The evolutionary role of AI on business trips

If AI teaches us anything, it means that the pace of change is increasing.

Returning the watch to the 2024 Business Travel Show Europe, sector executives told Phocuswire that they felt positive about the ability of artificial intelligence (AI) to improve travelers' experience and improve efficiency and productivity behind the scenes.

Since then, we have seen torrents of AI-related announcements and launches both in the wider world and in the business travel sector.

The world's first comprehensive AI law, the European Union's AI law, was created and entered into force. The Act outlines rules regarding the use of AI based on risk levels and prohibitions of vulnerable groups through AI, social scoring and biometric and facial recognition.

A number of new models have been released on tentatively, including Openai's GPT-4.1, O3 and O4-Mini, Google's Gemini 2.0 and 2.5, and Anthropic's Claude Opus and Sonnet 4. The launch that caused the biggest wave was the launch of China's Deepseek. This suggests that a low-cost model that is not entirely dependent on Nvidia's GPU's computing power is possible.

It was a busy year for AI too.

A recent Serko/Saber survey on AI in corporate travel revealed that 44% of corporate travel managers believe that AI will “having a major impact on the program” over the next five years, while 22% believe it will be “transformative and will restructure the industry at its core.”

The change has already happened

Last July, Altour unveiled a suite of AI-powered tools, including bookings, travel disruption management, and a natural language interface for travel managers.

Amex GBT, which launched its dedicated AI initiative in early 2024, recently announced AI enhancements to its Egencia platform. We have strengthened the virtual agent that was released in 2020 with new AI-equipped features, and added natural language query capabilities to Egencia Analytics Studio as a beta version. The full release is expected in 2026, and travel managers can access travel program data by asking questions in plain language.

Meanwhile, in early May, HRS launched an AI-powered Copilot platform. It combines the leading language model of humanity (LLM) with HRS Labs' proprietary specialist language model to enable travel and procurement managers to manage, interrogate and optimize hotel programs.

HRS has used AI and machine learning capabilities for many years to manage and optimize accommodation and conference programs for clients.

“While many people were talking about their booking experience, I thought the biggest advantage in a corporate travel space was in managing the program,” said Martin Biermann, chief product officer at HRS.

“We thought about ways to innovate the whole approach. What is the point of doing this program's optimization exercise once a year when the world is changing so fast? It's no longer a savings. It's sustainability, satisfaction, and safety.

Biermann believes that the true power of AI comes from enhancing program adoption rather than optimizing the program itself.

“While we can optimize our programs with AI, a big lever from a company's perspective is to encourage adoption into the program. This requires travelers to adjust the program to what they ultimately need. But that's a multidimensional issue.

The HRS Analytics Data Model examines various markets, supplier structure, customer business units, travel personas, employee profiles, spending behaviors and loyalty.

This model is optimized in three dimensions: reducing unmanaged spending, saving time for procurement or travel managers, maximizing compliance with sustainability, accuracy of rate availability, price correctness, and supplier compliance.

Companies are still concerned about AI risks

According to Biermann, data security and privacy bring the biggest questions.

“We have made sure our models are trained in our own infrastructure and our own cloud environment. We are not sending data client data anywhere. Action logs and inference are completely transparent for the company to provide transparency,” he said.

Keesup Choe, CEO of PredictX, recalls that the initial survey on AI was two pages long.

“The latest is books. So if you're a vendor who wants to offer these models for your business trip, the biggest hurdle is now getting IT security approval.”

Some companies want AI, but aren't sure how they'll actually roll out, according to Sophie Taylor, chief technology officer of the Graydoz Group.

Travel Management Company (TMC) holds regular innovation meetings, including senior executives and customers.

“All customers say, 'I want AI'. When I ask what AI means, the room is fatally silent,” Taylor said.

Cho agrees. “Travel managers and end users aren't using AI enough for their jobs for all sorts of reasons. Vendors can guide vendors on what to build. On the other hand, vendors honestly haven't worked long enough on the field so for both reasons it's a bit slower.”

There are also voices in companies calling for the rapid adoption of AI.

“Despite this appetite for AI from travel managers and bookers who believe AI will enhance its products and services, their information and security team wants to be extremely advantageous here,” Taylor said.

There are also very general concerns, particularly regarding accuracy and the answer to “hastisation.”

“Even a hallucination rate of 1% is unacceptable. There are many strategies to reduce hallucination rates to virtually zero. I think this can reach a fairly perfect 99.9999%, but that's a lot of work and special expertise needed to achieve this,” Cho said.

But according to Cho, development on the horizon is “huge.”

“There's a new set of new models that will be released. I recently played one, which is actually different from the current transformer models,” he said. “For example, a diffusion model for creating images comes from ether and becomes sharper and sharper. Instead of generating linearly from start to finish, everything comes at once.

He also believes that AI giant Google will make waves.

“I'm shocked, especially when Google doesn't announce the Gemini agent on mobile phones, especially when Google does exist on Google's flights. Do you need a TMC?”

Despite being bullish on AI, Gray Dawes' Taylor doesn't see it as an existential threat to TMC.

“Because the TMC is not when things get right, but when things don't go well,” Taylor said.

“We've closed the Heathrow Shutdown, not to mention Covid, hurricanes, volcanoes. All of these things when we need a TMC where AI doesn't work. That's what TMC is like.”