For years, booking a business trip meant opening a corporate travel portal, entering an origin and destination, clicking through pages of flight options, checking hotel rates, and hoping everything fit the company travel policy. In 2026, that familiar routine is starting to look surprisingly old-fashioned.
Artificial intelligence is moving into the center of corporate travel, turning booking from a sequence of searches and filters into something that increasingly resembles a conversation with a knowledgeable travel coordinator. Some systems can already interpret natural-language requests, apply company rules, surface appropriate flights and hotels, help make changes, and connect the resulting transactions with expense systems. The shift is significant enough that the Global Business Travel Association says artificial intelligence, data, and new retailing models are helping shape the next era of managed travel.
There is an important reality check, however. AI has not transformed every corporate travel program at the same speed, and some of the industry's most ambitious agentic features are still being introduced, expanded, or piloted. In a March 2026 GBTA survey of 269 travel buyers in North America and Europe, 58% said AI had made little or no impact on their programs so far. At the same time, 83% expressed interest in conversational booking, 85% in AI-powered traveler support, and 89% in automated disruption management and rebooking. The result is a business travel market caught between old workflows and a new model that is arriving remarkably quickly.
The Search Box Is Turning Into A Conversation
The most obvious change is happening right at the beginning of the booking process. Traditional online booking tools require travelers to translate a fairly simple goal into a series of form fields, filters, fare comparisons, and individual searches. Conversational AI reverses that relationship by letting the traveler describe the trip in ordinary language and asking the software to translate that request into travel options. KAYAK for Business, for example, introduced AI tools in 2026 that allow users to describe what they need while the system considers their traveler profile and company policies when suggesting an itinerary. Travelers can then review the options before making their selections.
That may sound like a cosmetic change, but it alters the work involved in arranging a trip. A traveler might care about arriving before an afternoon meeting, avoiding an unnecessarily early departure, staying near a client office, and complying with a hotel-rate limit. Conventional booking tools often require that traveler to express those preferences through separate searches and filters. A conversational system can potentially interpret the request as a whole, then narrow the available inventory before the traveler has to evaluate it. KAYAK says its business product can also handle questions and trip changes conversationally, with a handoff to a live agent when additional support is needed.
Other travel companies are moving in the same direction, but some are pushing beyond the travel website itself. American Express Global Business Travel said in its second-quarter 2026 results that it launched an Egencia AI connector in Anthropic's Claude. According to Amex GBT, the connector allows travelers and enterprise AI agents to book and manage policy-compliant air and hotel transactions from the AI environment rather than requiring travelers to begin the process inside a conventional booking portal. The company also said it had expanded Egencia's conversational AI into Google Chat and was piloting conversational AI in Microsoft Teams for Neo customers. That makes the booking interface less like a destination employees visit and more like a service that follows them into their normal workday.
Perk, the company formerly known as TravelPerk, made a similar move in August 2026. It announced that customers could book and manage trips and submit expenses through compatible AI assistants including Claude and ChatGPT, with company approval processes still applied behind the scenes. Perk also said its AI-connected workflow can route out-of-policy bookings or expenses into the employer's existing approval process. The broader pattern is becoming clear across several competing platforms. Instead of teaching employees how to navigate another piece of corporate software, travel companies increasingly want the software to understand employees where they already work.
Your Company Policy Is Becoming Part Of The Booking Engine
The real breakthrough in business travel is not simply that AI can recommend a flight. Consumer travel sites have been experimenting with recommendation technology for years, and generative AI has made itinerary planning much easier. Corporate travel has a more difficult job because a reasonable trip for an individual traveler may still be unacceptable to an employer. There can be negotiated airline and hotel rates, cabin restrictions, price caps, advance-purchase requirements, approval rules, preferred suppliers, and different policies for different employee groups. Modern corporate travel AI is increasingly being built around those constraints rather than treating them as an afterthought.
That structure may actually give managed business travel an advantage as agentic AI develops. Skift reported in August that Amex GBT and KAYAK for Business had both moved into conversational booking systems capable of completing policy-compliant transactions while keeping the traveler involved in final confirmation. Corporate programs already contain much of the information an AI agent needs to determine what it is authorized to do, including policies, negotiated rates, approval chains, and audit requirements. Those rules create boundaries around the AI's choices. In other words, the complexity that once made corporate booking frustrating can also provide guardrails for automation.
