It takes about twelve years for a pharmaceutical firm to research, develop, test, and launch a product. Several firms, including Pfizer, Novartis, an… - Thomas H. Davenport

Every decade or so, the business world invents another term for how it extracts managerial and decision-making value from computerized data. In the 1970s the favored term was “decision support systems,” accurately reflecting the importance of a decision-centered approach to data analysis. In the early 80s, “executive information systems” was the preferred nomenclature, which addressed the use of these systems by senior managers. Later in that decade, emphasis shifted to the more technical-sounding “online analytical processing,” or OLAP. The 90s saw the rise of “business intelligence” as a descriptor. In the middle of 2000’s first decade, “analytics” began to come into favor, at least for the more statistical and mathematical forms of data analysis.

In an attempt to develop a tool to help organizations begin to identify these errors, the Institute for Healthcare Improvement (IHI) developed a global trigger tool (GTT).9 Briefly, the GTT provides a standard methodology for reviewing patient records for triggers, or indicators, of potential adverse events.

One large US bank, for example — an aggressive adopter of AI — has announced a $350 million investment in reskilling related to AI-related job changes, and the bank is being both predictive and granular about the initiative.10 It’s working with researchers from MIT and elsewhere to understand — based on a “suitability for machine learning” (SML) assessment — which skills and jobs are most likely to be replaced by AI.11 The SML analysis will help the bank plan for changes in those jobs and help workers gain the skills they need to succeed in their modified jobs or transition to new ones.