happy-technologies-llc/hrsd-sentiment-analysis

Employee Sentiment Analysis

Overview

This skill analyzes employee sentiment across HR Service Delivery touchpoints including cases, surveys, chat interactions, and feedback. It helps you:

• Assess sentiment polarity (positive, neutral, negative) from HR case communications
• Aggregate survey responses and satisfaction scores for trend analysis
• Identify flight risk indicators based on case patterns and sentiment decline
• Flag cases with escalating negativity for proactive manager outreach
• Track department-level and organization-wide sentiment trends over time
• Correlate sentiment signals with employee lifecycle events (tenure, role changes, reviews)

When to use: When HR leaders need visibility into employee satisfaction trends, when agents need to prioritize cases with negative sentiment, or when identifying employees at risk of attrition.

Prerequisites

• Roles: sn_hr_core.manager, sn_hr_core.case_reader, or sn_hr_core.admin
• Plugins: com.sn_hr_service_delivery (HR Service Delivery), com.glide.assessment (Assessment/Survey)
• Access: Read access to sn_hr_core_case, sn_hr_core_profile, sys_journal_field, interaction, asmt_assessment_instance, and survey_response
• Knowledge: Understanding of your organization's survey cadence, HR case types, and attrition indicators

Procedure

Step 1: Retrieve HR Cases for Sentiment Analysis

Fetch recent HR cases for the target scope (individual, department, or organization).

Using MCP (Individual Employee):

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_case
  query: subject_person=[employee_sys_id]^ORDERBYDESCopened_at
  fields: sys_id,number,short_description,description,state,priority,hr_service,opened_at,closed_at,contact_type,resolution_code,satisfaction_rating,reopened,reopen_count
  limit: 20

Using MCP (Department-wide):

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_case
  query: subject_person.department=[department_sys_id]^opened_at>=javascript:gs.beginningOfLast90Days()^ORDERBYDESCopened_at
  fields: sys_id,number,short_description,state,priority,hr_service,opened_at,subject_person,satisfaction_rating,contact_type
  limit: 100

Using REST API:

GET /api/now/table/sn_hr_core_case?sysparm_query=subject_person.department=[department_sys_id]^opened_at>=javascript:gs.beginningOfLast90Days()^ORDERBYDESCopened_at&sysparm_fields=sys_id,number,short_description,state,priority,hr_service,opened_at,subject_person,satisfaction_rating,contact_type&sysparm_display_value=true&sysparm_limit=100

Step 2: Extract Case Communications for Sentiment Scoring

Pull work notes, comments, and additional details from cases to analyze language and tone.

Using MCP:

Tool: SN-Query-Table
Parameters:
  table_name: sys_journal_field
  query: element_id=[case_sys_id]^element=comments^ORelement=work_notes^ORDERBYsys_created_on
  fields: sys_id,element,value,sys_created_on,sys_created_by,element_id
  limit: 50

Using REST API:

GET /api/now/table/sys_journal_field?sysparm_query=element_id=[case_sys_id]^element=comments^ORelement=work_notes^ORDERBYsys_created_on&sysparm_fields=sys_id,element,value,sys_created_on,sys_created_by,element_id&sysparm_limit=50

Step 3: Retrieve Chat and Interaction History

Fetch interaction records to analyze sentiment from live conversations.

Using MCP:

Tool: SN-Query-Table
Parameters:
  table_name: interaction
  query: opened_for=[employee_sys_id]^ORDERBYDESCopened_at
  fields: sys_id,number,type,channel,state,opened_at,closed_at,short_description,wrap_up_comment,satisfaction,direction
  limit: 25

Using REST API:

GET /api/now/table/interaction?sysparm_query=opened_for=[employee_sys_id]^ORDERBYDESCopened_at&sysparm_fields=sys_id,number,type,channel,state,opened_at,closed_at,short_description,wrap_up_comment,satisfaction,direction&sysparm_display_value=true&sysparm_limit=25

Step 4: Retrieve Survey and Assessment Responses

Pull employee survey responses for structured satisfaction data.

Using MCP (Assessment Instances):

Tool: SN-Query-Table
Parameters:
  table_name: asmt_assessment_instance
  query: user=[employee_sys_id]^metric_type.nameLIKEemployee^ORmetric_type.nameLIKEHR^ORDERBYDESCtaken_on
  fields: sys_id,metric_type,taken_on,percent,state,user,category_scores
  limit: 10

Using MCP (Individual Question Responses):

Tool: SN-Query-Table
Parameters:
  table_name: asmt_assessment_instance_question
  query: instance.user=[employee_sys_id]^instance.metric_type.nameLIKEemployee
  fields: sys_id,instance,metric,value,string_value,actual_value,category
  limit: 50

Using REST API:

GET /api/now/table/asmt_assessment_instance?sysparm_query=user=[employee_sys_id]^metric_type.nameLIKEemployee^ORDERBYDESCtaken_on&sysparm_fields=sys_id,metric_type,taken_on,percent,state,user,category_scores&sysparm_display_value=true&sysparm_limit=10

Step 5: Fetch Employee Profile and Lifecycle Context

Retrieve the employee's HR profile to correlate sentiment with lifecycle events.

