Airbnb’s automation milestone: ai now resolves 33% of North American support issues
- Marijan Hassan - Tech Journalist
- 21 minutes ago
- 2 min read
CEO Brian Chesky vows "step change" in quality as platform pivots to an AI-native app that "Knows You".
During its Q4 2025 earnings call on February 12, 2026, Airbnb revealed a major shift in its operational DNA: approximately one-third (33%) of all customer support inquiries in the U.S. and Canada are now handled entirely by its custom-built AI.
The announcement signals that the travel giant has successfully moved AI from a "experimental chatbot" phase into a core fulfillment engine capable of managing complex transactions like refunds and booking modifications without human intervention.
Beyond the chatbot: A "custom-built" agent
CEO Brian Chesky emphasized that Airbnb isn't just using a generic off-the-shelf tool. The company has spent the last year building a proprietary "agentic" system.
The AI now resolves routine issues such as changing reservation dates, processing refunds, and troubleshooting host-guest disputes autonomously.
The push is being led by new CTO Ahmad Al-Dahle, a former Meta executive who led the Llama model team. Under his leadership, Airbnb is training models on its unique dataset of 500 million reviews and 200 million verified identities.
Global ambitions
After successful pilots in English, French, and Spanish across North America, Airbnb plans to roll out this AI support globally by the end of 2026, including new AI voice support for phone inquiries.
The "step change" in quality
Addressing common fears that AI leads to "shitty service faster," Chesky argued that automation actually improves the experience:
Unlike human agents who may interpret policies differently, the AI follows Airbnb’s global service standards 100% of the time. Better yet, Support is now truly 24/7 with zero wait times, a critical factor for travelers in different time zones facing "midnight emergencies" like lock-out issues.
The human safety net
Complex cases involving safety, discrimination, or high-value claims are still automatically escalated to specialized human teams.
The future: An app that "knows you"
The support milestone is just the first "pillar" of a broader 2026 roadmap. Chesky outlined a vision for an "AI-native" travel platform:
Conversational search: Moving away from rigid filters (dates/location), users are beginning to test natural language search (e.g., "Find me a quiet cabin with a chef's kitchen near hiking trails").
Proactive concierge: The app is evolving to act as a trip companion that suggests dining and logistics based on past preferences, effectively mimicking a high-end travel agent.
Competitive moat: Chesky argued that generic AI tools like ChatGPT can't compete because they lack the "fulfillment layer". That is the ability to actually process payments, verify IDs, and message hosts directly
Financials: AI as a margin protector
The move toward automation helped Airbnb post a strong $2.78 billion in Q4 revenue (up 12% year-over-year). By automating 33% of its most expensive operational cost - customer service - Airbnb is protecting its margins while forecasting low double-digit growth for the rest of 2026.
