AI manager coaching can shape performance review conversations

AI manager coaching performance reviews can affect how a manager prepares feedback about an employee’s work, pay or job status. The tools may help shape the language, examples and tone used in a high-stakes meeting, although the evidence provided does not show that they independently decide compensation, layoffs or final ratings.

CNBC reported Sept. 12 that managers can use AI coaching products to rehearse difficult conversations before speaking with an employee. That preparation can matter when a review informs a raise, a development plan or a decision that an employee is not meeting expectations.

The basic idea is role-play. A manager can ask a tool to simulate an employee who is angry or emotional, then practice responding. They can also request feedback on whether their message is too blunt, too vague, repetitive or filled with verbal pauses. CNBC reported that some tools can help managers test directness and concision, while an HR executive said a tool may flag potentially legally risky topics.

That does not turn the AI into the manager. Karan Kashyap, chief executive of AI role-play provider Posh, told CNBC that practicing different reactions can help because managers cannot know in advance how a real conversation will unfold. Carnegie Mellon business communication professor Emily DeJeu told CNBC that preparation is appropriate for consequential workplace discussions.

How can AI manager coaching affect a performance review?

AI can influence the process before the meeting and the written material around it. Culture Amp says its AI Coach can synthesize feedback and recognition from its platform, help produce a first draft of performance content, and generate talking points or role-play support for review discussions.

For an employee, that means a manager’s assessment may be organized or worded with AI assistance. It does not mean the system makes the final rating in every case. Culture Amp says its product does not recommend final performance ratings, and that the manager sets them. The company also says users must give the tool context and prompts rather than having it automatically write an entire review.

CNBC cited a July survey fielded by The Predictive Index, which sells a manager role-play chatbot. In that survey, 42% of managers said they knew what they wanted to communicate in difficult conversations but struggled with how to say it. Meanwhile, 74% of CEOs and business leaders said their managers were very confident handling difficult people conversations alone. Only 20% of managers said they preferred to prepare without a framework, coaching guide or HR input. Because the survey was conducted by a vendor of such tools, its results should be read with that context in mind.

Where should human judgment stay in the process?

CNBC reported that sensitive employee information, legal concerns and complicated cases still require human judgment. Culture Amp likewise says its AI Coach is not legal or compliance advice and directs managers to consult HR or legal teams on sensitive matters.

Employees can also use general AI tools such as ChatGPT or Claude to prepare more articulated responses if they expect a tense review conversation, CNBC reported. Culture Amp says its platform can assist employees with self-reflections and peer or upward feedback during an active review cycle.

Privacy terms depend on the product and the employer’s setup. Culture Amp says interactions with its AI Coach are private from managers and administrators, and says third-party large-language-model providers do not log its conversation data or train on it. Those statements are specific to Culture Amp’s product, not a rule for every workplace AI tool.

• Ask whether AI is involved in drafting reviews or preparing conversations.
• Ask what information the tool receives and who can view its outputs.
• Ask how employees can correct factual errors in a review or challenge an assessment.

This story draws on original reporting from CNBC.