Using MCP:

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_profile
  query: user=[employee_sys_id]
  fields: sys_id,user,department,location,employment_type,hire_date,manager,job_title,cost_center,last_review_date,last_promotion_date,years_in_role,employee_type
  limit: 1

Using REST API:

GET /api/now/table/sn_hr_core_profile?sysparm_query=user=[employee_sys_id]&sysparm_fields=sys_id,user,department,location,employment_type,hire_date,manager,job_title,cost_center,last_review_date,last_promotion_date,years_in_role&sysparm_display_value=true&sysparm_limit=1

Step 6: Analyze Case Type Patterns

Examine the types of cases filed to identify concerning patterns.

Using MCP:

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_case
  query: subject_person=[employee_sys_id]^opened_at>=javascript:gs.beginningOfLast12Months()
  fields: sys_id,hr_service,hr_service.name,opened_at,state,priority
  limit: 50

Map case types to sentiment-relevant categories:

=== CASE TYPE PATTERN ANALYSIS ===

High-Risk Case Types (negative sentiment indicators):
- Workplace Complaint / Grievance
- Accommodation Request (especially repeated)
- Policy Dispute / Exception Request
- Manager Escalation
- Exit / Separation Inquiry

Neutral Case Types:
- Benefits Enrollment / Change
- Address / Personal Info Update
- Payroll Inquiry
- General HR Question

Positive Case Types:
- Internal Mobility / Transfer Request
- Training / Development Request
- Recognition Nomination
- Promotion Processing

Step 7: Compile Sentiment Assessment

Assemble findings into a structured sentiment report:

=== EMPLOYEE SENTIMENT ASSESSMENT ===

Employee: Jane Smith | Engineering | Senior Software Engineer
Tenure: 4.8 years | Last Promotion: 18 months ago
Manager: Bob Johnson

--- Overall Sentiment Score ---
Composite Score: 42/100 (Declining - was 68/100 six months ago)
Trend: Negative trajectory over 3 months
Risk Level: ELEVATED

--- Case Sentiment Breakdown ---
Total Cases (12 months): 7
| Period        | Cases | Avg Sentiment | Dominant Tone     |
|---------------|-------|---------------|-------------------|
| Last 30 days  | 3     | Negative      | Frustrated        |
| 30-90 days    | 2     | Neutral       | Matter-of-fact    |
| 90-365 days   | 2     | Positive      | Appreciative      |

--- Survey Scores ---
| Survey                  | Date       | Score | Org Average |
|-------------------------|------------|-------|-------------|
| Q1 Engagement Pulse     | 2026-03-01 | 3.2/5 | 4.1/5       |
| Q4 Annual Engagement    | 2025-12-15 | 3.8/5 | 4.0/5       |
| Q3 Engagement Pulse     | 2025-09-01 | 4.2/5 | 4.1/5       |

--- Flight Risk Indicators ---
[!] Case volume increasing (3 cases in 30 days vs 2 in prior 60)
[!] Engagement survey score dropped 1.0 point in 6 months
[!] Filed policy exception request (potential dissatisfaction)
[!] No promotion in 18 months (above team average of 14 months)
[ ] No internal mobility applications detected
[ ] No exit-related case types filed

--- Sentiment Signals from Communications ---
- Case HRC0012456: "I've raised this issue multiple times..."
  Signal: Repeat frustration, unresolved concern
- Case HRC0012501: "The process seems unnecessarily complicated"
  Signal: Process dissatisfaction
- Case HRC0012523: "I need to understand my options"
  Signal: Ambiguous; may indicate exploration of alternatives

--- Recommended Actions ---
1. IMMEDIATE: Flag for manager 1-on-1 conversation
2. SHORT-TERM: Resolve open cases within SLA to rebuild trust
3. MEDIUM-TERM: Career development conversation recommended
4. TRACKING: Add to monthly sentiment watch list

Step 8: Flag Case for Manager Attention

If sentiment analysis reveals risk, add a work note and flag for review.

