Executive Briefing: Organizational design for effective AI

Description

Executives in every business, vertical, and function are being asked to “innovate with AI,” but the barriers to successful adoption for most enterprises are organizational, not technical. Mariya Yao explains why effective AI requires not only technical talent but extended interdisciplinary coordination between teams, investments in retraining your workforces at all levels, and cultivation of an experimental, data-driven culture.

Whether an organization chooses to build or buy AI, there are five critical areas that are nearly always underinvested in:

1. Goal-driven AI strategy: AI is not a magical technology which, when plugged in, automatically yields enormous business benefit. In many cases, poorly designed automation hurts your business and workforce. Executives must carefully assess where (and where not) to apply AI in their operations.
2. Business-driven data collection: Having the right data is more important than collecting high volumes of data, but data in most enterprises is collected without the development of machine learning models in mind. Overcoming this requires that the business and product owners collaborate closely with AI technologists and data stewards.
3. Technical training for front-line employees and middle managers: Executives are often well educated on emerging technologies, but enterprises inevitably stumble when their employees who are tasked to adopt new software and methods are not sufficiently supported.
4. Culture of data-driven, probabilistic decision making: Business leadership historically relied on instinct and vision, but in our increasingly complex world, executives need to make more nuanced decisions that balance analytics with experience.
5. Organizational design: Roles, responsibilities, and workflows change dramatically when automation assumes increasingly more human tasks. Few organizations have spent the requisite time rethinking how to design, operate, and scale the automated enterprise.