It has been just about eight weeks since ChatGPT was launched, and rarely have we seen a new software-based service make such a dramatic entry into global consumer and work environments.

I am not going to bore you with a long list of brilliant examples of what ChatGPT can be used for as there already are countless examples of this all over social media and internet sites. This article is focused on identifying some of the potential areas for usage of ChatGPT and similar next-gen AI platforms inside our beloved travel industry—and hopefully with a few unexpected twists and examples which you haven’t already thought about.

The obvious case for AI in travel – For the past two years we have seen stories about the global shortage of people to work in the travel industry combined with the hard fact that a large number of current employees are heading towards retirement and will be leaving the workplace in the next few years.

The combined effect of these two mega trends is obvious—the travel industry is running out of people fast, and we need to come with a plan B very quickly. Even if we manage to convince a new generation of younger employees to join us, we will lose them the moment they realize how old and outdated current systems are and how inefficient the overall customer service experience delivery platform is.

This is where AI comes in. What if we could automate 40 to 50 percent of all the transactions performed by human agents today? I’m thinking of the most manual and repetitive tasks such as reading

an incoming email and providing a written response or giving advice to a traveler going somewhere for the first time. The ability to make this happen now that ChatGPT has arrived is suddenly very real, and this seems to be a perfect example of the right technology at the right time.

The unexpected opportunity for AI in travel – With the obvious case out of the way let us sharpen the pen and be more creative. How about starting to apply AI in the area of travel supplier contract management?

Before you stop reading, consider the following question: How much time would you like to spend monitoring your supplier agreements on a weekly basis to ensure you get the best value from these contracts and identify any areas of underperformance as early as possible? While your answer might be a few hours per week, the reality probably is that you simply don’t have the time to monitor supplier contract performance on a weekly or even monthly basis, so either you are only doing it once or twice per year, or you have to resort to using an external paid consultant.

In either case, you are likely losing significant money unless you pay a small fortune in consulting fees, which end up reducing the overall value of your deals anyway. Imagine if you could activate an AI bot designed to scan all the bookings constantly and identify any supplier contracts that appear to be underperforming—either because the transactions are made with incorrect prices or because travelers make bookings with the wrong suppliers. An AI bot could literally provide you with an updated summary with recommended action every morning while you sip your first latte.

The dynamic travel policy enabled by AI – Until today, corporate travel managers around the world have followed a best practice where 80 percent of travel spend (typically the top 10 destinations) is managed tightly with complex and multi-layered travel policies often involving pre-trip approval workflows and detailed expense reporting procedures.

Imagine if this was replaced by a “continuous travel policy engine” powered by AI designed to analyze every single booking against the overall “wisdom of the crowd” in the form of a global data lake with millions of bookings.

An AI policy engine could immediately compare the latest bookings and determine if the prices paid and choices made by the traveler fall inside a box of “acceptable choices” and therefore can be approved for processing, while any booking falling outside the box will result in a fully automated alternative option being presented back to the traveler along with information about why the booking was declined.

The role of the corporate travel manager in this model will be to define the rules for the AI policy engine including the tolerance levels for deviation from the norm. Rather than having a policy that focuses on class of travel or preferred supplier, the AI engine could look at whether the chosen option is acceptable compared with the industry wide average either for similarly sized companies, for companies in a similar industry vertical or something else.

The virtual travel manager powered by AI – While the dynamic travel policy opportunity no doubt is intriguing for travel managers, the final option is the scary one - what if we can deploy a virtual travel manager designed to not only optimize supplier contracts and manage travel policy but literally capable of running the entire travel program?

For years we have talked about how new technology eventually will eliminate the role of the TMC, but perhaps the real threat comes from another angle—a virtual travel manager will be capable of performing the role 24/7 and will most likely deliver better financial results than a human controlled travel program.

And don’t forget how a virtual travel manager can mine the data lake constantly and identify the smallest opportunity for a qualified saving (say 100+ USD) in the form of flying 1 hour earlier or later, staying at an alternative hotel or moving the meeting somewhere else. No human travel manager is capable of doing this at scale, so the “AI Travel Manager” could quickly become a new favorite of the CFO.

But wait, I hear you say, what about the importance of the human touch in a travel program? Well, as nice as it sounds, I don’t think the average traveler is feeling the human touch on most of their trips. They typically book and manage the trip themselves and only call or email the corporate travel manager when there is a major disruption or a very serious duty of care incident.

Having a virtual travel manager capable not only of identifying major problems in advance, but also taking preventive actions and arranging large volume re-bookings actually sounds very attractive.

And finally, by having an AI-powered travel manager in place, corporations will be able to test new technology and services out much more often and at a lower cost. The only major requirement will be to train the bot on the new dataset in order to enable logical and sensible decision making.

A Step Too Far Or Moving Too Slow?

I will leave you with this big glaring question: Just because we now have technology capable of providing services like the above mentioned, does that mean we should deploy these solutions in order to improve overall financial efficiency? Or should we perhaps accept that travel is a human-centric service that creates incredible value in the form of human contact before, during, and after the trip and therefore needs to be managed and delivered by people with all the potential inefficiencies this brings?

Regardless of where you land on this question, it is very likely we will see AI becoming a fully integrated part of the human agent desktop and workplace—with particular focus on automating repetitive low value tasks and providing relevant recommendations for faster decision making to the travelers.

I guess the answer depends on your mindset, personality and culture. I, for one, have realized that I prefer some human contact as part of my overall travel experience. If nothing else, it gives me someone relevant to share my own travel adventures with—without using technology to do it!