Dalsa Booking Assistant
Budget: $15 – $25 USD
Build an AI-powered voice agent that lives inside our airline’s mobile app and speaks to travellers as naturally as a seasoned customer-service rep. The agent must:
• Understand and act on spoken requests in English, Spanish, French, Kiswahili, German and Turkish.
• Securely search live inventory to book one-way and round-trip itineraries, confirm seats, and push boarding‐pass details straight to the user.
• Manage check-in, deliver gate or delay updates in real time, and guide passengers through same-day rebooking or full cancellations without human escalation.
• Personalise every reply by drawing on the customer’s frequent-flyer profile and past trips.
• Remain responsive 24/7, even during peak loads.
Please integrate automatic speech recognition, natural-language understanding, dialogue management and text-to-speech into a single flow; common stacks such as Dialogflow CX, Rasa, Amazon Lex or a custom Python/TensorFlow pipeline are all acceptable so long as the final SDK drops cleanly into iOS and Android codebases and hooks into our existing REST/GraphQL booking APIs.
Acceptance criteria
1. End-to-end voice demo (mobile emulator or TestFlight build) showing booking, check-in, status query and rebooking paths.
2. Latency under two seconds for recognition and response on a standard 4G connection.
3. Unit tests plus a short deployment guide that my internal team can follow to move from staging to production.
If this scope is clear, let’s discuss timelines and the data you’ll need to get started.
• Understand and act on spoken requests in English, Spanish, French, Kiswahili, German and Turkish.
• Securely search live inventory to book one-way and round-trip itineraries, confirm seats, and push boarding‐pass details straight to the user.
• Manage check-in, deliver gate or delay updates in real time, and guide passengers through same-day rebooking or full cancellations without human escalation.
• Personalise every reply by drawing on the customer’s frequent-flyer profile and past trips.
• Remain responsive 24/7, even during peak loads.
Please integrate automatic speech recognition, natural-language understanding, dialogue management and text-to-speech into a single flow; common stacks such as Dialogflow CX, Rasa, Amazon Lex or a custom Python/TensorFlow pipeline are all acceptable so long as the final SDK drops cleanly into iOS and Android codebases and hooks into our existing REST/GraphQL booking APIs.
Acceptance criteria
1. End-to-end voice demo (mobile emulator or TestFlight build) showing booking, check-in, status query and rebooking paths.
2. Latency under two seconds for recognition and response on a standard 4G connection.
3. Unit tests plus a short deployment guide that my internal team can follow to move from staging to production.
If this scope is clear, let’s discuss timelines and the data you’ll need to get started.