“Machine learning and AI is a horizontal enabling layer. It will empower and improve every business, every government organization, every philanthropy — basically there’s no institution in the world that cannot be improved with machine learning.”It’s not surprising that firms all over the world are developing plans on how to harness this powerful technology to transform their businesses. But with all the hype around AI, how can organizations determine which AI use cases will have the biggest impact on their business? How can they develop a machine learning strategy that generates the greatest return? A new report from the McKinsey Global Institute seeks to help companies with this very issue. Entitled “Notes from the AI Frontier: Insights from Hundreds of Use Cases”, the report provides an analysis of traditional analytics methods—as well as a subset of machine learning called deep learning—and how these technologies can be applied to over 400 use cases to help solve a variety of business problems. According to the authors, “the value of AI is not to be found in the models themselves, but in organizations’ abilities to harness them. Business leaders will need to prioritize and make careful choices about how, when and where to deploy them.” Some of the key findings from the study are as follows:
- 69% of the AI use cases in the research study could be enhanced by deep learning, beyond what can be accomplished through traditional analytics tools
- The use cases where AI will provide the most value vary by sector and depends on where value is derived in an organization. For example, a retail company should choose to apply AI to marketing and sales processes, whereas a manufacturing company should apply AI to their supply chain, logistics and manufacturing functions.
- Several hurdles that could hinder an organization’s ability to realize AI’s full potential include:
- Access to a large volume and variety of labeled training data
- Internal skills gaps and overall organizational readiness
- Regulatory concerns, including data privacy
