Op Ed: Steve Clagg On Travel Programs As A Dead Node In Enterprise AI

Key Takeaways

• Travel platforms lack MCP servers, remain isolated from AI orchestration
• Sabre and Lumo demonstrate callable travel services via Model Context Protocol
• Vendors focus on traveler‑facing AI, not agent‑to‑agent interoperability
• Corporate travel directors must require MCP compatibility in RFPs
• Missing travel node forces AI teams to route around it, increasing costs

Pulse Analysis

Enterprise AI is moving beyond chat‑based assistants toward orchestration layers that can execute business goals autonomously. Platforms such as Microsoft Copilot Studio, Salesforce Agentforce and ServiceNow AI agents stitch together callable services—each exposing a standardized interface—so that a single request can trigger bookings, approvals, expense posting and more without human intervention. Travel management, however, has largely remained a siloed, portal‑driven function, leaving it invisible to the AI graph that powers the rest of the organization.

A handful of innovators are already proving the concept. Sabre launched an MCP server in September 2025, exposing its inventory as a machine‑readable service, while Lumo’s MCP endpoint delivers real‑time disruption probabilities and rebooking recommendations in a structured format. These examples illustrate how travel data can become a first‑class node in an AI workflow, enabling downstream agents to assess risk, trigger rebookings and feed expense data directly into ERP systems. In contrast, most vendor announcements—SAP Concur’s Joule assistant, Navan’s Ava, Amex GBT’s agentic search—still revolve around traveler‑facing chat interfaces, which do not satisfy the agent‑to‑agent requirement.

For travel leaders, the imperative is clear: shift procurement criteria from flashy AI features to proven interoperability. Ask vendors for an MCP server, agent‑callable APIs, and concrete policy‑parameter schemas that can be consumed by corporate AI agents. Embed these requirements in RFPs, secure a seat at the enterprise AI strategy table, and treat travel as a mandatory node in the workflow graph. Companies that embed callable travel services now will capture efficiency gains, reduce manual exception handling, and position themselves as essential partners in the next decade of AI‑driven enterprise operations.