Config Copilot
Turn a plain-language request into a plan from a fixed action catalog, visibly fill the real form and wait for Confirm.
Use permission-scoped AI on the real product screen to plan supported changes, fill approved forms, explain anomalies and assist implementation—without allowing autonomous payroll, leave or configuration decisions.
Configure the module around your structures, policies, ownership and approval model - not a generic process diagram.
Turn a plain-language request into a plan from a fixed action catalog, visibly fill the real form and wait for Confirm.
Ask about industry, size, state, employment mix, work pattern and overtime, then suggest real configuration masters.
Use tenant data and current-screen context through role-scoped agents rather than a generic detached chatbot.
Flag unusual patterns or variance for authorised human investigation.
Use deterministic checks first and AI when confidence is low, keeping the result reviewable.
Guide the implementation team using the selected tenant’s live configuration context.
Route tasks through rules, heuristics, local models or more capable models while preserving graceful degradation.
The example below is illustrative. Measures and thresholds should be configured around your operating definitions and review cadence.
Configure attendance, leave, payroll, approvals, permissions and notifications around the way your organisation actually works, while keeping every policy scoped, dated and reviewable.
ExploreConnect effective-dated salary structures, attendance and leave inputs, earnings, deductions, reimbursements, recoveries and statutory calculations in a payroll run that can be reviewed and rolled back before finalisation.
ExploreBring headcount, movement, people cost, attendance, payroll, performance and HR service data into role-based dashboards with consistent definitions and explainable drill-down.
ExploreGovern goals, appraisal calendars, eligibility, review windows, exceptions, calibration, development actions and salary-revision handoff from one effective-dated workforce foundation.
ExploreUse your actual workforce groups, approval paths and edge cases during product evaluation.
No. Those autonomous actions are not claimed. AI can prepare a supported action, but an authorised person must review and confirm it.
A self-learning training pipeline using customer data is not currently claimed. AI context is described as tenant- and role-scoped, with logged interactions and bounded actions.
Bring one representative workflow or current process challenge. The demo will show the operating path, controls and connected outcomes.