Use continuous performance evidence and AI-assisted merit bands to propose salary increment percentages, while requiring admins to apply changes through the standard payroll increment flow.
01Review scoreIncrement Recommendations process stepActive
02Generate recommendationIncrement Recommendations process stepActive
03Inspect rationaleIncrement Recommendations process stepActive
Every view stays structured and reviewable — nothing hides behind a tab.
Performance scores can be translated into a merit band for planning.
AI can recommend an increment percentage, but it remains advisory.
An authorized admin decides whether to apply, change or reject the recommendation.
Approved recommendations go through the normal payroll increment action instead of bypassing compensation controls.
Use continuous performance evidence and AI-assisted merit bands to propose salary increment percentages, while requiring admins to apply changes through the standard payroll increment flow.
Recommendations are based on scoreboard history and merit evidence instead of a single subjective opinion.
The engine can propose a merit band and percentage using configured performance signals.
Administrators approve and apply changes through normal payroll increment creation.
Performance signals can be grouped into recommendation bands.
The recommendation includes reasoning for human review, with a deterministic rule fallback when AI is unavailable — strong, standard and small bands map to set percentages, plus a tenure bump at two years.
The system never moves salary without admin approval.
Approved recommendations create standard increment records.
Recommendations look across roughly six months of evidence, and employees under three months of tenure are skipped.
Each employee holds at most one pending recommendation; regenerating replaces it rather than stacking duplicates.
Bands map the 0–100 composite onto an increment percentage. This is the deterministic ladder — the one that runs when AI is switched off or unreachable, so the recommendation never depends on a model being available.
A recommendation is a proposal, never an action. An admin can change the percentage before approving, and applying it goes through the same payroll increment flow as any other raise.
The 74.5 worked out on the tracking page lands in Standard → 8%, plus one point for two years' tenure → a 9% proposal, with the rationale attached.
No. AI recommends; admins approve.
Through the standard payroll increment flow.
Merit evidence, scorecards and scoreboard results provide the basis.
Map your current workflow to the right HRM setup, permissions and rollout plan.