Good morning!

Accenture chief leadership and human resources officer Angela Beatty had a problem. 

In an effort to get employees to provide each other with regular feedback, the company set up an automatic tool to nudge workers to leave their peers with some thoughts on how they did after a big project or meeting. While that certainly led to more responses, the feedback itself was often lackluster. It was clear that workers were more focused on getting the task done than taking the time to provide valuable constructive criticism. 

“There was a really decent uptake in the feedback, and it was timely and quick because it was part of the workflow,” she says. “But what we saw was the quality took a bit of a dip because people gave sort of more general feedback that wasn’t all that helpful.”

That’s why last year, her team started experimenting with ways to incorporate generative AI into the employee feedback process, with the hopes that it could help them produce specific feedback that is still timely, but more elaborate and constructive. In the roughly nine months since Accenture rolled out the tool, it has been used by employees more than three million times, increased their amount of feedback by 89%, and around 95% of workers say it’s helped them reduce the time it takes to think of and write out the commentary. 

The tool, which the company calls its “feedback coach,” is embedded into Microsoft Teams and Workday. After a project is completed, for example, employees are prompted to provide each other with immediate assessments, and then rate the other person’s performance on a scale from “great work” to “neutral” to “needs improvement.” Based on the answers provided, the AI will draft a response for them to edit and send. In the future, the AI will ask follow up questions for clarity, suggest ways employees can elaborate on a larger theme, or ask for more specifics about something that happened. 

Beatty says it’s made a huge overall difference in the quality of responses employees give each other. “The agent is kind of like a tutor, because it's coaching them to make feedback better, and as they’re using the tool, it's learning, and it's getting smarter too, so it's constantly getting better at helping others.”

She adds that the AI tool has also helped managers get a better overall picture of an employee’s performance throughout the year. Responses are saved over time, and can be reviewed at a later date through what’s called an “anytime performance summary.”

“A few times a year we'll do talent discussions, before the different promotion cycles and it’s really helpful to take these summaries of all of the different feedback and reflections and metrics and have that all rolled together when it comes time [for me] to have these kinds of discussions with our CEO.”

Brit Morse
brit.morse@fortune.com 

April 1, 2025: This story has been corrected to reflect Angela Beatty’s title, the name of Accenture’s AI tool, and clarify how the tool works currently, and how the company is planning to use the tool in the future. 

This story was originally featured on Fortune.com