How smart search works

1. Starting your search

For flights: Enter your origin, destination, and travel dates as usual.

For hotels: Enter any text describing where you want to stay — this could be a hotel name, a neighbourhood, or even a specific building or office address. You'll also need to specify your check-in and check-out dates and the number of travellers.

2. Sharing your trip context and preferences

Once you've entered your search details, a context panel will appear. You'll be asked to provide:

Trip Context: Describe your trip in plain language. You don't need to use specific terms — just write as if you were explaining your trip to a colleague. The AI reads your description to understand your intent and sets relevant filters automatically on your behalf.

Examples:

• "Flying to Bangkok for a client presentation. Need to arrive before 9 AM Thursday. Prefer Singapore Airlines."
• "Conference starts Wednesday at 2 PM — I'd like to arrive the night before to settle in."
• "I have a meeting at Google's HQ in Singapore on November 6th at 9am. I'll arrive the day before and need a room with breakfast. I'm looking for a 3 or 4 star hotel, with my nightly budget under USD 200."

The more specific you are, the more accurately the AI can filter results to suit your needs.

3. How the AI uses your information

Smart Search cross-references your trip context against the following inputs:

• Trip Context: For flights, the AI extracts your intent — when you need to arrive, which airline you prefer, whether you want a direct flight, or if you're flexible on timing. For hotels, it determines the most relevant search location and identifies applicable amenity filters (e.g. breakfast included).
• Travel Policy: Every result is checked against your company's travel rules before being shown to you. This includes budget limits, cabin class restrictions, advance booking windows, and approved airlines or hotel tiers. Options outside policy won't surface as top results.
• Lowest Logical Fare (LLF): LLF defines a reasonable fare or rate ceiling for your route and travel class. Smart Search factors this in automatically. If your preferred option exceeds the LLF threshold, the AI will flag the trade-off clearly so you can decide whether to proceed.
• Approvals: Some bookings may require manager sign-off — for example, if they exceed a budget or involve a cabin upgrade. Smart Search surfaces these flags upfront with a short reason, so you're never surprised at checkout.

💡 Note: The AI prioritises compliance first. Results that fall outside policy or LLF may not surface prominently, but will still be accessible if you choose to view all results.

4. Viewing your results

Flights: Instead of a long list, you'll first see your Top 3 AI-Recommended Flights, each with a short explanation of why it was chosen. For example:

• "Arrives by 8:45 AM — meets your deadline. Direct flight. Within policy and LLF."
• "Singapore Airlines as requested. Slightly above LLF — approval will be required."
• "Best value within policy. 1 stop, total travel time under 5 hours."

Your top recommendations stay pinned at the top of the page even when you click View all flights, so you can compare them against the full results at any time.

Hotels: Results are displayed with all AI-selected filters pre-applied (e.g. star rating, maximum nightly budget, breakfast included). Rather than a ranked shortlist, you'll see all available hotels that match your criteria, presented as: "These are the available hotels based on your preference/criteria." The choice is yours.

If the AI can't find results that meet all your criteria, it will explain why — such as policy restrictions, limited availability, or an LLF conflict — so you can adjust and search again.

5. What to expect (and watch out for)

• Compliance comes first. Results will always reflect your company's travel policy and LLF rules, even if that means a personally preferred option doesn't appear at the top.
• Transparency on trade-offs. If no option fully meets all your criteria, the AI will tell you why rather than just showing a confusing list.
• Be specific. The more precise your context, the better the results. For flights, "arrive before 9 AM" works better than "morning flight." For hotels, an office address or landmark gives the AI more to work with than just a city name.
• Flight results are recommended; hotel results are filtered. For flights, you get a curated Top 3. For hotels, you get a pre-filtered list — the ranking and final choice is up to you.
• Approval flags are informational, not a block (flights). Seeing an approval label means the booking will be routed for manager sign-off — it doesn't prevent you from proceeding.
• Your inputs are editable. For hotels, the AI pre-fills the trip context from your initial search entry, but you can always update it before running the search.