Performance Management

Software helps managers and employees prepare, structure, and support performance reviews with survey data, feedback, and review context.

Business impact

• Employee review participation — Glean reports a 7% increase at Zillow.
• Review-agent runs — Glean reports the review agent was used almost 10,000 times in its first review cycle.

Data requirements

• Employee survey results (tabular) — Used as review input for manager support.
• Review resources (text) — Used to provide quarterly review context.
• Feedback text (text) — Encoded as semantic embeddings and grounded in the ontology.
• People, skills, tasks, projects, and operational interaction data (graph) — Modeled in Neo4j for retrieval, inference, and similarity analysis.

AI methods and techniques

• Agent-based retrieval and assistance — Used to surface review resources and employee survey results for managers.
• Similarity search — Used in the graph-based system to infer profile similarity from connected people and skills data.
• Community detection — Used to identify communities in the workforce graph.
• Centrality analysis — PageRank is used to identify knowledge hubs.
• Graph embeddings — Structural and semantic embeddings are combined to compare people, roles, skills, and feedback.
• Ontology grounding — A proprietary ontology constrains LLM inference and anchors outputs to the graph schema.

Industries

Real-world evidence

3 documented case studies on record.

Companies using this: Salesforce, Zillow, Kipon.