Data Integrity & ATS Evaluation — Mandate Plans
Mandates PS-013 through PS-015: the reconciliation agent, workflow instrumentation, and the ATS decision.
PS-013: Monitor Data-Integrity Discrepancies
What this actually involves
The pay reconciliation agent (MRP-109) already handles much of this. It runs weekly and catches:
- Field integrity issues (blank fields, wrong domains, invalid values)
- Accrual/time-off bank mismatches
- Zero-hours-worked anomalies
- Benefit class FT/PT (Full-Time/Part-Time) inconsistencies
- Wage compliance violations
- Job change closeout gaps (Monday verification)
What it does NOT cover:
- Cross-system drift (Paylocity vs. Greenhouse vs. Lattice vs. Legion disagreeing about an employee)
- ACA/1095-C reconciliation (procedure not yet captured; out of scope)
- Real-time detection (runs weekly, not continuously)
Who does this today
- Syra — reviews the reconciliation agent's Exception Queue weekly (Monday after Sunday delivery)
- Grecia — owns compliance-level data integrity (wage, benefits, ACA)
- Beth (departed) — formerly co-owned biweekly Paylocity audits with Syra (by region split)
What's already built
The reconciliation agent is LIVE and running weekly. Full detail in brunz-mr/agents/pay-reconciliation.md.
Concrete tasks
- Review Syra's Exception Queue from the most recent week. Understand what findings look like, how she clears them, and what patterns repeat.
- Identify gaps in the recon agent's coverage: what discrepancies does Syra still catch manually that the agent misses?
- Assess cross-system drift: when Paylocity says one thing and Legion/Lattice/Greenhouse says another, how does the team notice?
- Review the auto-clear mechanism: when Syra fixes something in Paylocity, does the next agent run correctly mark the finding as cleared?
People to meet
| Person | Questions to ask |
|---|---|
| Syra | Walk me through your Monday audit review. Which findings are easy to clear and which require investigation? What do you still check manually that the agent doesn't catch? |
| Kyle | What's the current state of the recon agent rules? Any known gaps? What was the outcome of the Benefit Class Report audit decision? |
What to look for in mr-people
automations/recon/— operator manual, rules codebook, report contracts- Any Syra-facing runbooks or documentation
PS-014: Instrument Every Recurring Task
What this actually involves
This is the meta-mandate: while you do all the other work manually in August, you record every step. The instrumentation is what feeds the September automation builds.
For each recurring task, capture:
- What: exact steps performed, in what order
- Where: which system(s), which screens/fields
- How long: time per step and per full workflow
- Decisions: where you made a judgment call vs. followed a rule
- Exceptions: what went wrong, what was unexpected
- Volume: how many times per week/month
Concrete tasks
- Set up a task-logging system from Day 1. Could be a spreadsheet, a markdown log, or a structured form. The key is consistency.
- Log every onboarding, job change, termination, and leave entry you process in August.
- At end of each week, review your logs and categorize: which tasks were purely mechanical (automatable), which required judgment, which were one-off exceptions.
- By end of August, produce a "automation target list" ranked by hours-per-month and automateability.
People to meet
| Person | Questions to ask |
|---|---|
| Kyle | How did you instrument the workflows when building the existing agents? Is there a template or pattern I should follow? |
| Syra | Do you track how much time you spend on each type of task? Even roughly? |
PS-015: Provide Technical Read on ATS Decision
What this actually involves
The ATS (Applicant Tracking System) decision is between:
- Keep Greenhouse — renew the contract, pay ~$12K to rebuild the broken API
- Switch to Pinpoint — native Workday integration, Slack approval flow, better for high-volume retail hiring
- Wait for Workday Recruiting — part of the Workday HRIS migration (Q2-Q3 2027)
Team consensus leans heavily toward Pinpoint. The decision is due at the August 6 tech stack meeting (DAISY accountability matrix framework).
Key timelines
| Date | Event |
|---|---|
| August 6, 2026 | Tech stack decision meeting (ATS, HRIS, LMS decisions) |
| August 31, 2026 | Greenhouse Harvest API v1/v2 deprecated |
| September 2026 | If Pinpoint: migration scripting starts |
| January 13, 2027 | Greenhouse contract expires (hard deadline for migration) |
Factors for the decision
| Factor | Greenhouse | Pinpoint |
|---|---|---|
| Workday integration | No (needs API bridge) | Native |
| High-volume hourly hiring | Poor fit | Paradox (Workday-owned) upgrade |
| Slack hiring-manager approval | Not available | Native connector |
| Migration cost | $0 (keep current) vs. $12K API rebuild | Migration scripting (Daniel's scope) |
| Contract | Expires Jan 13, 2027 | New contract |
| Workflow configuration | Manual/rigid | Workflow-based |
Concrete tasks
- Attend the August 6 tech stack meeting or review the outcome immediately after.
- If Pinpoint is selected: explore Pinpoint's admin interface, understand its data model, assess what the migration from Greenhouse involves (offices, hiring managers, jobs, historical data).
- Evaluate from inside the workflow: as you process onboardings in August, note where Greenhouse helps and where it hurts. Your from-the-inside perspective is part of your value in this decision.
- Assess the API expiry impact: what actually happens when the Greenhouse-to-Paylocity feed breaks? Can you build a temporary bridge, or is manual entry the fallback?
People to meet
| Person | Questions to ask |
|---|---|
| Shlanda | What's your top concern about the ATS transition? What does the ideal ATS do that Greenhouse doesn't? |
| Lindsey | What do your recruiters need from the ATS? Where does Greenhouse fail for high-volume HCB hiring? |
| Kyle | What's the technical migration path for Greenhouse-to-Pinpoint? How much data needs to move? Is there a Pinpoint API for scripted setup? |
Key files in core
clients/madison-reed/work/MR-people-software-stack-build-vs-buy-map.md(Greenhouse section)meetings/clients/madison-reed/2026-07-31-people-team-tech-stack-review.md(DAISY framework, decision meeting)meetings/clients/madison-reed/2026-07-28-mr-cmprssn-weekly-working-time.md(Pinpoint decision framing)