Case Study · Human Resources

The case management system HR teams have been building in spreadsheets

An HRBP receives a complaint — discrimination, bullying, and a pregnancy disclosure all in one email. Somewhere in their inbox, there's a prior warning about the same employee from three months ago. The discrimination playbook requires a response within 24 hours. And the pregnancy means two separate processes need to run from day one. This is the problem peoplease is built to solve.

peoplease.ai HR platform Human Resources & Employee Relations Australia (Fair Work Act 2009) Enterprise HR teams
Hours Min
End-to-end case creation after a single capture
6mo+
Of entity history surfaced from the knowledge graph at case open
Zero
Fair Work playbook steps missed — every procedure covered automatically

The context

HR Business Partners manage ten to thirty active employee relations matters at any given time. The context for each one is scattered — emails, Slack messages, meeting notes, memory. Patterns across months go unnoticed. A verbal warning from six months ago sits in a folder somewhere. A repeated complaint about the same manager never gets connected to the first one.

When something goes wrong — a missed Fair Work procedure, a general protections claim, a pregnancy that should have triggered a separate process from day one — the cost is real: legal exposure, Fair Work Commission proceedings, damaged careers on both sides, and an HRBP left wondering what they missed and when.

The current toolkit is email, spreadsheet, and memory. Not because HRBPs don't know better. Because nothing has been built that actually fits how this work is done.


What we're building

01 · Capture & AI Triage

Paste the complaint. The AI does the rest.

The HRBP's one job is to capture the information — paste an email, a complaint summary, a meeting observation. No forms. No dropdowns. No taxonomy decisions. Just paste and move on.

The AI classifies the capture immediately: complaint type, severity, applicable legislation, matched playbooks, and every entity mentioned — complainant, subject, witnesses, manager. It checks the knowledge graph for prior history and surfaces what it finds: prior warnings, past complaints, patterns across months. A critical discrimination matter with a 24-hour Fair Work SLA gets flagged before the HRBP has read the second paragraph.

Case created in under 5 minutes · all parties identified · severity and SLA flagged · prior history surfaced automatically
02 · Knowledge Graph Memory

The context that doesn't fall through the cracks

Every capture, every entity, every relationship is stored in a knowledge graph. The system remembers everything: a verbal warning from six months ago, a complaint filed in October, a manager who appears across three separate matters. When a new case opens, that history surfaces automatically — not because the HRBP searched for it, but because the graph knows the connection exists.

Pattern detection runs across the full graph. When multiple captures reference the same people or describe the same behaviour, they're grouped into a cluster — the full picture of a matter, assembled automatically from individual observations over time. The HRBP sees what the data shows, not just what arrived this week.

Full entity history at case open · cross-case pattern detection · clusters assembled from months of captures · nothing buried in inbox
03 · Playbook-driven Documents

Fair Work procedures that never get missed

When the HRBP confirms a case, the AI generates the full document set from matched playbooks: investigation invitations with the Fair Work support person right included, acknowledgement letters, manager guidance emails, witness statement requests, show cause letters, and where a pregnancy or protected attribute is involved, a separation memo to split the processes from the start.

The AI drafts. The HRBP reviews, edits if needed, and sends. It never acts unilaterally. The playbooks encode what Australian employment law requires — procedural fairness, reverse onus awareness for general protections, 24-hour SLA obligations for discrimination — and every document reflects those requirements, every time.

Full document set generated on case creation · Fair Work procedures baked in · HRBP reviews before anything is sent

Why we're building this

Employee relations is one of the highest-stakes operational functions in any organisation — and one of the most under-tooled. A missed procedure creates legal exposure. A pattern unnoticed across months becomes a Fair Work claim. A document drafted without the right clause leaves the organisation exposed in a Commission proceeding.

The HRBP isn't the problem. The tools are. Inbox-based case management, manual document drafting, and memory as the knowledge layer is not a system — it's a liability. We're building the system that should have existed.

The AI's job isn't to decide what happens to an employee. It's to make sure the HRBP has the full picture — the history, the procedures, the risks — so the decision they make is the best one it could be. Judgment stays human. Context doesn't get lost.

Running HR or employee relations in Australia?

We're in early conversations with HR leaders managing employee relations at scale. If context loss, missed procedures, or manual case admin sounds familiar — we want to understand your version of it.

Talk to us → hello@peoplease.ai