Payment companies are looking for markets that are ripe for artificial intelligence technology, and one sector where AI payments are taking off is corporate travel.
The payments company Brex is using generative AI in an attempt to simplify corporate employee expense reporting and transactions. The company is betting the mix of booking trips, reserving hotels, scheduling ride shares, and filing expenses are a fit for large language models and other elements of generative AI, which uses machine learning to produce original content, answer questions or obey textual commands.
Brex is also in the midst of a strategy to expand corporate financial services, building on a boost in deposits it got when Silicon Valley Bank failed. And like most payment companies, Brex is interested in finding out if generative AI can cut costs and time from transaction processing and other internal and client-facing activities.
“Our vision is to give everyone a personal assistant,” said Cosmin Nicolaescu, chief technology officer at Brex. “An assistant who books trips, handles expenses and writes memos.”
The product, called Brex Assistant, uses “autonomous AI agents” that access data from employee and corporate calendars, internal organizational content, the employee’s past expenses and other information about a trip to automatically process corporate transactions in real time.
AI travel companion
Brex’s Assistant is designed to know when a person is traveling for work based on a range of data points. When the staffer books a flight, the product assigns the trip to the appropriate travel budget, digitally fills out the business memo and generates an itemized receipt. The expense is also automatically and proactively checked against the company’s policies to ensure compliance before booking.
“A lot of these travel policies are clunky and hard to process,” Nicolaescu said. “With generative AI, you don’t have to be super specific when making a request or asking for information. It can deduce what you’re asking.”
The “AI agents” are programmed, or “machine taught,” to answer questions such as: “How much am I allowed to spend per day at the off site?” or “What restrictions apply for this trip?” Employees can add prompts such as “Route all charges from DoorDash to my work-from-home meal stipend.” Employees can also enter queries such as, “What trip was I attending when I expensed this Uber ride?”
Brex, which sells lending, payments and other business services, has been testing generative AI internally for the past few months ahead of the Brex Assistant release. It did not say which clients use Brex Assistant.
The Brex Assistant will be embedded into Brex’s Empower spend management platform. Brex Empower launched in April 2022. as the primary suite of spend management tools from Brex. Brex Travel was built on top of the Empower platform, and and that was launched as a new portion of the Empower platform in March 2023.
Like other digital payment firms such as Klarna, Brex envisions generative AI as a decentralized technology, given the natural language tool’s potential to be more understandable and accessible to a non-technical audience.
“Large language models are an opportunity for everyone to use AI,” said Nicolaescu, who worked as an engineer on Stripe’s international expansion before joining Brex in 2018. “Not everyone could use earlier versions of AI.”
Travel is one of the largest applications of Brex Assistant, which also includes general expense management functionality. It will continue to expand with new functionalities for employees and finance teams.
“The opportunity to use generative AI is very large,” Nicolaescu said.
A travel AI boom
While OpenAI grabbed early headlines when its ChatGPT generative AI program debuted in late 2022, other firms such as Amazon and Google are quickly developing similar generative AI and large language models.
Like Brex, several financial institutions are also applying generative AI to travel. U.S. Bancorp launched a travel payments and management platform this year for middle market companies, or firms with $159 million or less in yearly revenue, the same general category of businesses that Brex is approaching. U.S. Bank is using AI from TravelBank, a fintech U.S. Bank acquired in 2021.
Travel fintech Navan, which partners with Visa and Mastercard to compete with travel payment startups, deployed generative AI in recent months to expand services for corporate travel users. And outside of generative AI, travel technology firm Xeni has added stablecoins, another new payment option that is seeking use cases, to power its outreach to a broader range of clients.
Beyond its complexity, corporate travel poses an opportunity for new payment technology because of its popularity among smaller businesses that require third parties to manage payment processing. Sixty-eight percent of companies with revenue below $1 billion expanded their travel budgets in 2023, according to Morgan Stanley, noting that this trend has held up despite inflation and concerns over an economic slowdown.
AI is very well suited to simplifying corporate spending and related processes, according to Gilles Ubaghs, a strategic advisor at Datos Insights.
“Expense management in general has been a huge competitive front for fintechs the last few years and that is essentially due to how manual and inefficient these processes normally are for businesses,” Ubaghs said. “Excel-based or other manual processes aren’t just annoying; they’re a resource sink that limits visibility or control over expense spending.”
AI can generate efficiencies in terms of the expense reporting itself, a much better user experience, and add more control and visibility over spending in real time, Ubaghs said.
“Given the high levels of competition in expense management, and the ongoing return to business travel, I think these sorts of capabilities should find a receptive audience — and I’d be surprised if we don’t see a lot of similar developments,” Ubaghs said.
