Xoxoday uses artificial intelligence across several Empuls features to help organizations run fairer, more consistent, and more effective employee engagement programs. This page explains the principles that govern how AI is designed and deployed, what protections are in place for your employees’ data, and how administrators control AI usage within their organization.Documentation Index
Fetch the complete documentation index at: https://empuls.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
Guiding principles
Xoxoday’s use of AI is governed by three foundational principles: Privacy by Design — AI is architected with privacy embedded across all data flows and processing layers. Privacy protections are not added on top of AI features as an afterthought; they are built into the architecture from the start. Data Minimization — AI processes only the minimum amount of data required to perform explicitly defined tasks. No data is collected or retained beyond what the feature needs. Enterprise Control and Accountability — Your organization retains full control over whether and how AI is used within your Empuls environment. Admins can enable or disable AI features entirely, and disabling AI does not affect core platform functionality or data integrity.How AI is used in Empuls
Writing assistance and award generation
The AI assistant, Em, helps employees and managers craft award citations, recognition messages, and milestone wishes. When you type a prompt, Em generates a suggested message aligned with your organizational values. This saves time and improves message quality — particularly for managers who want to recognize team members in a meaningful way but aren’t sure how to phrase it.Content moderation
Em automatically scans recognition posts and comments for content that is unsafe, discriminatory, or inconsistent with workplace standards. Flagged content is surfaced for review before it appears on the social feed, helping maintain a respectful and inclusive environment without requiring constant manual moderation.Skill mapping
When employees receive recognition, Em analyzes the language in award citations and extracts demonstrated soft skills — such as adaptability, communication, leadership, and critical thinking. These insights are aggregated and visualized in skill reports, giving HR and managers a continuously updated view of workforce capabilities without relying on self-reporting or static assessments.Sentiment analysis in surveys
Em analyzes open-ended survey responses to identify recurring themes, emotional tone, and areas of concern. Rather than requiring HR teams to read thousands of individual responses, Em surfaces patterns and priority areas automatically. All analysis is performed on anonymized, aggregated data.Recognition gap detection and bias prevention
The AI engine monitors recognition patterns across your organization to identify employees who are consistently under-recognized, detect circular recognition patterns (where employees exchange awards with the same small group), and flag potential favoritism. Administrators receive reports on these patterns and can configure governance rules — in either Advisory mode (users are nudged but not blocked) or Restricted mode (the action is blocked) — to promote fair and equitable recognition.Predictive analytics
Em uses engagement and recognition data to forecast attrition risks, identify disengaged cohorts (such as new hires or remote employees), and surface recommendations for improving program effectiveness. These insights are presented in aggregated form to support HR strategy without profiling individual employees.Protection of personal data in AI workflows
All Personally Identifiable Information (PII) — including names, email addresses, employee IDs, and contact details — is removed or anonymized before any data interacts with AI systems. AI models in Empuls operate solely on non-identifiable, sanitized, context-only inputs. Specific protections include:- Your organization’s data is never used to train public or shared AI models
- AI interactions are transactional and non-retentive — data is not stored by AI systems after the interaction concludes
- No cross-tenant data sharing occurs — one organization’s data never informs AI outputs for another
- AI-generated insights are delivered in aggregated or summarized formats to minimize re-identification risk
How admins control AI features
Administrators control AI usage through the platform’s AI settings:Enable or disable AI features
Toggle individual AI capabilities on or off. When AI is disabled at the organization level, no data is transmitted to AI systems.
Governance and human oversight
AI features in Empuls operate under role-based access controls. Human oversight is maintained for all AI-assisted outputs — Em’s suggestions are starting points, not final decisions. Award citations generated by AI require a human to review and send them. Content flagged by moderation is reviewed before action is taken. Xoxoday periodically reviews AI capabilities and controls to address evolving regulatory, security, and ethical requirements. New AI features undergo risk and governance assessment before release.Questions about Empuls’s AI practices? Reach out to the support team at cs@xoxoday.com.