Most of the stuff mentioned in the book we already know:
- how algorithms made possible high-frequency trading (HFT) in the stock markets
- how Google, Netflix, and dating sites make use of algorithms to make uncanny suggestions
- how music labels and movie producers predict pop hits and box-office blockbusters before they are released
- how algorithms are outperforming chess players, music composers, game show contestants, poker players, doctors, lawyers, pharmacists, radiologists, sports writers, etc.
There's one thing which is new to me though. When you call customer service, and the automated voice tells you that this phone call may be monitored or recorded for quality and training purposes, what do think is happening? I used to think that such calls are being recorded in case there is a disagreement, so management can go back to the recorded call and check who said what.
What is actually happening is that when you call, a bot listens in to the conversation. It is listening to your tone of voice, how you structure your sentences, what specific words you used, etc. It determines your "personality" within thirty seconds, and advises the call agent on how best to interact with you. Better still, the next time you call, the bot will be able to identify you, and route your call to an agent that best matches your personality traits for a faster resolution time.
The numbers don't lie. Data scientists took 1500 phone call recordings for a group of 12 customer service agents and crunched the numbers. When a caller gets paired with a agent with a matching personality, calls lasted for about 5 minutes with a 92% resolution rate. Matched with a "wrong" agent, calls lasted twice as long and the resolution rate drops to 47%.
With bots and their algorithms take over a lot of jobs (self-driving vehicles!), everyone should start thinking about getting into programming and computer science. Until the day bots start writing their own code.
No comments:
Post a Comment