Cutting through the hype of artificial intelligence in corporate travel

Start with the problem, not the technology

It's tempting to jump straight into AI in corporate travel. Everyone else has it, your technology providers are probably talking about it, and maybe leadership is asking what you’re doing in that space. But as Grant explains. "Don't worry about AI, worry about the problem that needs to be solved first. Is it policy compliance or negotiating a hotel program? At the end of the day, AI is amazing, but without important business problems, AI is just a shiny toy." Think about where the pain points sit in your corporate travel program:

• If travellers are spending far too much time searching for hotels, the opportunity might be a simpler more aligned booking experience.
• If travel managers are spending hours pulling reports together and getting little out of it, the outcome might be a tool that analysis data into user friendly dashboards.
• If teams are dealing with non-stop questions, the outcome might be more consideration for traveller needs to be met or a self-service way to access program information.  

• If disruption is causing a lot of panic and stress, the outcome could be identifying problems earlier and helping your team act before the situation gets worse.

   

AI powered corporate travel bookings

The idea of booking travel through a chat interface has been a topic of conversation for years. Instead of navigating a bunch of fields and filters, and endless scrolling until you find the one, you tell the tool what you want — just like how you would a friend or family member — and it works out the rest. "In the future, travellers and bookers will expect to type and talk in simple terms, not use their mouse or trackpad, and the actions needed will be done by AI,” said Grant.

For corporate travel, that could mean saying, “I've got a meeting in Sydney Tuesday morning, I need to be there by nine, and I'd like to be home Wednesday night” rather than specifying every search parameter.

How trust is becoming an AI feature

Early exposure to AI tools left a mark on a lot of people. Grant explains. "AI adoption has been rocky. In the early days it was fun - it hallucinated a bit, and we all had a good laugh at some of the results. But the use of AI has reached enterprise scale now, and these early trust issues have endured.”

The travel industry has spent the past 18 months trying to recover this trust, and with good reason. If your chosen AI travel management tool gives you a slightly odd restaurant recommendation, you can probably shrug it off. But if it gets a fare rule wrong, misinterprets policy, recommends an unsuitable itinerary or fails to recognise a serious disruption, the consequences are different. “Trustworthy AI is about providing the context behind why a decision has been made,” explains Grant. 

The response from serious players in the space has been to build AI that is transparent about how it works. Rather than just surfacing a recommendation, tools are increasingly explaining the reasoning behind it. Why this fare? Because it matches the traveller's usual routing and falls within the approved cabin class for this route. Why this hotel? Because it's on your preferred supplier list and sits inside the rate cap for Sydney this week. FCM is exploring this directly in the Sam booking experience. When Sam surfaces travel options, the experience can provide context around why something has been recommended. 

“Underneath all of this, though, is the data quality and it being trustworthy too. If the information you're feeding AI is wrong or incomplete, you can't expect the output to be reliable.” Added Grant. This is one of the reasons AI in corporate travel has taken longer to mature than predicted. Getting the data right is a prerequisite.

If you're assessing AI in business travel, look for:

AI shouldn’t make humans disappear

The estimated time saved for travellers and arrangers over that fortnight was around 1,100 hours. But the most interesting number isn't necessarily the 97%. It's what happened to the other 3%. That 3% of interactions aren't a failure of Sam. They're an important part of the intelligence layer and we designed it that way. As Grant explains, "AI will get very good at the mechanics - the search, comparisons, paperwork. What it won't replace is judgment in the moments that matter. When a traveller's flight is cancelled at 2:00am and they're stressed, they don't want a perfectly reasoned rebooking option, they want someone to tell them it's going to be okay. That's not a data problem, and I don't think it ever will be."

This intelligence layer means various things for our consultants, but for corporate travel programs, that might mean less time pulling reports, answering questions, chasing basic info, and more time influencing, working with stakeholders and suppliers, improving policy, and planning for what’s next.

How to introduce AI into your program 

“We couldn't do this two years ago. We couldn't do this twelve months ago. We probably couldn't even do this six months ago. AI in corporate travel is still moving incredibly quickly.” Explains Grant. Rather than trying to predict the future, focus on what AI can improve today.