Air India is expanding its use of Salesforce Agentforce to automate more customer service processes, moving beyond its initial deployment for refund processing. The airline is now using the artificial intelligence (AI) platform for multi-intent email requests, passenger name changes, and faster access to operational knowledge for service teams.
The expansion is part of Air India’s broader use of agentic AI across its digital operations. According to the company, its cloud-first digital ecosystem spans more than 140 enterprise systems, with more than 30 agentic AI initiatives underway.
Refund processing drops from days to hours
Air India’s first Agentforce deployment focused on refund processing, a process that can involve policy checks, multiple validations, backend systems, and dependencies between teams.
Agentforce coordinates these activities as an end-to-end workflow. Air India says this has reduced refund turnaround time from approximately 14 days to about four hours.
The airline says the results show how AI agents can move beyond answering customer questions and instead execute defined business processes.
AI handles multiple requests in one email
The latest deployment includes Automated Email Resolution. Customer emails can contain several requests involving refunds, baggage, loyalty, and other services.
Agentforce identifies the different intents in an email and uses specialized sub-agents to investigate requests, validate information, and carry out required actions. It then combines the results into a single response using an Air India-approved template.
For eligible cases, responses are generated only when the system reaches a confidence level above 95%. Human oversight is retained where required.
The system also uses customer interactions and service team interventions to generate insights that can be used to improve prompts, workflows, and model performance over time.
Name changes now take 30 minutes
Passenger name corrections can require several eligibility checks, validations, and ticketing actions. Air India says Agentforce now coordinates these steps, including initiating ticket reissuance and delivering the updated ticket to the customer.
The airline reports that the process has gone from approximately three days to 30 minutes.
Knowledge assistant supports service teams
Air India has also deployed a Knowledge Agent for customer service representatives. The AI assistant provides conversational access to policies, procedures, fare rules, and operational guidance.
Previously, service teams could need to search across multiple knowledge repositories to find relevant information. Air India says the new system reduces information retrieval time and hold times during live customer interactions.
The airline also reports improvements in response consistency, agent onboarding, and first-contact resolution rates.
Air India plans broader agentic AI use
Air India evaluates its agentic AI initiatives across four areas: increasing revenue, reducing cost, improving customer experience, and enabling capabilities that were previously not possible.
Dr Satya Ramaswamy, Chief Digital & Technology Officer, Air India, said the focus is shifting from AI that assists employees to AI that can understand, decide, and execute within defined guardrails.
Arundhati Bhattacharya, President and CEO, Salesforce South Asia, said the partnership combines customer data, AI agents, and human judgment to make service more responsive.
For Air India, the latest deployments represent another step in its shift toward automated service workflows. The airline says its focus remains on applying AI to customer needs while retaining accuracy, governance, and human oversight.