Skip to playbook

HR09·Employee Relations

Employee Relations Case Summary & Risks

Summarize employee-relations cases, cite facts against policies, and surface compliance risks and cross-case patterns.

Before you start

What you’ll need

  • Access to all employee relations case files for the review period
  • Familiarity with your organisation's employee relations policies and escalation procedures
  • Employee Relations Case Files
  • Organisation Policy & Compliance Framework
  • Escalation & Severity Criteria

Who’s involved

  • HR Specialist — Runs the case review workflow and produces the case trend summary
  • Employee Relations Manager — Reviews risk classifications and approves escalation decisions
  • Legal Advisor — Reviews compliance-related risks and confirms regulatory classification

Safe use

  • Data Handling — This workflow processes employee relations case files containing protected personal data (disciplinary records, grievance details, performance improvement plans). Do not paste these inputs into public or unapproved GenAI tools.
  • Verification — GenAI may conflate facts from separate cases or misclassify policy relevance. Verify every AI-generated case summary and risk highlight against the original case files and organisational policies before distribution.

Execution steps

Hybrid

Inputs

  • Confirm all employee relations case files for the review period are collected and accessible
  • Confirm engagement risk flags from the prior survey analysis are available
  • Confirm your organisation's policy and compliance framework is current and approved
  • Confirm escalation and severity criteria are documented and agreed upon
  • Confirm you are using an approved GenAI tool for this workflow

Prompt

Define structured template for case summarisation

CONTEXT
You will be provided with the following source documents:
1. Employee Relations Case Files
2. Engagement Risk Flags
3. Organisation Policy & Compliance Framework
4. Escalation & Severity Criteria

TASK
Define a structured markdown template for summarising individual employee relations cases. The template must capture case type, timeline, key parties by role title, documented facts with source references, relevant policy sections, and current case status.

OUTPUT FORMAT
Produce a template in this exact structure:

## [Case Reference]
### Case Overview
- **Case Type:** [Grievance / Disciplinary / Performance / Workplace Conflict / Other]
- **Date Opened:** [Date]
- **Current Status:** [Open / Under Review / Escalated / Closed]
- **Parties Involved:** [Role titles only]

### Documented Facts
| # | Date | Event | Source Document | Verbatim Excerpt |
|---|---|---|---|---|

### Relevant Policies
| Policy Section | Relevance to Case |
|---|---|

### Open Items
- [List unresolved actions or pending decisions]

CONSTRAINTS
Do not add fields beyond those specified. Do not include any personally identifiable information in the template example. Do not create scoring or automated decision fields.

Outputs

  • Case Summary Template

Verification: Confirm the template fields align with your organisation's case documentation standards before proceeding

GenAI

Inputs

  • Confirm sensitive personal identifiers have been redacted from case files before prompting

Data Handling: Do not paste case files containing personal identifiers such as home addresses or medical details into the prompt; redact these fields first

Prompt

Summarise each case with verbatim evidence

CONTEXT
You will be provided with employee relations case files and a case summary template defining how each case should be documented.

TASK
For each case file, populate the case summary template. Extract every documented fact with its date and source reference. Quote relevant passages verbatim from the case file. Identify all policies referenced or implicated.

OUTPUT FORMAT
For each case, produce a completed template following the exact structure provided. Every entry in the Documented Facts table must include the source document name and a verbatim excerpt. List all policies referenced or implicated in the Relevant Policies table.

CONSTRAINTS
Do not summarise or interpret facts — extract text exactly as written in the source case file. Do not add risk assessments at this stage. Do not include names of individuals; use role titles throughout.

Outputs

Verification: Verify every documented fact exists verbatim in the source case file before proceeding to analysis

GenAI

Prompt

Map case facts to specific policy references

CONTEXT
You will be provided with individual case summaries and your organisation's policy and compliance framework.

TASK
For each case summary, map every documented fact to the specific policy section it relates to. Classify each fact as a potential violation, a procedural observation, or informational context only.

