An AI Coach for Kinder, Better Feedback

In almost every profession, feedback is fraught. Some managers dread giving it. Employees brace for it. Across industries, feedback often lands as cruel rather than constructive.

Shrihari Sridhar, senior associate dean and Joe Foster ’56 Chair in Business Leadership at Texas A&M University’s Mays Business School, has seen the damage firsthand and created a tool that proposes a fix. Sridhar first outlined the EMPATHY framework in a Journal of Marketing editorial designed to help reviewers offer feedback that is rigorous yet kind. Building on that foundation, his colleagues Pallav Routh, assistant professor at University of Wisconsin-Milwaukee, and Ryan McCarroll and Ayush Garth, founders of AnswerThis.io, worked with him to transform the framework into EMPATHY AI — a tool that coaches users in real time on how to deliver motivating, constructive feedback.

“A user pastes their draft review or feedback into the system,” Sridhar explains. “The tool then analyzes the text across seven dimensions: Does it start with a clear end goal? Does it reflect a developmental mindset? Has the user engaged deeply with the material? Are critiques prioritized rather than scattered? Is the tone professional, not toxic? Does it balance strengths and weaknesses? And does it provide a roadmap rather than a jumble of comments?”

For each of those seven dimensions, the EMPATHY AI system offers prompts and suggestions. For example: If someone writes in a peer review, “The methods are weak,” the tool might prompt: “Can you suggest a specific improvement, such as expanding the sample or clarifying the justification?” And if the tone drifts into personal language, the tool will suggest reframing to keep the focus on the work. Sridhar said users can see how their language shifts from vague critique to precise, constructive guidance.

“What makes the tool powerful is that it doesn’t write reviews for people; it teaches them how to write better ones themselves,” he said. “It’s a mirror, reflecting where the feedback lacks empathy or clarity, and a coach, offering concrete ways to improve.”

A Tool Born From Frustration

Sridhar developed the editorial and co-developed EMPATHY AI after seeing the impact feedback can have in the academic peer review process. Peer reviews are the lifeblood of academia, he said, yet too often they discourage rather than develop. At top journals, rejection rates exceed 90%, he said, and anonymous reviews can sometimes bring out the worst in people — spitefulness, performative toughness or careless cruelty.

He noticed most reviewers are untrained, offering vague feedback without suggesting a path forward. Plus, their delivery is often blunt and lacking empathy. “They learn by imitation, often replicating the harshness they once endured,” Sridhar said. “The result is a system where feedback feels like punishment rather than mentorship.”

When Sridhar publicly shared his frustrations, the response was overwhelming. “Hundreds of scholars reacted with relief, anger and hope, echoing their own painful experiences of receiving reviews that were dismissive, sarcastic or vague,” he said. “The outpouring confirmed that this was not an isolated issue — it was systemic.” What began as a framework for reviewers has now been adapted into an AI system, thanks to collaboration, with broad applications across industries.

He emphasizes that the framework is not just for scholars — it’s for anyone who gives feedback.

Supervisors conduct annual performance reviews, managers coach teams, teachers evaluate students, doctors deliver diagnoses and lawyers advise clients. The challenge is the same in each of these contexts: how to be candid without being crushing, and how to motivate improvement without sugarcoating reality.

“Imagine a manager saying, ‘You didn’t meet expectations this year’ versus, ‘You’ve made progress on X and Y, but Z still needs improvement — here are three ways to get there.’ The difference is enormous for morale and development,” Sridhar said. “In health care, a doctor using EMPATHY principles might frame a diagnosis not as a blunt statement but as a roadmap for treatment and hope. In schools, teachers can provide feedback that encourages growth rather than discourages effort. In corporate law or consulting, advisers can guide clients with clarity and professionalism while maintaining trust.”

Human Communication Augmented With AI

Technology can strip nuance, leaving us feel disconnected. Paradoxically, AI can make us more human, Sridhar said. When used deliberately, he believes AI can serve as a tutor for humanity.

EMPATHY AI learns from patterns of feedback that are effective, precise and kind, and then coaches users on how to emulate those patterns.

“It helps us slow down and choose words that motivate rather than wound,” he said. “This matters because humans often hesitate to be kind — they fear it will be seen as weakness. AI removes that hesitation by modeling how kindness and clarity coexist.”

Sridhar believes that as automation takes over repetitive tasks, “the real differentiator will be our humanity.” Empathetic, motivating and professional communication will stand out.

“The more AI pervades our lives, the more valuable our human qualities become,” Sridhar said. “EMPATHY shows that AI is not about replacing human connection but about coaching us to bring out our best selves in every interaction.”

This article was originally published on Texas A&M Stories.