Exposure to AI has become ubiquitous, but it is far less common as a driver of investor value. In 2022 over 1.6bn people engaged with AI on a daily basis but, even though use in business processes is growing and there are now several AI unicorns, in 2019 it was estimated that 40% of European ‘AI start-ups’ didn’t actually use AI/ML in a way that was material to the investment.

Microsoft defines AI as ‘a way for a computer to think like a human and perform tasks on its own’, and Machine Learning as a way for a computer system to develop intelligence. By that definition, the vast majority of commercial data science, from predicting health to identifying credit card fraud, is actually ML.

There are three reasons why it is difficult to exploit AI and ML commercially. First is scalability: it is often impossible to acquire the volume of data needed to scale beyond the proof of concept, either in totality or within the projected time lines of the investment thesis.
The second barrier is cost: even with intelligent infrastructure that scales on demand, operating AI and ML at scale can be prohibitively expensive.

The third difficulty is organisational complexity: development requires innovation but careful governance, and investment without immediate ROI.