top of page

AI and the travel sector - Part 2: What small and medium businesses need to do now

What’s the reality of AI (artificial intelligence) in the travel sector today – and how can businesses take full advantage of it? In the second and final part of this special blog series, data specialist Mark Bush, a Firebird Director, explains what MDs and owners need to do now.

If you’ve read the first part of this miniseries on the travel industry and AI (artificial intelligence), you’ll already have a good grasp of what the term “AI” encompasses, what it relies on, and what AI tools can generally do. In this article, I want to look more closely at the practicalities around embedding AI in travel for small and medium enterprises (SMEs) especially.

Modelling AI can do the heavy lifting in a business”

 There are two ways travel SMEs typically use AI at the moment: for generative reasons, i.e. to produce content (textual/visual/audio), in the vein of Chat GPT; and for modelling reasons, i.e. to recognise patterns, predict behaviour, and so on – such as the effective recommendation tools used by Netflix, Amazon and others.


To anyone reading this series who still thinks either function of AI is an issue for the future – not for now – let’s consider the ways in which some businesses are seeing the benefits today.


Generative AI has already been proving super useful for many companies in the sector and elsewhere – helping teams draft marketing campaigns, conduct basic research and so on. The possibilities are worth exploring if you haven’t yet: while your staff are unlikely to find these tools transformational in their daily work, they may still them useful for time-saving, and for playing with new ideas.


Alongside the “sexier,” headline-grabbing, generative AI options, there are also various brilliant types of modelling AI that a minority of SMEs are already turning to. This AI can do more of the heavy lifting in a business, and be used for accurate forecasting, understanding customer behaviour, predicting churn among clients – including high-value clients specifically – and much more.


Modelling AI largely functions by capturing data in-house (through website visits, tracking, social media, email campaigns, etc) and the possibilities it opens up are impressive. Knowing how to prepare and use this kind of AI, however, is not something as widely discussed in travel as it should be. 

“The sooner an SME makes the right changes, the bigger the lead it could take in the current market”

 The reason AI preparation is not as widely discussed as it should be? That’s probably because the prep sounds boring and dry. It is also difficult, and takes time to get right. As a result, the work is not often being done effectively – if it’s even being done at all. On the plus side, that means now is the ideal time to rise above your competitors in the data stakes. All a business needs is to be proactive: the sooner an SME makes the right changes, the bigger the lead it could take in the current market.


To my mind, the best metaphor for the process of embedding AI effectively is building a runway. You can’t simply take off and enjoy the benefits of the best AI as soon as the mood grabs you. Instead you need a solid route forward; plus the time and resources to get things ready in advance.


As mentioned in the first part of this series, AI would be nothing without data. This means we need the appropriate knowledge to feed our models, in order for AI to do what we want it to. The first step, always, is to capture and organise all the relevant data.


At present, only a small number of SMEs are doing this – and while those who followed my advice from last year on GA4 will have had a headstart in terms of gaining dependable data, very few of the companies I’ve encountered in travel are really focusing on how to leverage that data for the optimum results. 

“AI is neither IT or finance – but its own specialism entirely”

While most companies want to be "data-led" – by which I mean: easily able to access precise, real-time information about customers, campaigns, finances and the market – and therefore make better decisions for their business – the reality is that the vast majority of SMEs don’t have any data leadership or strategy in place.


I sympathise: it’s difficult for owners and MDs at the moment. Part of the issue lies in finding the right people and/or training to undertake that significant task. Typically, most businesses (both in travel and outside it) see AI as falling somewhere roughly between IT and finance, and so will delegate data management to one or both of those departments. In fact, AI is its own specialism entirely and we cannot expect either department to have the niche AI knowledge required to excel.


Taking the relevant steps to embed AI also requires an element of “change management,” i.e. preparing teams – including owners/MDs – to adapt to processes and approaches that may well be radically different to what came before. The costs this involves, and the cost of embedding AI tools themselves, are another deterrent. Many companies find it difficult to justify the expense – particularly on the back of a turbulent time post-pandemic. 

“Happily, many companies already have strong data and analytics at their disposal”

The good news is that most AI technology is widely available, and more affordable than you might expect. GA4, for instance, costs completely nothing – yet provides immensely powerful solutions to understand website performance from a user-engagement stand-point, understand how marketing and adverts perform, and how successfully leads are converting.


Another piece of good news: many travel companies already have strong data and analytics at their disposal. They just aren’t yet using them in the way they’re intended.


Some excellent AI models don’t even require any in-house data supply. For example, sentiment analysis uses publicly available information from third parties. Its purpose is to ascertain what people think of a product or service by scraping text from public review platforms like Tripadvisor, social media mentions and so on. In this way, businesses concerned about their reputation can obtain invaluable insights around what customers like and don’t like, as well as measure the impact of campaigns designed to improve trust and customer interactions.


The average person would, of course, need substantial training on how to use these tools and processes properly – learning to make the most of AI requires specialist knowledge. However, there are increasing numbers of skilled data scientists, data engineers and data analysts out there who can offer support – I’m just one of them.


Moreover, while it’s true that building the team you need to oversee a great data strategy can turn expensive, that also doesn’t have to be the case: in many instances, after working with an expert to establish the right foundations, an SME can easily internalise its AI functions. Once the team has been inducted into the essentials, it won’t be a huge leap for them to analyse their data in-house.

“Most owners and MDs will be aware of the concerns around AI”

The last thing I want to do before wrapping up this miniseries, is touch on some of the common concerns around AI – of which most owners and MDs will already be aware. The first is that the rise of AI might lead to the automation of just about every workplace process, and so prompt the loss of many once-specialist jobs. This will happen to some extent, I’m certain – though I’m also sure new jobs will emerge in response.


Of more pressing concern in the short-term are the ethics around AI bias. Since every model completely depends on the data it is fed, when that data is at all skewed, the model itself will produce skewed results. One practical example of this, which might impact travel businesses, could arise if/when recruitment companies use AI modelling to target the “best” candidates for hire.


Historic data might tell that recruitment company – perhaps accurately – that privileged white men have done very well in similar roles before. It’s easy to see how this societally-led bias could then lead to only those types of people being categorised as “best.” Companies could unwittingly inhibit diversity as a result – and, with that, usher in a lack of diverse thinking, life experiences and staff-led innovations in the workplace.

“Services will only get more evolved from here”

What we want to achieve through AI is not to become more deeply entrenched in past patterns, but to take flight into more expansive territory. As with anything, open-mindedness, curiosity, common sense, and a focus on the wider picture should set the travel sector in good stead.


We are still in the early days of AI, and the innovations and services will only get more evolved from here. Start your journey sooner rather than later, take time to get the fundamentals right, and your team and business are likely to go far.


Mark Bush is a Director of Firebird with over two decades of experience in the travel and leisure sector, in senior IT, MD and CEO positions. In 2022, he deepened his knowledge of data science with an MSc in Business Analytics from NYU Stern.


Mark supports sector leaders to implement powerful Business Analytics and martech solutions in B2B, B2C and eCommerce, and is an expert in innovations employing MS Power BI, Tableau, Google Cloud Platform, BigQuery, GA4, Salesforce, Marketing Cloud, as well as programming in R and Python.


Learn more, and contact Firebird, at



Also by Mark Bush:



88 views0 comments


Commenting has been turned off.
bottom of page