The majority of businesses, from hotels to retail stores, have large-scale language models in place, with top use cases in search and customer support. However, these models are not perfect.
Consumers are becoming frustrated with AI chatbots that fail to answer their queries, and their patience is running thin.
Some consumers simply give up, according to Parloa research released Monday. 3 out of 5 people repeat auto-support once before abandoning it.
These friction points can be costly. Here are three numbers about the cost to customers of AI failure:
50%
A Kim.cc survey of 1,000 U.S. consumers released last week revealed the percentage of consumers who blame a company’s leadership when AI support fails to answer customer questions.
Consumers will not relent when faced with AI chatbots and tools that fail to resolve their inquiries.
“It’s important for leaders to understand that customers don’t see poor AI support as a technology problem; they see it as a company problem,” Sachin Jaiwal, co-founder and CEO of Kim.cc, said in an email to CX Dive.
“If a customer gives a wrong answer, gets stuck, or feels like the company is hiding behind a bot to save money, the loss is much greater than one wrong ticket,” Jaiswal says. “That can lead to loss of trust, loss of brand reputation and ultimately loss of customers.”
According to a Kim.cc study, more than a third of respondents reported that customer support becomes easily irritated when it is enhanced by AI, which may be because the friction caused by AI is so common.
91%
The accuracy of Google’s AI models, according to an analysis by AI startup Oumi published in the New York Times earlier this spring.
Google processes over 5 trillion searches a year. This means that we provide billions of incorrect answers per year, with an accuracy rate of 91%. Recently, one of those wrong answers made headlines. Earlier this month, a German district court issued a temporary restraining order against Google over false information generated by Google’s AI Overview feature. Two companies had filed suit alleging that certain searches led to Google’s AI overview falsely linking companies to fraud and questionable business practices. The ruling stated that the false and defamatory statements generated by the company’s AI overview feature were its own content and therefore would be liable. Google is contesting the ruling, but the tech giant is now expected to pay 80% of the legal costs. If the ruling stands, it would open the door to further lawsuits and defamation lawsuits against technology companies that rely on generative AI.
$5,000
The amount of money it cost a car dealer when a fraudulent AI chatbot claimed the car was worth more than it actually was.
When a Toronto man decided he wanted to sell his car, he contacted the BMW dealership where he bought it, CBC News reported last month. He received a text message from “Quinn” offering him a buyback of CAD 27,162.79 ($19,119). It turns out that Quinn is an AI chatbot and the offer was made in error. The car dealer declined the offer until CBC News reached out to BMW Toronto for comment. The dealership reinstated the original offer and paid the man about CAD 7,000 (about $5,000) more than what the dealership thought the car was actually worth.
