Payroll compliance is a hard problem. The SCHADS Award alone runs to hundreds of pages, and a single timesheet can contain dozens of edge cases — broken shifts, sleepover adjacency, minimum engagements, weekend penalties — that all need to be assessed together. A single AI call, no matter how capable the model, isn't built for that kind of work.

So we didn't build one. CrossVault now runs a coordinated team of AI agents on every upload — each with a specific job, each making the next one's work more precise. The result is a compliance check that's more thorough, more consistent, and faster than anything we could build with a single prompt.

---

Upload your timesheet

To get started, upload your timesheet export from ShiftCare, Deputy, or any other scheduling platform. We accept CSV and XLSX — no reformatting required. Before compliance analysis even begins, our system reads your file's structure and figures out what each column actually means — regardless of how your platform exported it. Column headers vary wildly between platforms, and we handle that automatically so you never have to manually map fields.

We tell you what we see

Every report opens with a short plain-English note about your data. Not a wall of technical output — just the things worth knowing before you read the results.

This matters because raw timesheet exports are often confusing. Status fields that look alarming are sometimes just an artefact of how a platform labels confirmed shifts. Zero-dollar expense columns might mean pay is processed in a separate system, not that workers weren't reimbursed. We surface these things upfront so you can read the compliance findings with the right context.

Your compliance report

The compliance report works through every row of your timesheet against the applicable award rules — checking minimum engagement periods, broken shift spans, overtime thresholds, penalty rates, and allowances. It doesn't guess. It shows its working.

Each issue in the report includes what rule was breached, which shift triggered it, and what the correct entitlement should be. At the top, a summary card gives you the headline numbers: total shifts checked, issues found, and categories of concern — so you can triage before reading the detail.

The whole process takes under a minute for most files. For larger exports — hundreds of shifts across multiple workers — it scales without degrading in accuracy, because the work is distributed across agents rather than piled onto one.

We think this is how compliance tooling should work: the AI does the tedious, rules-heavy analysis; you make the calls that need human judgment. Upload a timesheet and see for yourself.