Law Firm
Resultado
-80% time sorting email
AI email organiser for a law firm
Three lawyers who were losing court notifications among hundreds of emails now receive a daily summary of what matters, with attachments organised by case.
Ver caso
Volver a casos de uso
Travel Agency
AI knowledge assistant for a travel agency
Agents took 25 minutes per enquiry searching prices in PDFs and spreadsheets. Now they ask the AI assistant and have the answer in 3 minutes with the original source.
El problema
Qué estaba pasando
The agency works with 40+ suppliers, each with their own PDF catalogue, Excel price table and special conditions by email. When a client asks about a destination, the agent has to search 3-4 documents, cross-reference prices with seasons, verify cancellation conditions and check for live offers. The process takes 20-30 minutes per enquiry. Each agent stores information their own way, so when one is on holiday, nobody finds their files. New agents take about 3 months to learn all destinations and conditions.
La solución
Qué implementamos
We built a specialised AI assistant that ingests all catalogues, price tables, supplier conditions and emails with special offers. Documents are processed, converted to embeddings and stored in Supabase. Agents access the assistant from an internal web interface where they ask in natural language. The assistant responds with updated prices, specific conditions and a reference to the source document for verification.
Resultados
Lo que medimos
Números concretos después de la implementación. Medidos sobre un periodo de 3 meses posterior al despliegue.
Métrica 1
3min
per enquiry instead of 25 minutes
Métrica 2
2 wks
new agent onboarding (was 3 months)
Métrica 3
0
errors from outdated information
Stack
Tecnología utilizada
Herramientas probadas, fiables y sin vendor lock-in. Integraciones limpias que escalan con tu negocio.
RAG (embeddings)
GPT-4o
Supabase
PDF processing
Web interface
¿Tu negocio tiene un problema similar?
Analizamos tu caso sin compromiso y te decimos si la automatización tiene sentido para ti — con números, no con humo.
Caso de uso basado en implementaciones reales. Los datos han sido adaptados para proteger la privacidad del cliente.
