Said the Algorithm to the Human: "You Complete Me"
Way back in 2012, in the same issue as the now infamous article declaring “Data Scientist” the sexiest job of the 21st century, Harvard Business Review defined Big Data as a management revolution. The technical challenges of Enterprise AI (EAI) adoption are formidable, but even the most sophisticated Machine Learning (ML) project will die on the vine without a the right ecosystem supporting it.
A successful EAI adoption plan prioritizes the cultural and technical challenges equally. For some companies, long-held roles and responsibilities change, and fundamental ways of doing business literally turn upside down. Every single employee is impacted in one way or another, starting with the C-suite. Three critical characteristics of the business culture are required for success: data-driven decision making, a learning culture, and unprecedented levels of teamwork.
Data-driven organizations instinctively ask “What do we know?” before asking, “What do we think?” Senior leaders accustomed to making decisions based on personal experience and opinion may struggle to relinquish control, but leadership based on the “highest paid person’s opinion” is the opposite of data-driven decision making. Change comes only when business leaders are able to let data outweigh their intuition. Full integration of EAI sometimes requires organizations to empower employees further down in the organization to make decisions that have historically been made by upper management.
Learning Culture
Data-driven organizations encourage heaps of experimentation, all the time. By definition, machine learning models are learning continuously, constantly improving and changing. Insights often come from surprising sources and orthogonal data. Data science teams need the freedom and support to learn by doing: running experiments, and acquiring and creating new data sets. Visible and enthusiastic support of ML projects with both financial and behavioral backing recognizes that failure is a necessary part of the learning process. There is no faster way to shut down enthusiasm than to deny resources or ignore findings that challenge the status quo.
Unprecedented Levels of Teamwork
Data science is team sport that requires unprecedented levels of cooperation and collaboration throughout an organization. Turf battles and cross-functional data silos are among the most common roadblocks to successful implementation of EAI. Companies that address these potential impediments early in the planning stages – especially as roles and responsibilities shift – set a precedent of teamwork as a priority. Machine learning projects require participation from the highest levels of management to the final end-users, and all layers in between: middle managers, business process experts and data scientists.
The business community is on the brink of the single most exciting period in human history. EAI preparation and adoption doesn’t have to be a grueling exercise in change management and damage control. With the right planning and leadership, the transition can be leveraged to reenergize and invigorate an organization, even to reinvent a business.
Next Up: Is Your Organization Data Driven?