Intelligence
Ask your compensation data anything.
Evenpay Copilot answers compensation questions in plain language, backed by a regression engine that separates the pay differences your factors explain from the ones that need a decision.
45 minutes. Bring a question you cannot currently answer.
Intelligence, not another dashboard to interpret.
A colleague who has read all of your pay data
Ask questions the way you would ask an analyst: where are my biggest pay gaps, should this employee get a raise, how does this team compare. Copilot analyses, benchmarks, visualises and recommends across employees, job architecture, evaluations, transparency data and pay rationale. It is available as an early access feature that an administrator switches on for the organisation.
- Plain language questions, with the working shown
- Spans employees, architecture, evaluations and rationale
- It informs your decision, it never makes the decision
A twelve percent gap is not one number
Blinder-Oaxaca decomposition splits a raw gap into the part your objective factors account for and the part they do not. A twelve percent raw gap might be eight percent explained by role, experience, level and performance, leaving four percent that needs action. That four percent is the number worth your time.
- Explained and unexplained portions separated
- Factors reflect your own pay philosophy settings
- The unexplained share is defensible, not a raw average
Every figure shows its workings
Model quality, significance, confidence intervals and multicollinearity checks are all on display. When someone asks how you know, you can show them rather than assert it.
- R squared, p values and confidence intervals visible
- Multicollinearity checks so factors are not double counted
- Minimum of twenty employees for adjusted analysis
Explore gaps across any combination you like
Cross position, level, organisational unit, job family, grade, career track and custom fields, drill three levels deep, then click through to the actual employees behind a cell. Fairness Index, department heatmap, age and quartile analysis come alongside.
- Any dimension combination, three level drill down
- Click through from a cell to the underlying employees
- Fairness Index, heatmaps, age and quartile analysis
AI that never sees a name
Employee names are tokenized before any AI call, minimum group sizes are enforced, the organisation boundary is strict, all AI processing happens inside the EU, and your data is never used to train models.
- Names tokenized before any AI call
- All AI processing inside the EU, no training on your data
- Minimal risk classification under the EU AI Act
How an answer gets built
Every answer traces back to your data and your definitions.
You ask in plain language
No query builder, no report to configure first.
Your pay factors are applied
The analysis uses the criteria your pay philosophy defines.
The model separates the gap
Explained and unexplained portions, with statistics attached.
You decide
Copilot recommends and evidences. The judgement stays with you.
Common questions about Evenpay Intelligence
Intelligence runs underneath everything
Pay Gap Alerts & Anomalies
The same engine, watching continuously and telling you when to look.
Pay Gap Justifications
Document the objective factors the analysis surfaces.
Salary Recommendations
Predictions built on the same model, with gender left out by design.
Security & Privacy
Tokenization, EU processing and the guardrails around the AI.
Ready to ask better questions?
Bring a compensation question and see Copilot work through it.