Air India has expanded its Agentforce deployment across 30+ agentic AI initiatives, moving beyond initial refund automation to tackle multi-step customer workflows including email resolution, passenger name changes, and knowledge assistance. The airline's implementation demonstrates measurable operational gains: refund processing collapsed from 14 days to four hours, whilst name change requests dropped from three days to 30 minutes. Critically, Air India has engineered confidence thresholds at 95 per cent for autonomous responses, with human oversight retained for edge cases—a design choice that reflects the airline's evaluation framework balancing revenue growth, cost reduction, customer experience, and capability enablement. The scale of this rollout across 140 enterprise systems signals that Agentforce has matured beyond proof-of-concept territory into production-grade orchestration.
For CX teams already running Agentforce, Air India's approach offers a blueprint for moving beyond single-use automation into integrated agent ecosystems. The knowledge assistant component—surfacing policies, procedures, and fare rules through conversational interfaces—addresses a persistent friction point in support operations: agent context switching and policy lookup latency. However, the real strategic implication lies in how Air India has positioned these agents as force multipliers rather than replacements, explicitly retaining human judgment in high-confidence scenarios. This raises a sharper question for your own implementations: are you designing confidence thresholds and escalation logic that genuinely reflect your risk tolerance, or defaulting to vendor-recommended settings? The airline's four-parameter evaluation framework also suggests that isolated metrics—resolution time or cost per ticket—obscure the fuller picture of whether AI initiatives are actually driving revenue or merely shifting costs between channels.
The deployment's scope across a complex, regulated industry with high-stakes transactions (ticketing, refunds, name changes) establishes Agentforce as credible infrastructure for mission-critical workflows. For teams evaluating platform consolidation or expansion, Air India's success in integrating agents across legacy systems without wholesale replacement signals that Salesforce's approach to API-first orchestration is delivering on its promise. Yet the absence of detail around failure modes, customer satisfaction metrics, or agent acceptance rates leaves open whether these speed improvements translate to genuine experience gains or simply faster wrong answers.
Air India expands use of Salesforce's AI platform to streamline customer services ET Infra
Air India expands use of Salesforce's AI platform to streamline customer services inkl
Air India expands use of Salesforce's AI platform to streamline customer services ET Infra