OUTPUT FORMAT
For each case, produce:

### [Case Reference]
#### Policy-Mapped Facts
| Fact # | Documented Fact | Policy Section | Classification | Explanation |
|---|---|---|---|---|

Classification values: [Potential Violation / Procedural Observation / Informational Context]

Explanation: one sentence stating why the fact maps to that policy section and classification.

#### Unmapped Facts
- [List any facts that do not map to a specific policy section, with a note explaining why]

CONSTRAINTS
Do not classify a fact as a potential violation unless the policy section explicitly addresses the behaviour or event described. Do not speculate on intent or motive. Do not omit facts that lack a policy mapping — list them separately under Unmapped Facts.

Outputs

  • Cited Evidence Sheets
  • Every fact from the individual case summaries appears in the policy-mapped or unmapped section
  • No fact is classified as a potential violation without a specific policy section reference
  • Unmapped facts are listed with clear reasoning for the lack of mapping

Verification: Verify the AI has not fabricated policy section references — confirm every cited policy section exists in your framework

GenAI

Inputs

Prompt

Link engagement risk signals to case patterns

CONTEXT
You will be provided with cited evidence sheets from employee relations cases and engagement risk flags from a prior engagement survey analysis.

TASK
Cross-reference the engagement risk flags against the case evidence to identify where engagement signals correlate with documented case patterns. Note which risk flags have corresponding case evidence and which remain unsupported.

OUTPUT FORMAT
Produce a cross-reference table:

| Engagement Risk Flag | Related Case(s) | Supporting Evidence | Correlation Strength |
|---|---|---|---|

Correlation Strength values: [Strong — direct evidence / Moderate — circumstantial link / Weak — thematic only / None — no supporting case evidence]

Below the table, list any engagement risk flags with no corresponding case evidence under an Unmatched Flags heading.

CONSTRAINTS
Do not infer correlation without citing specific case evidence. Do not treat engagement survey themes as confirmed case findings. Do not create new risk flags beyond those provided in the input.

Outputs

  • Case-Engagement Cross-Reference
  • Every engagement risk flag from the input appears in the table or under Unmatched Flags
  • Correlation strength ratings are supported by cited case evidence
GenAI

Inputs

Prompt

Highlight compliance and practical risks per case

CONTEXT
You will be provided with cited evidence sheets mapping case facts to policies, and the escalation and severity criteria used by your organisation.

TASK
For each case, highlight compliance risks (potential regulatory or policy violations) and practical risks (operational disruption, reputational exposure, escalation likelihood). Assign a severity level per the escalation criteria for each risk.

OUTPUT FORMAT
For each case, produce:

### [Case Reference]
#### Compliance Risks
| Risk | Policy Reference | Severity | Evidence Citation | Recommended Action |
|---|---|---|---|---|

#### Practical Risks
| Risk | Impact Area | Severity | Evidence Citation | Recommended Action |
|---|---|---|---|---|

Severity values: [Critical / High / Medium / Low]

#### Overall Case Risk Rating
- **Rating:** [Critical / High / Medium / Low]
- **Basis:** [One-sentence justification]

EXAMPLE
#### Compliance Risks
| Risk | Policy Reference | Severity | Evidence Citation | Recommended Action |
|---|---|---|---|---|
| Repeated conduct documented without formal warning | Disciplinary Policy S4.2 | High | Three incidents noted in file between March and June with no written warning issued | Initiate formal documentation review with the relevant manager |

CONSTRAINTS
Do not assign severity without citing the specific escalation criterion used. Do not recommend termination, suspension, or other employment actions — present risks and evidence only. Do not speculate on outcomes of unresolved items.

Outputs

  • Every risk cites the specific escalation criterion or policy reference used
  • Severity levels match the escalation and severity criteria definitions
  • No employment action recommendations appear — risks and evidence only

Verification: Verify the AI has not assigned severity levels inconsistently across similar findings in different cases

Hybrid

Inputs

  • Risk Highlight Report
  • Case-Engagement Cross-Reference

Prompt

Identify cross-case patterns and escalation triggers

CONTEXT
You will be provided with the risk highlight report covering compliance and practical risks for all reviewed cases, and the case-engagement cross-reference mapping engagement signals to case evidence.

