Donna Rufo
BTN's survey showed most companies remain in the testing in exploring stage in their deployment of AI in travel management, but some organizations have advanced to more comprehensive utilization. One is IT services and consulting firm EPAM Systems, which has embedded AI tools into employee workflows to guide travelers' choices, surface policy requirements and reduce reliance on manual intervention.
"The goal is to use AI to create a better experience for our employees and at the same time deliver measurable value and efficiencies for our business," said EPAM head of global travel, expense, credit card and remote work Donna Rufo.
The company's first foray into an AI-powered travel application was a chatbot in the Microsoft Teams platform named Grace Helper, designed to offer employees instant answers to travel questions at any time of the day. Grace went live after a three-month beta test in which "we just asked tons of questions to her, and everything she got wrong, we would go back to our knowledge base or intranet page and change how we phrase our information so she's able to pick up the right answers," Rufo said.
The chatbot now "saves about 120 minutes per employee per month because we do measure our results," Rufo said.
Rufo's team internally has developed AI-powered employee-facing calculators and guided workflows around trip planning, visa requirements and other administrative tasks to save time and reduce friction.
The team has embedded an AI-powered calculator in its pre-trip travel request form, which requires a spending estimate. The AI sources travel prices on the open market to provide such estimates, saving what Rufo said was 70 percent of the would-be traveler's time.
The company's booking tool will pop up policy reminders at the time of booking if the planned purchase appears noncompliant, she said. It won't halt the booking—as EPAM provides consulting services, some travel is client-billed.
The team also has integrated AI into the company's expense and payment processes. EPAM policy requires travelers report expenses monthly on their company-paid corporate card; when they do not, AI tools will recognize it, cap their card spending limit at one dollar, then revert it once expenses are received, Rufo said. It also flags noncompliant or incomplete expenses.
Most EPAM AI tools are developed internally, with select third-party purchases that are heavily customized, Rufo said.
In the company's road map for the next 12 months is "an end-to-end orchestration where AI can connect travel, expense, HR and finance systems to trigger actions automatically like the card controls, like approvals, like reimbursements," she said. "Where all the different departments that touch a trip [are involved], that's my end-to-end goal."