The Challenge

Every sprint, the process broke before the feedback could mean anything.

• 2–3 PM spent hours per sprint just to collect responses
• Numbers reflecting bias, not actual performance
• Employees had no structured view of where they stood
• Short-form fields, short-form answers
• Results sat in spreadsheets nobody maintained

The Objective

Feedbackly was built to close every one of those gaps.

• Eliminate PM coordination by automating the full feedback lifecycle
• Replace numeric ratings with AI-derived scores from written text
• Give employees visibility into where they stand against career level expectations
• Surface skill gaps and readiness signals for targeted development
• Enable leadership to analyze trends without manual data work

The Solution

Automated Feedback Cycles: EventBridge and Lambda open each sprint cycle, identify reviewers, and close with a final reminder. PMs are out of the loop entirely.

AI-Derived Scoring: Written feedback is validated for authenticity and relevance. Responses that pass are scored across KPI’s Reliability, Productivity, and Positive Energy, 1 to 5 in 0.5 increments, derived from text, not gut feel.

Conversational Intelligence: A Strands agent reasons across peer feedback, skills profiles, 360-review records, and career level expectations - letting employees, mentors, and leadership ask questions about performance, surface growth gaps, and get grounded answers about what to do next.

The Impact

The shift wasn't just in participation. The data itself became worth using.

• PM coordination time per sprint: eliminated - cycles open, run, and close automatically
• Performance scoring: subjective numbers replaced with AI-derived, text-grounded metrics
• Employees can query what they need to demonstrate to reach the next level and get a grounded answer.
• Most feedback systems collect data. Feedbackly is built to make it mean something. Feedback goes in as writing, comes out as scores, trends, and answers. The scores build a dataset. The dataset is one you can actually talk to. That is what an AI-native feedback system looks like.