Travel
Travel runs on supplier feeds, rate caches and a search box, and every one of them is a place to lose money or a customer. We build the models and agents that sit in that path: which suppliers to call, how to rank what comes back, how to map one hotel's rooms to another's, how to turn a conversation into a booking. They run inside the OTA's or aggregator's own stack, against live traffic.
Built for Travel.
- Supplier routing and ranking
- Recommendation engines
- Natural-language hotel search and reranking
- Hotel content enrichment
- Conversational booking agents
The work in Travel.
6 production models, deployed by non-technical teams.
+7%
conversion uplift, A/B tested
6
models live, run by PMs and analysts
60% lower infrastructure spend through smarter supplier routing.
-60%
infrastructure cost
A corporate booking platform with a policy-aware chat agent, in production.
~10,700
Automated API tests
52
Serverless functions
Room mapping for a hotel marketplace, live and watched for drift.
600
Room names per request
1,000
Hotels per batch job (pilot)
A live hotel search box, and a demo of searching hotels by sentence.
224
Countries covered by Autosuggest
5,000 requests a day
Free tier
Claude drafts the first reply to travel-agent support messages.
32
Request types
122
Unit and integration tests
Hotel records a traveller's own words can search, with evidence for each value.
55
Fields in the registry
44k
Stored answers in the shared cache
Your story next.
Bring us the decision or workflow you want AI to own. We'll show you the path to production.