Custom Travel Solutions (CTS) has introduced a platform designed to enable AI assistants to access live travel inventory and complete bookings within a single interaction.
The service, known as RouteStack, allows users of tools such as ChatGPT and Perplexity to search for hotels and be directed to a pre-filled checkout page with selected dates and pricing, where bookings can be completed.
The platform provides real-time availability and pricing data to large language model agents, addressing a limitation where AI systems have been unable to access live travel inventory. Until now, users searching for travel through AI assistants were typically redirected to external websites to complete bookings.
RouteStack connects AI applications to live supply, starting with hotel inventory. Additional services, including flights, car hire, holiday rentals and activities, are expected to be added from April. Cruises, rail and transfers are also planned.
The system is built using Model Context Protocol servers designed for AI environments. It enables AI tools to query inventory and return structured results within conversational interfaces.
A key feature is the use of deep links that direct users to pre-populated booking pages, allowing transactions to be completed without restarting the search process on another platform.
CTS said the platform is intended to support travel companies and developers seeking to integrate booking capabilities into AI applications. It allows suppliers to connect inventory while retaining control over pricing, customer data and fulfilment.
The launch comes as AI tools become more widely used in travel search and planning, while booking processes remain largely tied to existing distribution systems.
“This isn’t about building another chatbot,” said Mike Putman, CEO of CTS. “We’ve built the commerce layer AI agents were missing. The APIs, the deep links, the checkout infrastructure – those pieces already existed inside our broader ecosystem. We’ve now assembled them specifically for AI. This is the moment where conversational discovery becomes transactional.”
