National Airline handles 7 million calls with a Quiq AI agent, solving support calls 20% faster

This National Airline’s previous chatbot was built on rigid, menu-driven conversation flows: pre-defined decision trees that could only handle one issue at a time, in a fixed order. That structure created problems that compounded as the airline grew:

• Dead-end conversations. Simple questions became lengthy, transactional back-and-forths, and almost any complex issue immediately routed to a live agent, since the bot had no way to reason through a request it hadn’t been explicitly scripted for.
• A heavily regulated environment with no room for improvisation. As an airline, National Airline needed a partner that could navigate complex legal and technical requirements, including when a request had to be handled in English versus Spanish, without relying on rigid, predefined paths that couldn’t adapt to nuance.
• Fragmented channels with no shared context. Customers moved between SMS, voice, web chat, and messaging apps, but each touchpoint operated in isolation. A customer who called in couldn’t easily continue the conversation by text, and agents had no unified way to see what had already happened across channels.
• Reporting scattered across systems. With interactions split across multiple tools, the team had no single, aggregated view of conversation volume, topics, or AI performance, making it hard to spot trends or prove out ROI.
• A stalled Voice AI pilot. National Airline had already tried deploying voice automation with a different vendor, but the project failed to deliver a workable product, leaving the team wary of vendor promises and cautious about trying again.

The National Airline partnered with Quiq to build an agentic AI agent grounded in the airline’s own knowledge base, reservation systems, and business rules, replacing scripted flows with real reasoning.

• Quiq makes it easy to get data AI-ready: The Quiq platform automatically restructured the airline’s existing knowledge base and datasets for AI use, stripping unwanted HTML, adding related questions, and keeping every source in sync with the original material, without requiring the team to manually reformat anything.
• Deep platform integration: The AI agent connects bi-directionally into their core operational systems, including its Sky Speed reservation platform, AWS for sensitive passenger record data, Microsoft Dynamics for analytics, and its IVR system, so it can take action, not just answer questions.
• An orchestration layer built on the airline’s own rules: Quiq’s platform selects the right “Process Guide,” a defined set of instructions and tools, for each conversation, deciding when to escalate to a human agent, which language to respond in, and when to shift a customer between channels mid-conversation.
• Agentic reasoning across multi-turn conversations: The AI agent evaluates each message against conversational context, the active Process Guide, and live system data, letting it independently complete tasks like adding a checked bag or changing a flight rather than just describing how to do it.
• Consolidated Voice AI on the same platform: After the airline’s earlier Voice AI pilot with another vendor failed to reach production, Quiq built voice capability on top of the same agentic architecture already powering chat and messaging, letting them retire a separate vendor relationship instead of maintaining two disconnected systems.
• Centralized reporting: The team can now analyze conversations individually or in aggregate, across every channel, to track topic trends, AI performance, and where the experience needs tuning, instead of piecing together reports from separate tools.

• Adding a stroller to a reservation: When a customer messages about adding a stroller, the AI agent identifies this as a reservation modification, pulls up the correct Process Guide, verifies the customer’s identity, and updates the booking directly rather than pointing the customer to a generic FAQ page.
• Switching from voice to text mid-call: When a customer calls to review flight options, the AI agent recognizes that a list of times and prices is easier to read than to hear, and texts the options directly to the customer’s phone instead of reading them aloud over the call.
• Resolving IROP refunds end to end: During irregular operations, like weather-driven cancellations, the AI agent handles refund requests directly, reasoning through the airline’s specific policies rather than routing every disruption-related question to a live agent.
• Handing off with full context: When a request does need a human agent, that agent receives the complete conversation summary and relevant account details automatically, so the customer never has to repeat what they already explained to the AI agent.

For their guest care team, the shift wasn’t just about volume, it was about what kind of work was left once the AI agent took over the routine requests. Reservation changes, bag additions, and flight modifications that used to consume agent time now resolve automatically in most cases, freeing the team to focus on the complex, judgment-heavy situations that still need a person.

Escalations also changed shape. Because the AI agent hands off full conversation context, agents no longer spend the first few minutes of a call re-establishing what the customer already needs, which is a meaningful part of why escalated conversations now run 20% shorter. And with one platform now powering voice, chat, and messaging, the team gets a single, unified view of performance across every channel instead of stitching together reports from separate systems.

By replacing a rigid, script-based chatbot and a stalled Voice AI pilot with one agentic AI agent, this National Airline turned its self-service strategy into a measurable, scalable part of its customer experience operation.

• 40% and growing automated resolution rate across voice, chat, and messaging
• 16% reduction in average conversation time, driven by the AI agent’s ability to resolve multi-step requests without escalation
• 20% reduction in conversation time for escalated inquiries, since human agents now start with full context instead of from scratch
• 7M+ conversations handled through the AI agent in a single year, across a fully consolidated set of channels
• One vendor, one system replacing a previously failed, separate Voice AI pilot, simplifying both operations and reporting