Travel buyers appear much more comfortable with AI when those boundaries are clear. GBTA found that 95% of the buyers it surveyed were comfortable with AI recommending flights and hotels based on negotiated rates, while 92% were comfortable with AI generating custom reports. Their comfort declined when AI was given more sensitive information or more autonomy. Only 64% said they were comfortable with AI accessing employee calendars to recommend travel, and 57% were comfortable with AI changing or canceling a booking. Those numbers suggest that travelers and companies are not simply asking whether AI works. They are also deciding exactly how much authority it should receive.
SAP Concur is also using AI to change the way companies create the rules themselves. At its Fusion 2026 event, the company announced tools designed to let administrators create travel rules from natural-language instructions or uploaded company policy documents. SAP also described an integration between its Sales Cloud and Joule Booking Agent in which planned client visits can be connected to personalized travel options for a salesperson. Instead of maintaining policy logic only through specialized administrative interfaces, companies are beginning to turn written policies and business context into machine-readable instructions. That could eventually make corporate travel policy less visible to employees precisely because more of it is being enforced automatically during the booking process.
Business Travel Is Moving Into The Apps Where Work Already Happens
One of the biggest changes in 2026 is not what the AI can do but where employees can use it. Corporate travel has traditionally lived in a dedicated ecosystem of booking tools, travel-management-company websites, mobile apps, approval portals, and expense systems. That creates friction every time an employee has to leave the software being used for actual work and enter a separate administrative workflow. Several travel companies are now trying to remove that boundary. Their goal is to make booking a trip something an employee can initiate from the same conversational software used to manage other work.
Amex GBT's Egencia connector for Claude is one of the clearest examples. The company says the integration can give travelers and enterprise AI agents access to policy-compliant air and hotel transactions without forcing them to leave the AI environment. Its expansion into Google Chat and its Teams pilot point in the same direction. KAYAK for Business says its AI Connector is designed to let travelers access its corporate booking capabilities from everyday work tools. Perk announced comparable capabilities for external AI assistants and said it planned to bring its agent into Slack in September 2026.
WavebreakmediaMicro, Adobe Stock
Navan is building around the same idea from an infrastructure perspective. In August, the company described its Model Context Protocol implementation as a standardized bridge connecting external AI tools with Navan's travel and expense platform. Navan argues that a general-purpose AI model does not inherently know an employer's travel policies, actual travel inventory, or internal expense information. Connecting the AI to those systems gives it the business context required to answer company-specific questions or eventually take actions within authorized boundaries. The significance for travelers is that the visible interface may increasingly become interchangeable while the travel-management platform operates behind it.
This helps explain why Model Context Protocol, commonly shortened to MCP, has become such a prominent piece of the industry's AI conversation. Phocuswright reported in February 2026 that 61% of travel businesses in its research were experimenting with or scaling agentic AI, including 6% that were actively scaling and another 22% beginning to scale. The research specifically highlighted MCP as one of the developments that had helped set the stage for agentic systems in travel. The practical consequence is that an employee's preferred AI assistant may increasingly act as the front door while specialized travel platforms supply inventory, policy, payment, and servicing capabilities behind the scenes. The old question of which booking website an employee uses could eventually become a question of which AI interface the employee prefers.
AI Is Starting To Manage The Trip After You Book It
Booking is only a small part of what makes business travel difficult. Flights get canceled, connections become impossible, meetings move, hotel stays change, and travelers sometimes need assistance when they are already standing in an airport. That is why travel companies are putting substantial AI investment into servicing rather than concentrating only on trip search. In GBTA's 2026 survey, automated disruption management and rebooking attracted interest from 89% of travel buyers, slightly higher than the 83% interested in conversational booking. The numbers suggest that some of AI's greatest value may arrive after the reservation has been made.
