Employers are creepily analyzing your emails and Slack chats to see if you’re happy

Who is reading your email?
Who is reading your email?
Image: Reuters/ Neil Hall
We may earn a commission from links on this page.

Big Brother isn’t just watching. He’s hanging out on Slack, too.

Most office workers know, on some level, that their work emails and messages aren’t private. They may not realize, however, the extent to which their communications are being analyzed and parsed for signs about employee happiness and satisfaction.

AIR, a Tokyo-based software company, is marketing software that scans conversations on workplace communication tool Slack to gauge team morale. Their product, called Vibe, looks for keywords and emoji, then sorts a team’s mood into five emotions: happiness, irritation, disapproval, disappointment, and stress. There’s even a bot that will notify managers of real-time changes in team morale.

Sentiment analysis, as it’s called, has been around for almost a decade, but its main commercial application has been scanning social media for clues about consumers behavior. IBM markets its AlchemyLanguage software, for example, as a tool marketers can use to sift through social media and produce reports on how products are received.

Those tools are now being directed at employee communications as corporations increasingly fret about each worker’s “engagement”—a nebulous measure of commitment and motivation—as they look to drive down costs, reduce staffing (paywall), and retain hard-to-please high-value employees. In a Harvard Business Review survey of 568 executives from large companies, employee engagement was ranked ahead of innovation or efficient productivity as factor in their businesses’ success.

No wonder, then, that software firms and consulting firms are pitching products and services designed to measure and improve engagement. Deloitte, for example, sells services that promise to diagnose failings in corporate culture and engagement through surveys, employee interviews and focus groups.

Of course, the accuracy of the analysis is only as good as the software performing it, and algorithms can be led astray by ambiguous language.  To a robot, “I hate my job” may look very similar to “I hated my last job.” The smaller the pool of comments being analyzed—such as in one team’s Slack channel—the more easily it can be skewed by rogue words out of context.

For workers not eager to be surveilled, it might be time to take your gripes offline and back to the water cooler.

Update: An earlier version of this article said Deloitte sells a service that scans employee emails to monitor sentiment. Deloitte has said an executive spoke in error when describing this product, which does not scan work emails.