Cross-functional applications are where robots shine.
On May 23, 2017, the MIT Sloan School of Management hosted the 14th annual CIO Symposium: “The CIO Adventure: Now, Next and… Beyond.” The one-day event brought senior IT executives together to discuss key technologies, including IoT, AI, blockchain, Big Data, DevOps, cloud computing, and cybersecurity. The main idea was to help prepare these tech leaders for challenges they face, including shepherding ongoing digital transformations, building a digital organization, and managing IT talent.
This series highlights insightful sessions from the event.
How is business actually using artificial intelligence? Tom Davenport, the president’s distinguished professor of information technology and management at Babson College, cofounder of the International Institute for Analytics, a fellow of the MIT Initiative on the Digital Economy, and a senior advisor to Deloitte Analytics, discussed his research on this question, drawing on insights from his forthcoming book The Cognitive Company.
The research, based on analyses of 160 AI projects, identified three primary groups of business applications.
Robotics and Cognitive Automation
Forty-four percent of these projects center around the mostly robotic tasks AI can perform when data transfer occurs across multiple systems: changing addresses, replacing lost credit cards, and automated reporting in fields including wealth management.
Nearly half of the organizations in Davenport’s sample use AI to deliver what he calls cognitive insights, which occur when machine learning detects patterns. Technology identifies matches of similar data across databases. Applications include identifying credit or claims fraud in real time, or combining supplier data from disparate systems (an effort recently undertaken by GE).
Just shy of 10% of the 160 projects use AI for cognitive engagement, or interaction with customers and employees. Examples include virtual digital assistants and corporate HR or IT services. Davenport notes that this use of AI is the hardest to manage, as it relies on speech recognition, the accuracy of which is improving but does not yet match human capability.
To date, none of these AI applications have wiped out any jobs, but what the successful initiatives have in common is that they involve cross-functional teams and often combine functionality (such as blending robotic process automation and deep learning) to perform higher-level tasks, which, in Davenport’s estimation, may threaten jobs within the next 20 or so years.