SAP Concur and Amex GBT are building that kind of assistance into their Complete platform. In March, SAP Concur said an integrated chat system with automated handoff to a live travel counselor was in pilot, with SAP's Joule AI being added to help travelers answer common questions using Amex GBT marketplace information. The companies have also connected Egencia booking information with Concur Expense so eligible itinerary and e-receipt data can move into the expense process automatically. The idea is to connect booking, service, and post-trip administration rather than treating them as separate tasks. For travelers, that means a conversation that begins with finding a flight may increasingly continue through changes, support, and reimbursement.
KAYAK for Business takes a similar hybrid approach. Its 2026 AI tools can be used to ask questions or request changes, but the company says the conversation can transfer to a human travel agent when additional help is required. The agent receives the trip details and previous chat context so the traveler does not have to begin again from scratch. That handoff is important because business trips can become complicated quickly once cancellations, international fares, multiple passengers, or unusual ticket rules are involved. For now, the emerging model looks less like AI replacing the corporate travel agent and more like AI deciding which problems do not require one.
Song_about_summer, Adobe StockSome companies are aiming for an even more proactive model. Navan announced TravelClaw in March 2026 as an upcoming agentic layer intended to monitor trips and contact travelers proactively rather than waiting for them to ask for help. The company said the planned system could use traveler preferences and trip context to provide information or help resolve issues through messaging, email, or Navan's own applications. Because Navan described the technology as an upcoming rollout, it is better understood as a sign of where the industry wants to go than as evidence that fully autonomous trip management is already universal. Still, the direction is noteworthy: corporate travel software is being designed to notice problems instead of merely responding when a traveler reports them.
Booking And Expensing Are Becoming The Same Workflow
Ask frequent business travelers what they dislike most about work trips and the answer is not always the airport. Administrative work after the trip remains a significant pain point. The 2026 State of Corporate Travel and Expense report from Skift and Navan found that 71% of surveyed travelers spent at least 30 minutes completing an expense report. The same research found that 76% trusted AI with straightforward travel and expense tasks, up from 59% in 2023. Those figures help explain why the race to reinvent booking is increasingly inseparable from the race to automate expenses.
The traditional corporate process creates a strange duplication of work. An employee selects a flight, hotel, or train inside one system, pays for parts of the trip through one or more payment methods, then later reconstructs those transactions inside an expense tool. Integrated platforms are trying to turn the booking itself into the beginning of the expense record. SAP Concur and Amex GBT describe Complete as a combined system for booking, servicing, payments, and expensing, while their Egencia integration can automatically move eligible booking and itinerary information into Concur Expense. Amex GBT said in its second-quarter update that the Egencia and Concur Expense integration was live for all customers.
Perk's August announcement takes the concept into conversational AI. The company said users can submit expenses from compatible AI assistants in addition to booking or managing trips. Managers can also interact with approval workflows through the AI layer, while company permissions and existing controls remain part of the process. That means a traveler could eventually handle searching, booking, canceling, and expense submission without thinking of them as four different software tasks. The experience starts to look like a continuous conversation about one trip rather than a collection of unrelated administrative transactions.
AI is also moving deeper into the approval side of the process. SAP Concur announced an AI-assisted approval manager in August that highlights higher-risk expense submissions and provides information intended to help approvers focus their attention. It also introduced an AI-assisted dashboard for employees who arrange travel and handle approvals or expenses for other people. These capabilities matter because the future of business travel automation is not simply about saving a traveler a few clicks. The larger goal is to connect the employee, travel manager, finance team, booking system, payment data, and expense process so fewer transactions have to be manually reconstructed later.
Personalization Is Getting Smarter, But So Is Oversight
AI also promises to solve one of corporate travel's oldest tensions. Employees want travel options that fit their preferences, while employers want travelers to use preferred suppliers, negotiated rates, and approved spending limits. Those goals do not always align when a booking tool presents hundreds of options with little understanding of what an individual traveler actually values. AI systems can narrow those choices using profile data, previous behavior, company policy, and travel context. KAYAK for Business says its conversational booking system considers traveler profiles and company policy when recommending itineraries, while SAP Concur and Amex GBT describe Complete as a system designed to personalize booking while incorporating policy-compliant options.
