

Adopting AI for the future  

By embracing innovative AI solutions, airlines can position themselves as industry leaders, setting new standards for customer engagement and satisfaction. Along with streamlining operations, AI offers airlines the opportunity to build stronger, more personalised relationships with customers. In an increasingly competitive market, this differentiation can be a key driver of customer loyalty and positive brand perception, ultimately boosting the airline’s revenues.  

The hackathon demonstrated just how swiftly technology can be applied to revolutionise customer service and achieve these goals. Strategic AI innovations not only can enhance customer satisfaction but could also lead to significant operational efficiencies and improved employee wellbeing. 

Ensuring human oversight in AI solutions 

While the benefits of AI in customer service are substantial, it is crucial to maintain human oversight to manage potential risks, such as AI hallucinations or inappropriate responses. Our approach integrates a policy knowledge base and a ‘human in the loop’ mechanism, where human agents review and monitor AI-generated responses to ensure accuracy and brand alignment.  

Key measures include: 

1. Retrieval Augmented Generation (RAG) mechanism: Our AI agent is strictly bound to answer using only the information from the available knowledge base leveraging a RAG approach. 

2. Regular audits and reviews: Human agents periodically audit AI responses to ensure consistency and quality. 

3. Fallback mechanisms: In cases where AI encounters complex or ambiguous queries, it escalates the issue to human agents. 

Claudiu Stancu , our EMEA Client Partner Lead for Travel, said: “Seeing the power of the latest AI technology was absolutely impressive.  

“As the teams proved during the hackathon, the challenges that the airline and the wider travel industry have traditionally faced can be solved differently, and with solutions that provide a wider impact for their businesses.  

“This is still the beginning of a major transformative journey for the industry, and we already see the benefits of adopting AI technology in different parts of the business. 

“We’re already discussing with our clients the positive impact AI can have on their businesses, and we’re lining up multiple hackathons, proofs of concepts (PoC) and projects aligned with their business ambitions.” 

Stay tuned for more detailed materials and insights from this exciting event. The future of airline customer service is bright, and these AI-driven solutions could be at the forefront of this transformation.