Australian client
Meet the HR Assistant That Deflects 80% of Repetitive Inquiries — and Gives HR Time to Lead Again
Meet the HR Assistant That Deflects 80% of Repetitive Inquiries — and Gives HR Time to Lead Again
2025 | PH
MY ROLE
Full Stack Designer
TEAM
1 Project Manager,
2 Developers.
DELIVERABLES
Reframed Problem Statement,
Chatbot Scope & Intent Map (High-Level),
Security & Compliance Protocol Summary,
LLM Architecture Diagram,
Action/Transaction Blueprint,
Failure & Fallback Logic Flowchart,
Rich Media UI Element Library
TIMELINE
1 month
How to change my Personal Information?
Show me the remote work policy
When is my next performance review?
How much accrual do I have left?
CONSTRAINTS
API Rate Limits
HRIS API
Data Retrieval (RAG)
CLIENT OVERVIEW
Turning a Trusted Platform into a Smarter HR Experience
This Australian client has been one of our company’s most loyal partners over the years. Our collaboration began with the creation of “Love Work,” a human resource platform designed to simplify HR processes and improve employee engagement within their organization.
After successfully adopting the platform, the client came back with a new challenge — they wanted to introduce an AI-powered chatbot that could handle Level 1 HR inquiries such as leave requests, policy clarifications, and basic employee concerns. Their goal was to reduce repetitive manual tasks for the HR team while giving employees quick, round-the-clock support.
This addition wasn’t just an upgrade — it was a step forward in making their HR operations more intelligent, conversational, and human-centered.
PROBLEM
When People Wait, Productivity Suffers
HR Managers spend 60% of their time answering repetitive questions (e.g., "What's my PTO balance?", "How do I change my emergency contact?"), leading to burnout and delayed strategic work. Employees wait hours or days for critical information.
REFRAME PROBLEM
Turning Everyday Questions into Intelligent Interactions
How can we design a Trusted, Compliant, and Instant information retrieval and action system that deflects 80% of Level 1 HR inquiries while providing the HR Manager with actionable insights?
USER PROFILE AND NEEDS
We focus on two primary personas to ensure a holistic design:
Persona A: The Employee (End-User)
TRAITS
NEEDS
GOAL WITH ASK HR ASSISTANT
Anxiety Level
High (especially around pay, benefits, and sensitive issues).
Instant, accurate answers without talking to a human, and a clear path for complex needs.
Key Use Cases
Check PTO balance. 2. Get policy document link. 3. Submit a minor change request (e.g., direct deposit update).
Efficiency and Discretion: Get the task done in under 60 seconds.
Persona B: The HR Manager (Primary Stakeholder)
TRAITS
NEEDS
GOAL WITH ASK HR ASSISTANT
Anxiety Level
High (due to compliance risk, security, and workload).
Security, compliance, and time savings. Must trust the bot's answers are correct and that sensitive data is protected.
Key Use Cases
Monitor high-volume questions. 2. Be alerted for urgent escalations. 3. Generate a report on policy engagement.
Strategic Focus: Free up time for complex employee relations, recruitment, and culture initiatives.
LLM IMPLEMENATION
The LLM Architecture & Enterprise Guardrails
To meet the high-stakes requirements of HR, the LLM implementation is heavily customized:
LLM COMPONENTS
HR MANAGER ASSURANCE
EMPLOYEE EXPERIENCE
Retrieval-Augmented Generation (RAG)
Compliance: The RAG system is trained only on the official, timestamped Love Work documents. This prevents "hallucination" and ensures all answers cite the current, auditable source.
Accuracy: The confidence in the answer is high because it's directly from the company's official policy
Role-Based Security Layer
Security: Before fetching any personalized data, the system verifies the user's role and authorization level (e.g., Managers can't see peer salaries; Employees can only see their own accruals).
Privacy: Employees feel secure knowing the bot won't accidentally share their personal information (e.g., health enrollment data) with unauthorized users.
Orchestration & Action
Efficiency: Bot is configured to use APIs for simple actions (e.g., checking a PTO balance, initiating an address change workflow) instead of just retrieving policy text.
Utility: Turns a conversation into a transaction. Asking "Update my address" results in a guided form within the chat, not a link to another page.
THE BREAKTHROUGH
UI/UX Design Process & Solutions
The design process prioritized Clarity, Trust, and Guided Action.
The data shown in videos are for placeholder purposes only and is intended for prototype demonstration.
Phase 1: Conversational Mapping & Scripting (UX Writing)
Define the "Fall-Forward" Strategy: Instead of dead-end "I don't understand" responses, we designed the Fallback to guide the user.
Sometimes, users don’t know exactly what to ask — and that’s okay.
Ask HR Assistant listens, learns, and responds with your ask/s.
It’s like talking to someone who just gets it.
Because true intelligence means keeping the conversation going.
The Tone & Empathy Protocol: The persona is professional but warm. During sensitive queries, the bot's response shifts from instructional to empathetic and immediately offers an Escalation Button to a human HR representative.
Sometimes, an answer isn’t enough — it needs empathy.
When employees search for sensitive topics like “grief leave,” the chatbot recognizes the moment.
Its tone softens, its language becomes more human, and an Escalation Button appears for direct connection with HR.
Because true intelligence understands not just what’s asked — but what’s felt.
Phase 2: Interface Design (UI Solutions)
The conversation starts with clarity.
Instead of typing, users tap — guided by rich, intuitive buttons for the first responses.
It’s faster, simpler, and beautifully human.
Trust is built through transparency.
To help employees feel confident in every policy answer, we added small, tappable source tags.
It keeps the conversation credible — no second-guessing, no misinformation.
Every conversation tells a story.
And now, HR Managers can see that story unfold — instantly.
The Manager Dashboard brings everything into focus: the Top 10 Unanswered Questions and Escalation Volume, all displayed in a beautifully simple, real-time view.
Because when insight feels this natural, you’re not managing anymore — you’re leading.
THE OUTCOME
A Smarter Way to Support People.
Ask HR Assistant streamlined how information flows — cutting resolution times from hours to seconds and empowering HR teams to focus on people, not processes.
Faster Resolutions
Reducing time spent on repetitive inquiries and improving response efficiency.
before
4 hours
After
10 seconds
99% time reduction
Freeing employees from long waits and delivering answers in seconds.
Time Saved
Automating FAQs and streamlining workflows to reclaim valuable hours.
before
20hr/week
After
16hr/week
80% of repetitive inquiries deflected automatically
Allowing HR Managers to focus on strategic, high-impact initiatives.
Higher Satisfaction
Enhancing employee confidence through accurate, timely responses.
before
7/10 score
After
9/10 score
+2 satisfaction score
Building trust and creating a more connected employee experience.
Ask HR Assistant demonstrated that by grounding an LLM in a secure, compliant enterprise environment and applying human-centered design principles—especially guided action and transparency—you can create an AI solution that radically improves efficiency for both the employee and the HR manager, transforming HR from an administrative bottleneck into a responsive service.
Let's be friends!
Let's be friends!
Made with coffee after gym.
© 2025