What did the HR Workflow Automation redesign improve?

It replaced technical builder language and the blank-canvas start with HR native prompts, scenario based templates, and AI assisted workflow creation, so HR admins could reach a working automation without dropping off.

What was Vu Nguyen’s role on HR Workflow Automation?

Vu worked as the product designer, reframing the builder around the HR admin’s mental model, defining the template-first entry, and shaping the AI-assisted (Lumi) creation flow.

vu nguyen

HR Workflow Automation

A redesign of the HR workflow builder, reframing it around HR native language, ready made templates, and AI assisted creation so HR admins can automate onboarding, leave, and offboarding without a blank canvas.

Role

Product Designer

The problem

The workflow builder lets HR teams automate the processes they run every day, such as onboarding, offboarding, and leave management. It is the first value moment in the product, so it drives activation, reduces manual HR workload, and shapes long term retention.

But HR admins struggle to create a workflow at all, which blocks them from reaching that core value. The current builder speaks in system oriented language and opens on a blank canvas, a model that does not match how HR admins actually think or work.

Technical terminology

Terms like “trigger” and “action node” are language HR users don’t recognize, increasing cognitive load.

Blank-canvas entry

Starting from nothing gives users no foothold, which causes decision paralysis and high drop-off.

No AI assistance

Users have to already know exactly what they want before they can build it.

Define and remind the team about target user

The primary user is a non-technical HR admin responsible for people operations. They think in everyday scenarios and tasks, not abstract logic flows.

Solution

Translate system language to HR language

Replace system terms with human readable prompts and group them by HR scenario rather than trigger type. This aligns the interface with the user’s mental model, cutting cognitive load and decision time.

Workflow builder using HR-native prompts grouped by scenario — System terms rewritten as plain questions, grouped by HR scenario.

Template first entry

Surface predefined workflow templates — Onboarding, Offboarding, Leave Request — as the primary entry point, with the blank canvas kept as a secondary option. Users start from a familiar, real scenario instead of an abstract canvas, which removes the blank-state problem.

Introduce Template gallery in workflow builder — Introduce template at the screens, where user can easily find templates they need.

AI assisted creation flow

Users describe a workflow in natural language and AI generates a draft to review and edit. This lowers the knowledge barrier for people who don’t know where to start and speeds up setup for common cases.

AI touch point in workflow builder — To streamline creation, the main view introduces AI prompting and templates as primary entry points for instant generation or customization.

AI generating a draft workflow from a natural-language prompt — AI translates natural language descriptions into editable draft workflows, allowing users to accept the results or iterate until it fits their needs.

Outcome & impact

Stakeholders signed off on the solutions across the board. The full redesign went into the backlog since engineering didn't have the bandwidth right now,

Takeaway

This one showed me why it pays to dig past the obvious problem. Instead of patching one broken onboarding flow, I reworked the whole builder to match how HR admins mental model. Turning a single clunky path into three clear ones (templates, AI, or a blank canvas).