McKinsey has released a new discussion paper called, “Notes from the AI frontier: Applications and value of deep learning” by Michael Chui, James Manyika, Mehdi Miremadi, Nicolaus Henke, Rita Chung, Pieter Nel, and Sankalp Malhotra.
McKinsey collated and analyzed more than 400 use cases across 19 industries and nine business functions. The goals was to identify specific sectors where deep neural networks can create the most value, the incremental lift that these neural networks can generate compared with traditional analytics, and the data requirements that must be met for this potential to be realized.
Of the 400 use cases McKinsey analyzed:
- 15% of use cases can capture full value without AI
- 16% need AI to capture value
- 69% of use cases could improve performance using AI
The majority of use cases analyzed, 69 percent, could improve performance using deep neural networks instead of other other analytics regimes. Neural networks were the only solution for adding value in 16 percent. The remaining 15 percent see limited performance gains from using artificial neural networks – often because of data limitations that made these cases unsuitable for deep learning.
Read many more words on McKinsey.com, Notes from the AI frontier: Applications and value of deep learning.