Daniel Senyard is SVP of commercial platforms and innovation at travel management company FCM Travel.
A travel manager at a midsized enterprise client recently told me about a habit her team had picked up. Once a month, she exports a spend report from our reporting platform, uploads it into an AI chat tool, and asks it questions her dashboard was never built to answer. What changed in hotel spend in Germany last quarter? Which cost centers are drifting off policy? Draft me a memo explaining it.
Nobody built that workflow for her. She rebuilds it herself every month: exporting the data, explaining what every column means, telling the AI how to calculate each metric. The tool she actually wanted—one she could talk to, not just click through—didn't exist yet inside her travel platform. It existed, in pieces, everywhere else, and it left all the setup work sitting with her. And forget about cross-referencing that against her company's sales numbers to see if increased spend correlated with greater ROI!
That's the gap Model Context Protocol (MCP) technology is starting to close. Strip away the acronym and it's a simple idea: a secure, standard way to plug a business' own systems into an AI assistant so it can work with them directly, instead of a person exporting data and re-explaining it by hand each time. It's plumbing, not a product in its own right—but it's the plumbing that turns "upload a spreadsheet and hope" into "just ask". The industry has started calling this "headless," and the name is right—the value stops living in the interface and starts living in what's underneath it: the data, workflows, and content.
Enterprise travel reporting has always been bounded by what someone built in advance. Dashboards and report builders are useful, but only ever answer questions someone anticipated when designing them. What's changing is that a travel manager can ask a question nobody anticipated and get an answer that arrives already interpreted—not just a number, but what it means and what to do about it.
FCM Travel tested what that looks like. Ask our reporting layer to show hotel statistics with no date range given, and it defaults sensibly to the last year. Push further—"recommend a three-tier nightly rate policy from our actual spend, and justify it"—and it returns three priced tiers, each grounded in that client's real booking data. Close the loop: ask it to draft the traveler memo, summarize it for your CFO, and prep an email for review—and it's all sitting there, ready to send. One conversation, from question to policy to communication, without ever switching between your reporting tool, Word or Outlook.
The best businesses in a category aren't those with the best website or app, but those that set the standard or create the protocol and become the system of record. The authoritative and exclusive source of truth for a category of data. Salesforce is that system for customer data. Workday is that system for HR data. FCM Travel the system of record for the business traveler: the only place where company policy, stated traveler preference, real booking behavior, and preferred supplier terms converge into a single traveler profile. That's where the value lies in a headless future: the record persists even when the interface becomes optional, and the ability to query and combine the data in any way empowers the user to achieve whatever they want, without the limitations of the tool.
If a traveler can get a policy-compliant answer to a travel question or book a trip inside Microsoft Copilot, Teams or Claude without ever opening a dedicated app, the app was never really the point. The point was whether the answer was right, whether it was grounded in real negotiated rates and real booking history, and whether someone stands behind it when it isn't—because an instant wrong answer is still wrong.
Around 70 percent of FCM Travel clients' reporting is financial, and that's deliberately where we've focused first—numbers pulled from a governed source, not calculated on the fly by a guessing model. Traveler safety and duty of care sit outside that loop entirely, on purpose. AI isn't deterministic; a traveler's location during a crisis has to be.
The clients getting the most from this aren't asking about travel spend alone. They're asking cross-domain questions nobody could easily answer before: do the sales teams who travel more close more revenue? Do regions with flexible travel policy retain staff better? That used to mean a data warehouse project and months of work. Increasingly it just means two systems that can both be asked a question in one go—and someone willing to ask it.
That's the real promise underneath the talk of protocols and connectors: not that travel gets a chatbot, but that travel data stops being a silo. The technology making that possible should stay in the background. What it opens up for the businesses using it is anything but.