Airlines ride AI: Lesser workload for customer service agents, human-like responses for passengers

New Delhi:  From human-like responses to reduced workload for customer service agents to utilisation of trillions of parameters for answering air passengers’ queries, artificial intelligence is providing the technological tailwinds for Indian airlines as they cater to rising traffic.

For Air India, which has embarked on a five-year transformation plan, Artificial Intelligence (AI) use will be “pervasive” and its generative AI virtual agent AI.g handles over 1,300 topics.

And the country’s largest airline IndiGo has AI chatbot 6Eskai that has 1.7 trillion parameters, allowing it to answer questions with ease. Also, the bot can understand written, typed language, and verbal instructions using speech-to-text models.

Akasa Air, which describes itself as a “cloud-native and digital-native brand”, said it will continue to invest heavily in proven technology solutions across all business functions.

India, one of the world’s fastest growing civil aviation markets, is seeing rising domestic air traffic, and airlines are also expanding their fleet, including with wide-body aircraft.

Across industries, including aviation, the adoption of new-age technologies like AI and Machine Learning (ML) is on the rise.

Air India’s Chief Digital & Technology Officer Satya Ramaswamy told PTI that “AI-infused” is an integral part of the airline’s technology approach and that AI has enabled it to reduce the variance in responding to customer service requests.

“Once developed and trained, our AI agent AI.g delivers consistent responses. It has helped us streamline and sharpen our contact centre functions. It has helped our contact centre agents focus more on queries and requests that require complex and detailed interactions with customers and areas where a response cannot be pre-determined,” he said.

According to him, computer vision, which refers to analysis of images and videos using various algorithms, will be an integral part of the operations, be it in airports, maintenance bases or in airplanes.

An IndiGo spokesperson said that several enhancements are in the pipeline to elevate the capabilities of its AI chatbot 6Eskai. “These include the integration of Google’s Gemini, a tool which can easily engage in conversations across various formats such as pictures, text, speech, music, and even computer code. Additionally, we plan to incorporate the IndiGo Avatar and voice features to streamline the booking process seamlessly,” the spokesperson told PTI.

Besides, another feature is being planned to offer passengers more options in the event of operational issues.

“The AI chatbot efficiently addresses customer queries in 10 different languages and offers a first-of-a-kind platform for booking tickets across the entire network. The implementation of 6Eskai has led to a remarkable 75 per cent decrease in the workload for customer service agents,” the IndiGo spokesperson said.

Currently, 6Eskai — which boasts an impressive 1.7 trillion parameters, allowing it to answer a diverse range of commonly asked questions — is transferring only 2-9 per cent of chats to experts.

Akasa Air, which has been flying for less than two years now, said it will continue to invest in appropriate technologies such as optimisation and AI to help gain insights into its customer behaviour and operational parameters.

“We leverage these insights to make informed and data-backed decisions to serve our customers better. Additionally, we have also deployed AI-based cyber threat prevention capabilities that leverage the power of AI to identify and thwart cyber threats in real time,” Akasa Air Co-Founder and Chief Information Officer Anand Srinivasan told PTI.

He also noted that all operational and commercial decisions are backed by technologies that leverage data, mathematical models and, where appropriate, AI and Internet of Things (IoT) .

“These include the pricing of our flights, choice of routes, deployment of aircraft, rostering of operating and ground crew, scheduling engine maintenance etc”.

According to Ramaswamy, the containment ratio of AI.g is 93 per cent, which means only 7 per cent of the queries are escalated to a human agent. It handles over 1,300 topics and provides human-like responses to the various ways in which customers ask questions related to these topics.

Citing an example, he said the airline had injected pet policy in the AI system. “When customers ask ‘can I take my labrador with me?’, AI.g reasons that labrador is a dog, a dog is a pet, and hence the laid down pet policy applies. This is a quantum jump from the earlier generation of technologies where one would have to train the AI model with every possible type of dog breed,” he noted.

At Air India, Ramaswamy said a generative AI agent has been deployed recently for answering employee queries.
“Our use of generative AI in CoPilots for a variety of use cases will grow as enterprise users are loving the ability to interact in natural language in dynamic ways. We anticipate a time when generative AI may front-end all of our data,” he added.

Going forward, Akasa Air’s Srinivasan said the airline will continue to utilise advanced technologies that will provide improved efficiency and passenger experience.

Ganesh Gopalan, CEO of conversational AI company Gyani.ai, said there are many opportunities for AI in the aviation sector. For instance, for rescheduling of tickets, generally the waiting time has been long and an AI application can help in reducing it, he added.