TASK
Analyse all cases together to identify recurring patterns across case types, common escalation triggers, and areas where engagement risk flags reinforce case trends. Produce a pattern analysis highlighting systemic themes rather than individual case details.

OUTPUT FORMAT
Structure the analysis as:

## Cross-Case Patterns
| Pattern | Cases Involved | Frequency | Supporting Evidence |
|---|---|---|---|

## Escalation Triggers
| Trigger | Description | Cases Where Present | Recommended Monitoring Action |
|---|---|---|---|

## Engagement-Linked Themes
| Theme | Engagement Risk Flag | Related Cases | Implication |
|---|---|---|---|

CONSTRAINTS
Do not identify patterns supported by fewer than two cases unless the single case involves a critical-severity risk. Do not attribute intent or motive to any individual or team. Do not recommend specific disciplinary or employment actions.

Outputs

  • Pattern & Escalation Analysis
  • Every pattern is supported by evidence from at least two cases or one critical-severity case
  • Escalation triggers reference specific case evidence and monitoring actions
  • No individual names or personally identifiable details appear in the pattern analysis
Hybrid

Inputs

  • Pattern & Escalation Analysis
  • Individual Case Summaries

Prompt

Produce final case trend summary with next steps

CONTEXT
You will be provided with the pattern and escalation analysis covering all reviewed employee relations cases, and the individual case summaries for reference.

TASK
Produce the final Case Trend Summary document. Consolidate cross-case patterns, escalation points, and risk-informed next-step options into a single document that HR leadership can use for policy review and resource planning.

OUTPUT FORMAT
Structure the document as:

# Case Trend Summary

## Executive Overview
[3-4 sentences summarising the overall case landscape, key patterns, and risk posture]

## Case Volume & Distribution
| Case Type | Count | Open | Closed | Escalated |
|---|---|---|---|---|

## Key Patterns
For each pattern identified:
### [Pattern Name]
- **Description:** [One-sentence summary]
- **Cases Involved:** [Case references]
- **Risk Level:** [Critical / High / Medium / Low]
- **Policy Implications:** [Which policies are affected]

## Escalation Summary
[Numbered list of active or recommended escalations with justification]

## Next-Step Options
[Numbered list of recommended actions for HR leadership, each with a rationale tied to the pattern analysis]

CONSTRAINTS
Do not recommend specific disciplinary actions against individuals. Do not include personally identifiable information. Present trends and options only — final decisions rest with HR leadership and legal counsel.

Outputs

  • Executive overview accurately reflects the aggregate patterns and risk posture
  • Case volume table accounts for every case in the confirmed review scope
  • Every pattern cites specific case evidence and policy implications
  • Next-step options are tied to pattern analysis with clear rationale

Verification: Verify the summary does not contain disciplinary recommendations — it must present trends, risks, and options only

Reference

Guardrails

  • Factual Citation Only — Only cite documented facts from official case files — never let the AI infer undocumented events or assume intent from ambiguous notes.
  • Separate Fact From Interpretation — Keep factual case summaries strictly distinct from risk assessments and recommendations to prevent conflation in downstream decision-making.
  • Policy-Anchored Assessment — Every risk classification must reference a specific policy section or escalation criterion, not general judgment or subjective severity.

Pitfalls

  • Pasting employee relations case files containing protected personal information into a non-approved GenAI tool
  • Treating AI-generated risk classifications as final compliance determinations without HR or legal review of the source findings
  • Allowing the AI to infer undocumented events or assume intent from ambiguous case notes and incomplete records
  • Merging findings from separate cases into a single summary, obscuring individual case boundaries and accountability trails

Definition of Done

  • Every case summary cites the specific source case file and documents facts with verbatim excerpts
  • The Case Trend Summary covers all cases from the confirmed review scope with no omissions
  • Every risk highlight references a specific policy section or escalation criterion from the organisation's framework
  • No hallucinated facts, dates, or policy references appear in any AI-generated output

GET FULL ACCESS

Unlock every step, prompt, and downloadable example — for this playbook and the rest of AGASI AiOS, our GenAI capability framework.

We'll send a magic link — no password needed.

AGASI AiOS · HR09 v1.0 · Apr 7, 2026