A personal assistant to ease the burden of navigating work systems

A mobile app that helps employees find the relevant and meaningful information about their job

Job ecosystems are complex and disconnected from each other, especially when rarely used. We wanted to normalize these systems into a single interface by providing an easy to access, personalized experience.

MY ROLE

I helped leadership imagine what this could become, tested our ideas, and worked with engineering to deliver to production.

PLATFORM

Mobile app

TIMEFRAME

2019

RESPONSIBILITIES

UX research
User interviews
Prototyping
UI design
Handoff to engineering
Information Architecture

Unboxing what we don’t know

There wasn’t anything in the market like it, putting a lot of stress on discovery, framing, and exploring the limits of technology. So here was the general approach for unboxing the vague idea into a buildable product.

Approach

Embrace the pain points

The pains are the opportunities acting as the cornerstone for reaching success.

Put on the user’s shoes

Understand the context of the problem by framing it around the user’s life.

Platform constraints

Synthesize the functionality from APIs of the various services into a simplified approach.

Make the complex simple

It needs to be easy to understand and easy to use, using one UI to interact with services from multiple external systems.

High level goals

Quick access

Provide quick access to systems that would otherwise be difficult. Systems would include services around HR, IT, etc.

Alignment

Engage users with items that require action like polls, required reading, etc, bringing employees and the business closer together.

Human

Work tools are painful; create an enjoyable experience by being friendly and inviting.

Extensible

Build for the future - the interface needs to be simple, flexible, and extendable.

KISS it

“Keep It Simple Stupid” — Work systems are already unnecessary complex. One of the critical values was avoiding this natural tendency to solve everything. When needing to add functionality, never sacrifice clarity. This emphasized the need to establish reusable patterns early.

Making it extensible

A chatbox interface (aka Conversational Interface) offered a lot of leeways. As services are added and new commands are introduced, the user experience would stay the same, keeping the learning curve low while capabilities grow.

Personalized delivery

There was also the need to engage the users with notifications, news, required reading, polls, and automated service pulls, allowing for passive browsing, offering value without requiring cognitive energy from the user.

Discovering commands

The biggest challenge with any chatbot is that you have to know what the bot can do. These services would be unique down to user role, permissions, etc. User research showed us organizing commands by service was most natural to the way users would approach exploring a service.

Personality

The emotional connection between the user and the company was crucial for success. Modern workplace apps overlook this. We wanted it to be friendly, warm, and welcoming. The lasting memory after the first experience should be positive. Additionally, by white-labeling the app, the user is brought closer to the company’s brand.

User testing

I performed three rounds of user testing with 7-8 people each time. Things I wanted to understand and iterate on:

Meet your personal assistant

The default view lets the user know what has happened since they were last in the system and pushes essential information necessary for taking action.

Go deep to discover

When users want to run a specific command, they can type in the command directly or find the command by service, making discovery easy.

In conclusion

Results

Approx. 2 external integrations were added every quarter. After 2 years, ten integrations were fully supported, including ADP, Workforce, and Zendesk. Additionally, customers could create custom search terms for their internal file systems.

Next steps

The system is contingent on the company using it to set up and activate the integrations for their employees. Once the system reaches a certain threshold, the discovery flow will need reevaluation to ensure users gain the most value from relevant commands available.

© 2024 ✦ All Rights Reserved