Lockey
Business travel sounds simple: book a flight, take the trip, file the expenses. In practice, it often involves delays, unclear policies, and slow approvals that make even routine trips harder than they should be. Lockey emerged from Lockton's challenge as the single, trusted companion that carries that cognitive load across the entire journey so the traveler can focus on the work that made the trip necessary.
What It Does
Lockey is a proactive AI copilot that guides the traveler through every stage of a business trip in one continuous, context-aware conversation.
Before the Trip
Lockey outlines key requirements like visas, passports, and vaccinations, and explains company policy in plain terms. It builds a personalized checklist and ranks flight and hotel options based on cost, compliance, and flexibility, with clear explanations so the traveler can choose confidently.
Approvals
Lockey fills out approval requests, shows status updates in simple terms, and if something is rejected, explains the reason and suggests how to fix it quickly. The traveler does not need to dig through policy documents.
During the Trip
Lockey tracks live flight updates and responds when plans change. If there is a delay, it highlights rebooking options, explains what is covered, and provides the right contacts to move forward.
Issues & Exceptions
When something goes wrong, Lockey summarizes the situation, lays out next steps, and points the traveler to the right person when direct action is needed.
After the Trip
The traveler can scan receipts with their camera, and Google Gemini extracts key details like merchant, amount, and currency. Lockey then files the expense report and updates the trip record in MongoDB.
How We Built It
Next.js
Next.js runs both the frontend and backend in one place, keeping the app fast and consistent. It handles data updates on the server so the interface stays responsive, even when AI features are working in the background.
Google Gemini 1.5 Pro
Google Gemini 1.5 Pro is used to understand and act on travel-related input. It takes natural conversation and turns it into structured trip details like dates, destinations, and purpose. From there, it helps compare options using company policy and user constraints, and produces clear explanations rather than raw results. It also supports key workflows like summarizing policies, generating rebooking guidance during disruptions, and extracting expense data from receipts, all tied directly to the trip context so outputs stay grounded and usable.
MongoDB Atlas
MongoDB Atlas stores the entire trip as one evolving record. Instead of separate systems for each stage, everything builds on the same data structure as the trip progresses. This allows flights, hotels, approvals, receipts, and expenses to stay connected in one place. It also enables fast policy search, location-aware hotel results, and live flight tracking, so the system always reflects the current state of the trip.
ElevenLabs TTS
ElevenLabs adds a voice layer that adapts to context rather than staying static. The assistant uses different tones depending on the situation, calm for planning, upbeat for approvals, empathetic for issues, and urgent when timing matters. This makes the same information feel more appropriate to the moment and helps the assistant come across as responsive rather than mechanical.
Challenges We Ran Into
Bringing everything into one place was difficult. Business travel is split across many third-party systems, which makes it hard to track things like flights and hotel bookings. To keep the experience consistent, we had to connect these systems and, in some cases, hard-code parts of the flow.
The project was also complex because of the number of possible scenarios. We had to plan for many different inputs and outputs to make sure the system behaved correctly. Even then, we ran into integration issues and merge conflicts that took time to debug and fix.
Accomplishments That We're Proud Of
What stands out most is that Lockey feels alive. This was a different type of system than we had built before. It required combining several tools and APIs, which made development complex, but seeing everything work together in the end made the effort worthwhile.
What We Learned
Good retrieval depends on good data. How we split and indexed the handbook mattered as much as how Google Gemini was prompted. Well-structured data leads to accurate, grounded answers. Poor structure leads to confident but wrong results.
Voice also shapes how the system is experienced. Using ElevenLabs with the right tone changes how the same information is received. A calm, empathetic voice feels more natural and builds trust, showing that tone is part of the user experience, not an add-on.
Privacy & Safety
| Principle | Notes |
|---|---|
| Minimal Collection | Only essential trip data is stored. Sensitive information like card numbers, CVVs, and SSNs is masked before storage |
| Sanitized AI Context | Sensitive identifiers such as passport numbers are excluded from AI input. Receipt PII is masked before being sent or stored |
| Sanitized Persistence | Receipt data is cleaned before being written to the database to ensure no raw sensitive data is stored |
| Access Control | All sensitive routes are protected with authentication checks and reject unauthorized requests |
What's Next for Lockey
A native iOS and Android app brings the full copilot experience into the traveler’s pocket, designed for quick, one-handed use while moving through airports or on the road. In the same flow, push notifications keep users updated on gate changes, check-ins, approvals, and expense deadlines so important details are not missed in real time.
Trip plans also extend into the user’s existing tools through Google Calendar, where confirmed flights, hotel stays, and meetings are automatically synced into a single view. At the same time, existing calendar events are checked to prevent conflicts before anything is booked. Each itinerary also includes live flight tracking, so status updates are always available without switching to external tools.
Beyond scheduling, the system supports end-to-end trip execution by including hotels and conference tickets alongside flights. This allows full travel planning and booking to happen within one continuous conversation instead of separate tools and steps.
Built With
- elevenlabs
- google-gemini
- mongodb
- next.js
- nextauth.js
- openweathermap
- serpapi
- tailwind
- typescript
Log in or sign up for Devpost to join the conversation.