I had a fascinating conversation with one of our non-Exec directors this morning about life in general, and for a point about AI in particular. He by day is the CFO of a travel business, a world that has become consumed by sophisticated pricing algorithms in the last 20 years.
As he pointed out, though, 20 years of prior history comes to nought when considering what pricing should be applied to UK travel accommodation in this year. If a brief trip up to the Lake District last month is anything to go by, the school holiday season this year is going to be complete carnage in the tourist areas of the country, and British families look to spend their holidays in Great Britain rather than overseas. How does traditional pricing history help you when you are modelling a pattern of sales behaviour that has never happened before?
The airline industry, again a realm where pricing algorithms to maximise yield have been fine tuned since the 1990s, must be facing similar challenges? Assumptions about how and when people travel have changed beyond recognition (and presumably won’t really ever “go back” to where they were before).
Moreover, what other areas are going to see algo-kookiness in the coming months and years as 18 months of outlying data merges into the historical trends? I’m surprised more people aren’t talking about this at the point we are with extrapolation-based AI technologies still being touted as the latest sliced bread.