Architecture

How the measurement system works

People Analytics does not lack dashboards. It lacks a reliable bridge from strategy to definitions, data, evidence, and action. This product is designed around that chain.

Strategy-first

Start from the business outcome, not a file upload.

Human + AI co-design

AI proposes frameworks; people confirm definitions.

Deterministic analytics

Code calculates. AI explains.

Local-first privacy

Raw HR rows stay in the browser by default.

Semantic metric layer

Approved definitions become reusable knowledge.

Visible uncertainty

Confidence, limitations, and data gaps stay on the page.

AI vs code

AI understands and explains

Clarify strategy and metric trade-offs
Propose mappings and analysis plans
Write executive narrative and slide storylines

Code calculates

Stage contribution and readiness scores
Joinability and data-health checks
Chart data, PPTX export, and audit objects

Default privacy rule

Raw employee and candidate rows stay in the browser. Remote AI calls receive table names, column names, aggregate profiles, approved metric definitions, and statistical outputs. No row-level People data is sent by default.