Using MCP:

Tool: SN-Add-Work-Notes
Parameters:
  table_name: sn_hr_core_case
  sys_id: [latest_case_sys_id]
  work_notes: "SENTIMENT ALERT: Employee shows declining sentiment trend (score 42/100, down from 68). Flight risk indicators detected: increasing case volume, declining survey scores, no recent career advancement. Recommend manager outreach and career development discussion."

Tool Usage

MCP Tools Reference

| Tool | When to Use | |------|-------------| | SN-Query-Table | Query cases, profiles, surveys, interactions, journal entries | | SN-NL-Search | Natural language search for cases with specific sentiment keywords | | SN-Get-Record | Retrieve detailed single records for deep-dive analysis | | SN-Add-Work-Notes | Document sentiment findings on case records | | SN-Update-Record | Update profile risk indicators or sentiment scores |

REST API Reference

| Endpoint | Method | Purpose | |----------|--------|---------| | /api/now/table/sn_hr_core_case | GET | HR cases for sentiment extraction | | /api/now/table/sys_journal_field | GET | Case communications and comments | | /api/now/table/interaction | GET | Chat and phone interaction history | | /api/now/table/asmt_assessment_instance | GET | Survey/assessment response data | | /api/now/table/asmt_assessment_instance_question | GET | Individual survey question responses | | /api/now/table/sn_hr_core_profile | GET | Employee lifecycle context | | /api/now/table/sn_hr_le_case_type | GET | Case type classification |

Best Practices

• Use multiple signals: Never rely on a single data point; combine case sentiment, survey scores, and interaction tone for accuracy
• Establish baselines: Compare individual sentiment against department and organization averages for meaningful context
• Track trends, not snapshots: A single negative case is less concerning than a sustained declining trend
• Respect privacy: Sentiment data is sensitive; share only with authorized HR personnel and the employee's direct chain
• Avoid bias in scoring: Focus on observable patterns (case frequency, survey scores) rather than subjective interpretation of isolated comments
• Calibrate flight risk thresholds: Work with HR leadership to define what score thresholds trigger specific interventions
• Consider lifecycle context: New hires in their first 90 days may show different patterns than tenured employees; adjust analysis accordingly
• Document methodology: Record the scoring criteria and data sources used so results are reproducible and auditable

Troubleshooting

"Survey responses not found"

Cause: Surveys may use a different assessment type name or may be in a separate survey plugin Solution: Query asmt_metric_type to list available assessment types. Also check survey_response table as an alternative storage location

"Journal entries return too much data"

Cause: High-volume cases may have hundreds of journal entries Solution: Filter by date range using sys_created_on>=javascript:gs.daysAgo(90) and limit to element=comments for employee-facing communications

"Satisfaction rating field is empty"

Cause: Not all case types collect CSAT; it may be captured post-closure Solution: Check sn_hr_core_case for close_notes and check if a separate satisfaction survey is triggered on case closure via asmt_assessment_instance

"Department-level query times out"

Cause: Large departments with many cases over 12 months can exceed query limits Solution: Break the query into monthly chunks or use sysparm_query=subject_person.department=[id]^opened_at>=YYYY-MM-DD^opened_at<=YYYY-MM-DD with pagination

Examples

Example 1: Individual Employee Sentiment Check

Input: "Analyze sentiment for employee Jane Smith over the past 6 months"

Steps:

1. Look up employee sys_id from sn_hr_core_profile
2. Query cases, interactions, and surveys
3. Score sentiment from communications
4. Compile individual sentiment report

Example 2: Department Sentiment Dashboard

Input: "Show sentiment trends for the Engineering department"

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_case
  query: subject_person.department.name=Engineering^opened_at>=javascript:gs.beginningOfLast12Months()
  fields: number,subject_person,hr_service,opened_at,satisfaction_rating,state,priority
  limit: 200

Aggregate by month and case type to produce trend visualization data.

Example 3: Flight Risk Report for HR Leadership

Input: "Identify employees with highest flight risk indicators"

Tool: SN-Query-Table
Parameters:
  table_name: sn_hr_core_case
  query: hr_service.nameLIKEexit^ORhr_service.nameLIKEseparation^ORhr_service.nameLIKEresignation^opened_at>=javascript:gs.daysAgo(90)
  fields: number,subject_person,hr_service,opened_at,state,subject_person.department
  limit: 50

Cross-reference with declining survey scores and increasing case volumes to prioritize the list.

Related Skills

• hrsd/case-summarization - Detailed case context for sentiment-flagged cases
• hrsd/chat-reply-recommendation - Adjust reply tone based on sentiment analysis
• hrsd/persona-assistant - Personalized support for at-risk employees
• reporting/trend-analysis - Visualization of sentiment trends over time
• reporting/executive-dashboard - Executive-level sentiment metrics