Projects with a progress score
Each project records its type (payroll or time), country, cycle, and source and target systems. Completion comes from what exists: two mappings, two pay datasets and one completed run, each worth 20%.
Features
SmartCompaire is payroll comparison software for parallel testing. Load source and target pay data from spreadsheet exports, pair it through your own employee and pay code mappings, grade each difference against your thresholds, and track every open item until it is resolved or explained.
A project holds one payroll or time comparison. Its four setup steps (employee mapping, pay code mapping, source and target pay data) each accept a file, typed rows or pasted rows. Data arrives as exported files; there are no live connections to payroll systems.
Each project records its type (payroll or time), country, cycle, and source and target systems. Completion comes from what exists: two mappings, two pay datasets and one completed run, each worth 20%.
Upload .xlsx, .xls or .csv files of up to 100 MB by default; the first worksheet is read. Headers match regardless of case or punctuation, and leading zeros in IDs are kept.
Learn moreFiles over 2 MB and pastes over 500 rows import in the background, with progress and a cancel option. Each import is one database transaction, so a failed import writes nothing.
Each rejected row is reported with its spreadsheet row number and a plain reason, while good rows are saved. A downloadable CSV keeps the original cells, with formula-like cells neutralized for Excel.
Amounts are stored as exact decimals, not floating point. Parsing accepts $ signs, thousands separators, accounting negatives such as (12.00), Excel date serials and common date formats.
Each setup step lists past imports: when, which file, who, rows imported and rows rejected. Imports add to existing data; replace mode is limited to admins.
Matching and grading follow fixed rules, applied the same way on every run. When a match is not certain, the row stays unmatched instead of being guessed.
Rows match on tax authority plus pay code, or on pay code alone if it is mapped once. Ambiguous rows stay unmatched. Several codes mapped to one are compared as a total.
Learn moreA pay code missing from either system still appears as a result row. Unpaired rows are labeled No employee mapping or No paycode mapping.
See how many rows will match a mapping, the unmatched pay codes with the reason, and employees without a mapping, with plain-language fixes. It uses the same rules as the run.
Set an absolute amount per Bucket and Bucket code type; a row passes when its Amount difference is within it. Organization defaults are copied into each new project for editing.
Label rows by comparing the difference with the threshold or a fixed number, using operators such as <= or >. Organizations can add rules and override or switch off the defaults.
Each run re-matches against current mappings, so a fix takes effect without re-uploading. Runs keep a copy of the thresholds and rules in force, so later edits leave past results unchanged.
Source amount, target amount and difference per employee and mapping group, plus Number and Rate differences. Only Amount is graded, and all periods in a project are totaled together.
Reviewers record what each difference means, and open items stay visible until closed.
Filter the rows on screen by any column, including comparisons such as >=100; filters apply to the loaded page of up to 1,000 rows. Export the whole run to Excel with the columns you choose.
Learn moreSet a status, an override, a linked defect or KD/AD record and a comment on any result row. Bulk status changes are limited to admins and to chosen rows or one run.
Defects record bucket, employee group, type, priority from Low to Critical, assignee, due date, status and resolution comments.
KD/AD records capture expected differences, with a justification document link and an approval status. Rows are linked one by one; later runs are not reclassified automatically.
A project must be archived before deletion, which requires typing its exact name and is written to the audit log.
Payroll comparison results roll up into charts and six report types for sign-off.
Totals per project or run, and outcomes grouped by bucket, code type, employee or pay code with counts and total variance.
Chart over-threshold counts across up to 60 runs of a project to see whether each new run closes the gap.
Count defects by status, priority or assignee. The dashboard filters by country, cycle and status, and saves as a PDF snapshot.
Dashboard, Comparison, Analytics, Defect, KD/AD and Executive Summary reports, as PDF or Excel, for a project, a run or all active projects. The Comparison report lists up to the 500 largest over-threshold rows.
Reports can be downloaded for 14 days after they are generated, download links last 15 minutes, and downloads are counted.
Hirra is in beta and can make mistakes, so check figures that matter. Matching and grading never depend on it; it explains results the fixed rules produced.
Ask what is wrong with this project or what to do next. Hirra knows which project is open and answers from its setup, readiness, latest run and recent imports.
Learn moreHirra uses your sign-in, so it sees and changes only what you could by hand, inside your organization. It cannot edit pay amounts or delete data.
Members and above can have Hirra raise and resolve defects, create KD/AD records, edit a row's review fields and queue reports. Runs, imports, bulk status changes, thresholds, status rules, project settings and archiving stay on their own screens for now.
A written review of one run with ranked findings. Findings are checked against the run's figures before you see them, and percentages are computed in code.
Hirra draws bar, line, donut, scatter and heatmap charts from data it fetched. English voice input uses the browser's speech recognition.
Requests go from the SmartCompaire backend through its Corra AI service to OpenRouter; keys stay on the server. To answer, Hirra sends the data it reads, which can include employee numbers and amounts, to OpenRouter, which chooses the model provider for each turn. Each tool call Hirra reports is written to an append-only AI audit log.
Access is set per organization and per role, and the API checks it on every request.
Database row-level security keeps each organization's data apart; a query with no organization set returns no rows. SmartCompaire platform staff can switch into an organization for support, and each switch is audited.
Learn moreViewer, member, admin and owner are presets over 28 named permissions, and the API checks the role or permission on every request. Owners can adjust some permissions per person, such as projects, defects, KD/AD, reports, comparison runs and people.
Invite people with a role; invitations expire after 72 hours by default. Email goes through your own SMTP server, or you share the link. No one can grant a role above their own.
Email and password (12 characters minimum, locked for 15 minutes after 8 failed attempts) or Google sign-in, which SmartCompaire can make mandatory for an organization.
Sign-ins, role and permission changes, invitations and project events are logged. A database trigger rejects edits and deletions. The log is kept by SmartCompaire and is not yet viewable inside your organization.
No. SmartCompaire imports spreadsheet or CSV files exported from SAP, Workday, Oracle or ADP, and also accepts typed rows or rows pasted from a spreadsheet. There are no live connectors, scheduled pulls or SFTP folders.
No. SmartCompaire finds, grades and records differences; it writes nothing back to the source or target payroll system.
Bring a source and a target extract from your parallel run; we will walk through the comparison with you.
Request a demo