Google Workspace application
OpenAI embeddings and RAG make performance-review evidence available through natural-language queries.
← Portfolio
Case study 02 / Performance-review intelligence
PRISM
Turning performance-review evidence into a conversation HR leaders can query.
01 / Problem
The question behind the work.
Discussions about promotion readiness and expanded responsibilities need access to relevant performance-review evidence. Leaders need a practical way to ask questions of that material.
02 / My ownership
What I took on.
I built and deployed PRISM as a performance-review intelligence application on Google Workspace, using OpenAI embeddings and retrieval-augmented generation.
03 / Solution
From problem to working solution.
- 01
Use embeddings and RAG to make performance-review evidence accessible through natural-language queries.
- 02
Bring relevant review material into HR leadership discussions about promotion readiness and expanded responsibilities.
- 01Review evidence
- 02Embeddings & retrieval
- 03Natural-language query
04 / Evidence
The work, with context.
05 / Results
What changed.
Enabled HR leaders to query review evidence in natural language to support discussions on promotion readiness and expanded responsibilities. HR leaders retain the judgement in those discussions.