Evidence-first staffing.
Built for people.
A single workspace where credentials are verified, roles match on evidence, and the economy of every placement stays visible. No opaque scores, no autonomous decisions.
Warehouse operative · Birmingham
Role boardFour honest layers.
Evidence makes matching defensible. Matching makes operations accountable. Operations make the economy visible. Each layer builds on the one before.
Trust Layer
Credentials with provenance and explicit states. A declaration is not a confirmation. Documents expire on their own. Contestations leave a trail.
Role Board
For each role and candidate, which requirements are met, which are missing, which have expired. Actions have owners and deadlines.
Operations
Request, propose, accept, confirm, submit, approve, invoice, receipt. Decimal money, UTC-stored shifts, DST handled without errors.
Economy
Per-request contribution derived strictly from recorded data. Unknown costs flagged, never silently zeroed. Honest numbers only.
Exceptions
Seven live detections: unconfirmed shifts, contested timesheets, overdue invoices, mid-engagement expiry. Each links to the record to act on.
AI Assist
Optional Claude integration. Drafts internal notes and suggests actions you review before saving. Hard token budget, full audit, admin-only.
A credential has a state. A declaration is not a confirmation.
Every credential travels an explicit path — declared, document received, in verification, confirmed. Rejected credentials, expiries, and contestations each leave a visible trail. The verifier is always a human; the AI can propose, never confirm.
- Right to Work, DBS, SIA, CSCS, First Aid and more — ready to use
- Immutable audit history on every state change
- Contestation process with upheld / dismissed resolution
- Employer only ever sees eligibility — never the private document
Priya Shah · Credentials
Private · controlled by PriyaWarehouse ops · Economy
Request #142 · recorded data onlyHonest numbers. Visible costs. Nothing silently zeroed.
Every staffing engagement reads from the same ledger: hours in each state, revenue invoiced and received, labour pay for completed work. The resulting contribution is derived strictly from recorded data — and the costs we do not record are listed explicitly, so the figure is understood as an upper bound.
- Decimal money in pence, immutable invoice snapshots
- Overnight shifts tracked in UTC, DST-aware
- Credit notes and refunds preserve the original record
- Receipt reversal and refund with audit trail
Three commitments we design for.
Opaque algorithms are not safer because they are automatic — they are harder to contest. Our product is built around visibility, not inference.
No employability score.
People are not scored. The system shows evidence, lists what is missing, and asks a human to decide.
No autonomous decisions.
AI drafts and suggests. A person always confirms. The platform works unchanged without any AI connected.
Honest about unknowns.
If a cost is not recorded, we say so. If a credential is not verified, we say so. Nothing is silently assumed.
Open the workspace.
Explore the full product. Credentials, roles, operations, economy and exceptions — all live.