Home

Work

Resume

R.S.

Home

Work

Resume

AI Assistant

Role

UX designer, conversational experience design

Collaboration

Product management, engineering, and service stakeholders

Constraints

Enterprise environment, existing Salesforce chat architecture, engineering limitation

Overview

The TriNet Assistant reduces friction in high-volume HR interactions by helping users get quick answers, complete common tasks, and reach out to live support when needed.

The project focuses on defining realistic user scenarios and conversational flows within an enterprise service environment, supporting payroll, benefits, and general HR use cases.

Role

UX designer, conversational experience design

Collaboration

Product management, engineering, and service stakeholders

Constraints

Enterprise environment, existing Salesforce chat architecture, engineering limitation

Overview

The TriNet Assistant reduces friction in high-volume HR interactions by helping users get quick answers, complete common tasks, and reach out to live support when needed.

The project focuses on defining realistic user scenarios and conversational flows within an enterprise service environment, supporting payroll, benefits, and general HR use cases.

Role

UX designer, conversational experience design

Collaboration

Product management, engineering, and service stakeholders

Constraints

Enterprise environment, existing Salesforce chat architecture, engineering limitation

Overview

The TriNet AI Assistant reduces friction in high-volume HR interactions by helping users get quick answers, complete common tasks, and reach out to live support when needed.

The project focuses on defining realistic user scenarios and conversational flows within an enterprise service environment, supporting payroll, benefits, and general HR use cases.

Problem Snapshot

Today users encounter search experience in three different places, each with its own interface, capabilities, and limitations.

Three tools, three different experiences

Global header

Search bar

Keyword based search in the global header. Returns links and page results

Keyword based search in the global header. Returns links and page results.

Limitations

Limited to navigation, return links, not answers. Users get a list of pages, then have to figure out which one helps. No conversational follow up

Dashboard

DD: AI Search

Natural language available in Dynamic Dashboard. Powerful, but only available in one context.

Limitations

Hidden and contextual: Only available in DD. No access from other parts of the platform and no conversational follow up.

Support Widget

Chatbot

Support focused chat bubble. Escalates to human agents.

Limitations

Support only perception. Users see it as help desk, not assistant. Limited to FAQs. No task completion or cross context awareness.

The Experience

The assistant is designed to feel fast, clear, and reliable inside the existing TriNet service experience. Conversations emphasize clarity, confirmation, and seamless escalation when human support is needed.

Core Flows

The assistant is designed to feel fast, clear, and reliable inside the existing TriNet service experience. Conversations prioritize clarity, confirmation, and smooth escalation when a human is needed.

Quick answers with supporting context
Guided task completion in chat
Clear boundaries and graceful escalation to live agents

Knowledge Base:

Users can ask questions about payroll, benefits, and HR policies and receive clear answers with supporting resources.

Automate Tasks:

Common HR tasks can be completed directly in chat, reducing back-and-forth and time to resolution.

Connect to a live agent:

When a request needs personalized guidance, the assistant helps route users to a specialist without breaking the conversational flow.

Handling out of scope questions:

When a question is not supported, the assistant responds transparently and redirects users toward the best next step.

Impact and Feedback

The TriNet AI Assistant was designed to automate high-frequency payroll inquiries and reduce operational friction across support channels. The system focused on structured, low-complexity workflows ideal for AI-driven resolution.

2700+

Annual routine payroll inquiries targeted

1500+

Direct deposit changes eligible for automation

1200+

W-2 inquiries structured for instant resolution

Design Principles

Task completion over conversation: Prioritized structured workflows that resolved payroll tasks directly rather than generating conversational responses
Trust through clarity: Designed calm, predictable interactions suited for sensitive payroll and benefits use cases.
Escalation by design: Integrated seamless handoffs to live agents to maintain confidence and avoid dead ends.

Back to Selected Works

Thanks for reading!

Thanks for reading!