New Worker Woe: AI Impatience

When it comes to AI, if at first you don’t succeed, it appears the next best step is to give up and go back to the old way.  

Despite pouring money and resources into artificial intelligence, leaders are finding that the results they’d hoped for continue to elude them. And here’s one reason why: In a new report, a shocking nine out of 10 AI users admit to abandoning the technology and going back to a non-AI method because it was easier or provided a better outcome. In part, this is because people expect instant gratification from technology, and become impatient when they don’t get it. The typical AI user expects a satisfactory answer within two prompts, according to the report, and will walk away after a fourth unsuccessful one. “People expect the first output from AI to be great, and the reality is that it usually isn’t,” says Shanda Mints, vice president of AI strategy and transformation at Korn Ferry. 

One major reason people aren’t getting the results they want: They don’t take the time up-front to refine and optimize prompts. The average user in the study earned a “C” grade for the quality of their prompting, scoring 57 out of 100. Their mistakes included not specifying the tone or style desired, not asking for assertions to be challenged or justified, and not identifying a target audience, among others. “People don’t realize how much time and information they need to put into a prompt to get the output they want,” says Mints.  

That’s where corporate culture and a “fail to learn” mindset come into play, says Bryan Ackermann, head of AI strategy and transformation at Korn Ferry. “Working with AI effectively is a process,” he explains. To be sure, people can get better outcomes quickly by asking AI itself to review and improve their prompts, but less than half of users in the study did so. Moreover, even if the fault lies with the AI tool, not the prompt, improvements and advances are happening so rapidly that yesterday’s bad prompt may succeed today.  

Experts say there are ways to reduce AI frustration, such as taking the few things that have worked and sharing and replicating them widely enough across the enterprise that the payoff eventually outweighs the time expended on it. Instead of abandoning tasks that don’t provide the desired results, Mints suggests creating a “not yet” list and trying again every few weeks or months. “In practice, you may try ten things and only get real value from two of them,” says Mints. “The difference is what happens next.” 

Learn more about Korn Ferry’s Organizational Strategy capabilities. 

Read more This Week in Leadership articles