Steve Endacott believes the big calls should always be left to humans
At first glance, the idea of chief executives being replaced by artificial intelligence in businesses may seem far-fetched, but recent experimental studies have shown that it’s not beyond the bounds of possibility.
To explore AI’s potential as a chief executive, a real-world experiment was conducted from February to July 2024. Involving 344 participants and GPT-4o, a large language model created by OpenAI, this experiment simulated the decision-making challenges chief executives face in a gamified environment and used various metrics to track the quality of their choices.
Participants were tasked with navigating a digital twin of the US. automotive industry, making corporate strategy decisions over several simulated fiscal years. Their performance was tracked through key metrics such as market share, profitability, and board approval. They incorporated mathematical models based on real-world data from car sales, market trends, historical pricing strategies, and price elasticity, alongside broader influences such as economic conditions and the impact of the Covid-19 pandemic. The top two student and executive participants were then compared against GPT-4o, with surprising results.
How AI Fared in the role
GPT-4o excelled in many aspects, outperforming human participants on nearly every metric. It designed cost-effective products, responded effectively to market signals, maximising appeal while keeping tight cost controls, and gained a solid competitive edge.
How would this work in travel?
You can easily imagine AI doing an excellent job of identifying swings in consumer demand created by currency changes, such as a pound collapse against the Euro boosting the popularity of Turkey, which is contracted in sterling or dollars.
It may even be better at monitoring the performance of other European source markets, such as Germany, which use the same bed stock and have a large impact on European hotel pricing, as it can understand 42 different languages and ingest huge amounts of data.
Finally, it will undoubtedly be able to manage pricing yield management and flight capacity at a much finer level of detail than many current chief executives.
But what about the big calls?
The travel sector’s vulnerability to global incidents and market disruptions, exemplified by events like the September 11 attacks or Tunisian beach shootings, necessitates rapid, judicious responses. These decisions must prioritise human safety while mitigating potential brand damage stemming from public perception of the company’s actions. I very much doubt AI would get these big calls right.
Prioritising short-term gains without considering long-term consequences on strategic hotel partnerships can jeopardise future access to crucial inventory.
The unprecedented impact of Covid-19 on the travel industry demonstrated the value of human leadership, as more jobs and companies survived than initially anticipated, largely due to thoughtful decision-making by industry executives.
What about a hybrid?
Just as in politics, AI could have a significant supporting role, allowing 24/7 365 communication with voters; AI will have a major role in supporting chiefs in travel businesses.
As my recent experience with AI-driven fantasy football team selection revealed, purely logical and risk-averse decision-making often results in mid-table mediocracy, leaving me trailing behind more daring competitors willing to take more risks.
AI may be great for avoiding the last position, but it’s currently unlikely to give you a market-leading position when pitted against innovative leaders willing to back their gut and take risks.
Fortunately, this is a trait that most of the leading travel chief executives demonstrate in spades, so in my opinion, AI is best-suited to be an executive assistant summarising the data and leaving the big calls to their human masters.