| Primary strength | Bookable supply and price comparison | Broad knowledge and open-ended answers | Destination discovery and itinerary creation | Turns a conversation into a trip you can follow |
|---|
| Trip structure | Bookings sit in separate order flows | Usually gives a text answer | Builds a visual itinerary | Chatbot and Trip Canvas share one confirmed state |
|---|
| China-local execution | Strong for hotels, tickets and transport in its inventory | Gives general advice, but execution varies | Depth depends on the destination | Keeps language, payment, transport, rules and recovery in one trip |
|---|
| Long-term traveller memory | Travel history and saved preferences | Conversation memory can be inconsistent | Preference filters for the current trip | Shows where each memory came from, then lets you confirm, edit or forget it |
|---|
| Sources and rechecks | Strong inside its own inventory | Does not always show its sources | Depends on the product and destination | Shows a source, confidence and recheck time for practical claims |
|---|
| Inventory and transactions | The strongest option for live inventory and checkout | No owned travel inventory | Depends on its booking partners | Has less inventory than an OTA. Sends users to official channels and can add channel integrations later |
|---|
| When the plan breaks | Customer service usually focuses on the booking | Can suggest options, but may not update the trip itself | May rebuild the itinerary | Today suggests a recovery plan, then shows a Canvas diff for approval |
|---|
