Why And How AI Projects Are Failing In Firms
As AI propagates its way through our economy, entering industry by industry leaving a trail of lower costs of production and, in some cases, unemployment (read my article about this here). Firms who can, but have not, began to use AI in their production process (“Laggards“), will see a decrease in demand for their products as AI adopting competitors "Front-runners" will see a fall in costs of production, and therefore can undercut the non-AI adopting firms.
This idea is summarised in the below quote by the Program VP of Artificial Intelligence Strategies at IDC, Ritu Jyoti:
The result of this has been an increase in firms attempts to use AI in their day to day operations, from banking to medicine to education, AI has found its place in increasing productivity and lowering costs.
The below chart shows the exponential growth that has/ will happen due to the expansion of AI into several industries worldwide:
However, attempting to incorporate AI does not guarantee success. Research from IDC latest report on AI, "Artificial Intelligence Global Adoption Trends and Strategies", concludes that half of AI projects fail for one in four companies on average. Understanding why the failure of AI initiatives occurs is crucial to help businesses achieve a higher success rate when striving to incorporate AI in their systems.
The report considers a lack of required skills and unrealistic expectations as the main reasons for AI project failures.
Of the 2473 organisations questioned, there were some interesting findings that could shed light on other reasons for the failure of AI projects. Some examples are only two-thirds of firms are in the process of establishing an ‘AI-first’ culture and even less see AI as a priority at half of all firms.
In conclusion, along with the increasing investment in AI systems has also come increased failures in implementation, this can have the negative consequences of decreasing other firms confidence in AI and lead to lower levels of firms investing in AI projects. This problem can be solved through government intervention, with the simplest method being a subsidy paid to firms that attempt AI projects. This could be accompanied by specific specialised guidance, or maybe even partnerships with certain universities in order to give the projects the highest chances of success therefor encouraging more projects in similar firms.