Personalization is also changing what travelers may be able to select. GBTA's research found that 51% of travel buyers believed attribute-based hotel shopping would improve the booking experience. Instead of simply choosing a room category and attempting to arrange details afterward, a traveler could potentially select features such as room location or other attributes during booking. Buyers also expressed interest in purchasing additions such as breakfast, parking, early check-in, and late check-out within the corporate booking flow. That matters because the price of a hotel room is not necessarily the full cost or full value of the stay.
The same technology that makes personalization possible also gives companies more visibility into employee behavior. SAP Concur announced leakage reporting for Complete in August 2026, designed to combine travel data from different sources and help managers identify bookings outside the managed program. GBTA's research illustrates why companies care about that issue. Only 12% of global travel buyers surveyed said they had a consolidated view of their travel program from a single data source, while 63% cited a lack of consolidated reporting as a major challenge. Fragmented data makes it harder to understand spending, negotiate with suppliers, support travelers, and evaluate whether the travel program is working as intended.
S Fanti, Adobe StockThis is where AI becomes more consequential than a better search interface. A sufficiently connected system can potentially know that a traveler prefers an aisle seat, must reach a meeting by a particular time, has access to a negotiated corporate hotel rate, requires managerial approval above a certain price, and should use a specific payment method. It can then apply those constraints before presenting a shortlist. That can reduce the cognitive load on the traveler, but it also means companies need to make careful decisions about which personal and corporate data AI systems may access. The future of personalized business travel therefore depends as much on permissions and governance as it does on the quality of the recommendation algorithm.
The Biggest Challenge Is Trust, Not Technology
Employees are not waiting for every corporate AI policy to be finished. SAP Concur's 2026 global survey found that 75% of business travelers were already using AI-powered tools to support elements of their trips. Seventy-two percent said they had used or would use AI tools that were not approved or supported by their employer to plan or book business travel. SAP Concur also found that 85% of CFOs were concerned about employees using unauthorized AI tools for business travel. The company surveyed 3,300 business travelers across 21 markets, along with 800 travel managers and 700 CFOs in separate surveys conducted in April 2026.
That creates a new version of an old corporate travel problem. Employees have always been tempted to leave managed booking channels when consumer tools appear easier, cheaper, or more flexible. AI raises the stakes because employees may now share trip details, company information, preferences, or other data with systems that their employer has not approved. SAP Concur argues that organizations need clear governance and trusted tools if they want to reduce this kind of "shadow AI." Its survey found that 30% of travelers who would use non-approved AI said they would do so because their company did not offer AI tools for the task.
Travel buyers are also signaling where they want humans to remain involved. GBTA found broad comfort with AI recommending negotiated flights and hotels, but substantially less comfort with allowing it to change or cancel bookings. The same study found that buyers evaluating travel-management-company partners assigned 54% of their consideration to technology and 46% to servicing, while 26% said servicing was the more important factor. That is hardly a vote for removing people from the process altogether. Instead, it points toward a model in which AI handles routine interactions while human specialists remain available for complicated, unusual, or high-stakes situations.
The most successful AI booking systems will therefore have to do something harder than simply generate an impressive itinerary. They will need to demonstrate where their travel inventory comes from, how company policy is applied, what information the AI is allowed to access, when approval is required, and what happens when the system cannot safely complete a request. They will also need clear escalation paths when a traveler needs a human. Those requirements may slow down the dream of completely autonomous business travel, but they are also what can make the technology genuinely useful inside a company. Corporate travelers generally need convenience, but their employers also need control, accountability, and an auditable record of what happened.
What makes 2026 different is that these questions are no longer hypothetical. Amex GBT is connecting business travel transactions with enterprise AI, KAYAK for Business is introducing conversational corporate booking, Perk is enabling travel actions through general-purpose AI assistants, Navan is building infrastructure for agentic travel workflows, and SAP Concur is embedding AI throughout booking, policy, servicing, and expense processes. At the same time, survey data shows that many companies are still early in adoption and are working out how much autonomy they actually want to give these systems. Business travel has not reached a world where an AI agent handles every trip from calendar invitation to final expense approval without human involvement. But the familiar booking portal is no longer the only front door, and that may turn out to be one of the biggest changes in corporate travel since booking moved online in the first place.
You May Also Like:
AI Travel Assistants Are Coming. Here’s What They Could Do During A Trip
Why Team Offsites Are Replacing Some Traditional Business Trips
