FareHarbor runs the booking engine for more than 23,000 tour, activity, and attraction operators: whale watching out of Honolulu, canal cruises in Amsterdam, climbing trips in Colorado. When one of those operators cannot take a booking, they do not lose the sale until later. They lose the departure.
Support on a Departure Deadline
Most software support can wait an hour. FareHarbor’s cannot. An operator who finds their checkout misconfigured at 7am has until the 9am departure to fix it, and the seats on that boat cannot be resold once it leaves the dock. The clock is set by the schedule, not by an SLA.
The operators are small businesses, and the person contacting support is usually the owner — between tours, on a dock, on a phone, often with a group already waiting. They are not at a desk composing a ticket. Why is the 9am not showing as available? Why did that booking not go through?
Demand is seasonal and geographic at the same time. A Colorado ski operator and a Sydney reef operator peak six months apart, and FareHarbor works from six offices — Denver, Amsterdam, Boston, Honolulu, San Francisco, and Sydney — spanning nearly every timezone. The Honolulu morning rush is the Amsterdam night shift.
Then there is weather. One storm cancels a day of departures, and the operator needs dozens of bookings rescheduled or refunded at once, before their own customers start calling.
Open’s AI Agent
The AI agent handles support across email, chat, and voice. Voice carries more weight here than in most industries: an operator standing on a dock with a group boarding can call and get an answer, which is not something a ticket form can do.
For booking problems, the agent opens the reservation, works out whether the cause is availability, pricing, or a checkout setting, and corrects it. For payments, it verifies the transaction, explains where a deposit or refund is sitting, and processes what it can.
Weather cancellations are handled in bulk rather than one message at a time. The agent works through the affected departures, applies the operator’s own cancellation policy, and issues the reschedules or refunds.
Setup questions — availability rules, seasonal pricing, resource assignment, embedding the booking flow into an operator’s own site — are answered against that operator’s actual configuration, not a generic help article. Custom API integrations, contract questions, and payment disputes go to FareHarbor’s team with the full thread attached.
The Results
Open’s AI agent now handles FareHarbor’s operator support:
| Metric | Impact |
|---|---|
| Automation rate | 87% |
| CSAT score | All-time high |
| Response time | Under 30 seconds |
| Channels covered | Email, chat, and voice |
What matters is whether the answer arrives inside the operating window. An operator whose checkout is fixed before the 9am departure keeps the revenue from that departure. One who hears back that afternoon does not.
The Impact
87% automation across email, chat, and voice lets FareHarbor carry 23,000+ operators through seasonal peaks without scaling headcount proportionally. When a region enters high season, the load follows it automatically.
FareHarbor’s own team moved to the work that needs them: complex integrations for large operators, and the platform changes that come out of what support keeps seeing. Through the Booking.com integration, the operators they serve reach travelers in 115 countries and more than 5,000 cities.
