The International Institute for Analytics (IIA), a research firm based in Portland, Ore., recently discussed ten predictions for 2015. Some were conventional — Prediction #7: Hadoop will go mainstream. Some were thought-provoking — Prediction #2: Storytelling will be the hot new skill in analytics. Should CIOs consider hiring journalists to do that job?
And one stood out because it seemed, well, ominous — Prediction #9: Analytics, machine learning, cognitive computing will increasingly take over the jobs of knowledge workers. Tom Davenport, co-founder of the IIA, professor of management and information technology at Babson College and analytics thought leader, said — and has been saying for years now — that business leaders need to be preparing for this now. They should consider how to “prepare knowledge workers to augment the work of smart machines rather than be automated by them,” he said.
Automation is already happening. Journalists, lawyers and even teachers are standing by while parts of their job descriptions are being taken over by things like predictive coding, knowledge-based curriculum design or automated earnings reports. While the technology is “still quite fragmented,” Davenport said during the IIA 2015 predictions webinar, “there’s probably not a knowledge worker problem out there that can’t be addressed by some system.”
There are benefits to the advancing tech. In many cases, as fellow IIA faculty member Robert Morison pointed out, “what we’re doing is better equipping people, and if we could do that at scale, it could make an enormous difference,” he said.
Jeremy TerBush, vice president of analytics at Wyndham Worldwide Corp. explained in the call that his team developed internal pricing systems that rely on algorithms to project tomorrow’s vacation rental prices. The cognitive computing program has not had an impact on the company’s workforce. “We’ve seen it hasn’t automated away any jobs,” he said. “It’s just allowed us to be more focused on us managing our inventory better.”
The system works about 80% of the time. “But 20% of the time, the prices are overridden by our revenue management team, who is closer to the market and picks up on things the algorithms are missing,” he said.
Automation can provide efficiency, help businesses make better decisions and save on costs. But (cue the sounds of dismay) there is the other side of the coin businesses may not be considering: What are they at risk for losing? Will automation simply deepen the divide between haves and have-nots?
Said Davenport: “I suspect the people who you need to do that are your most experienced and expert pricing analysts — and not the ones fresh out of school. Because, as we were saying, oftentimes the entry level work can be done by computers, it’s the hard cases humans need to override or augment.”
The question is, said Morison, “How does someone become an experienced pricer when all of entry level work is done by machines? Who learns to be the experienced expert?”