[PODCAST] You’re Deploying AI Wrong and Here’s How to Fix It Pt. 1 w/ Rowan Trollope and Jonathan Rosenberg
The assumption that computers don’t make mistakes is completely false. Especially with AI, which is actually learning, there’s no question that computers are going to mess up. Hopefully not in a way that freaks out customers, though.
Being strategic about where you can allow inaccuracy is key.
BOT: “Sorry, I don’t understand that request.”
CUSTOMER: “Now I can’t trust anything this bot does! How do I go back like three steps?”
Whereas, if an AI assistant makes a dumb suggestion to the human employee working with a customer, the employee can just ignore it. The cost of inaccuracy in agent assistance is way lower than turning a bot loose directly on the customers.
“Hey, I've got new tech that shaves five minutes off of every agent call, but only half the time.”
“Hey, I’ve got new tech that misunderstands you 50% of the time.”
(Well, they’re actually the same tech.)
Rowan Trollope, our CEO at Five9, and Jonathan Rosenberg, CTO and Head of AI, came on this week’s episode of That’s Genius!, and here’s what they said:
Think about how you can strategically place AI to be helpful instead of just inaccurate.
To hear the full interview, click here.