BCD Travel introduces MCP: AI in business travel solves complicated API integrations
One of the world's largest corporate travel management companies has integrated an innovative open-source framework into its Tripsource platform, opening the door to autonomous AI agent collaboration and simpler data management.
In the world Travel tech industry, connecting various systems – from airline and hotel inventories to internal corporate travel policies – has meant hours and hours of programming effort for years. Each new integration required specific API keys, custom code, and constant maintenance. But, BCD Travel decided to change that dynamic by implementing the Model Context Protocol (MCP) framework within its Tripsource platform. This technological leap allows systems to communicate and exchange data with each other using natural language. This means complicated and expensive custom integrations are slowly becoming a thing of the past.
How MCP works as a universal data adapter
To understand the importance of this change, the easiest way is to draw a parallel with travel power adapters. Instead of carrying a specific charger for every country you travel to, MCP functions as a universal plug which enables artificial intelligence to connect to any system instantly. This applies no matter how it is developing in the background. Yannis Karmis, Executive Vice President of Product Planning and Development at BCD, points out that the aim is to completely change the way the industry consumes travel services and analytics. They are also focusing exclusively on the end result for the client.
This approach directly affects the architecture of modern Travel Management Companies (TMCs). Instead of rigid IT structures, a flexible ecosystem is created where applications and AI agents can draw information from various sources. This is possible without the need for constant code rewriting.
Real-time integration of bookings and travel policy
One of the most concrete benefits for corporate clients is visible in the search and purchasing process itself. Traditional systems often function separately: the traveller first checks flights or hotels, and only at a later stage is it verified whether these options comply with the company's internal policy. On the other hand, BCD's new portfolio model merges flights, hotels, car rentals – and soon rail travel – into a single interface.
Within this interface, content and corporate travel policies are integrated simultaneously. AI agents automatically filter options, displaying only compliant and preferred alternatives to the user. This bypasses the lengthy “shopping” and comparison phase, shifting the focus directly to quick and secure bookings.
Chatting with the database instead of writing code
Apart from the travellers themselves, this technology brings a significant change for managers who oversee travel programmes. The analysis of vast amounts of information within corporate databases (data lakes) has previously required advanced analytical tools or the assistance of the IT department. By using natural language processing (NLP) through the MCP interface, analytics becomes accessible to everyone.
Decision-making now boils down to simply asking a question. Systems can directly extract deep insights into consumption, trends, or passenger behaviour based on a simple text query. This dramatically speeds up decision-making and real-time budget optimisation.
The added value lies in so-called “agent-to-agent” collaboration. Within the system itself, various specialised AI agents – for instance, those responsible for finance, security, or logistics – can share context with each other and solve complex tasks without human intervention. By opening the platform to clients and partners, BCD Travel clearly shows that the future of business travel does not belong to isolated systems. Instead, it belongs to open platforms that enable rapid and collaborative innovation.
