HR Services +

AI Bot

HR Services + AI Bot

HR Services + AI Bot

Saudia Technic – aircraft maintenance company

One-stop-app app where employees can have access to company services as hazard report creation or add an HR request, all supported by AI assistant.

HR Services + AI Bot

Saudia Technic – aircraft maintenance company

One-stop-app app where employees can have access to company services as hazard report creation or add an HR request, all supported by AI assistant.

What?

Cross-platform employee app that covers HR services and uses AI chat-bot

Client

Aircraft maintenance company from Saudi Arabia

For whom?

For the employees of Saudia Technic. The ones who work in the fields (technicians) and the office folks

My contributions

As a Product Designer, I joined a team when they needed to implement an AI-powered HR chat and support AI-powered flow to submit a Hazard Request

Light and dark modes of the app

The idea came while brainstorming how to make the app feel more innovative. The client was happy to let the team implement AI

The idea came while brainstorming how to make the app feel more innovative. The client was happy to let the team implement AI

The Goal

The Goal

What AI promised:

Quick answers

Reduced workload

Fewer repetitive tasks

24/7 availabnility

What AI promised:

Quick answers

Reduced workload

Fewer repetitive tasks

24/7 availabnility

The AI assistant recieved a bulb icon that I made in Figma to be recognizable across the app

The AI assistant recieved a bulb icon that I made in Figma to be recognizable across the app

Design-Development Iterations

Design-Development Iterations

While I knew as designer what I need to achieve for the outcome, it should be corrected by what development team could realisticly serve.

While I knew as designer what I need to achieve for the outcome, it should be corrected by what development team could realisticly serve.

Part of the linear chat flow made in Figma

The AI Weakness

Like most AI bots, ours came with a big limitation: understanding natural language. We trained it on a pile of HR documents. On paper, it “knew” the policies. But in practice? Even with the right information available, it could not comprehend the questions and fall off to the error.

The quick solution was obvious: reduce the bot’s freedom. Yes, it went against the “AI magic” hype. But it guaranteed a smoother process, fewer dead-ends in conversations, and (most importantly) less frustration for employees who just wanted a simple answer without throwing their phone out the window.

Screens of the AI-led chat with examples of predefined options to answer

Not Just AI

At the first use the employee may confirm or reject the usage of AI. They can change this in the settings any moment.

User support

User support

User support

Every chat session has special features to help user navigate through the conversation

Every chat session has special features to help user navigate through the conversation

Every chat session has special features to help user navigate through the conversation

1

Chat can be closed any moment

2

Color coding helps clearly distinguish between AI and employee’s answers

3

Some employee answers are predefined by system for clarity

4

Employee is offered options to provide feedback on AI’s aqnswers, or refresh it to get alternative result

5

AI responses are marked to remind that the conversation doesn’t go with HR department employee

6

Employee can edit AI-generated requests before creating tickets.

7

Some steps are hardcoded to ensure employee can provide or necessary information

1

3

4

7

2

5

6

Color coding helps clearly distinguish between AI and employee’s answers

Chat can be closed any moment

Some employee answers are predefined by system for clarity

Employee is offered options to provide feedback on AI’s aqnswers, or refresh it to get alternative result

Some steps are hardcoded to ensure employee can provide or necessary information

AI responses are marked to remind that the conversation doesn’t go with HR department employee

Employee can edit AI-generated requests before creating tickets.

Color coding helps clearly distinguish between AI and employee’s answers

THE LEAVE REQUEST

THE LEAVE REQUEST

THE LEAVE REQUEST

Bringing Leave Request flow to the one-stop-app

Bringing Leave Request flow to the one-stop-app

Bringing Leave Request flow to the one-stop-app

A cunning copycat

A cunning copycat

A cunning copycat

I had to follow the Leave Request functionality that pre-existed in client’s HR system. But make it more digestable on the UI side for the app users.

I had to follow the Leave Request functionality that pre-existed in client’s HR system. But make it more digestable on the UI side for the app users.

I had to follow the Leave Request functionality that pre-existed in client’s HR system. But make it more digestable on the UI side for the app users.

Client’s internal tool examples

Client’s internal tool examples

Screens of the Leave Request

Part of the Leave Request flow made in Figma

Human language

The statuses in the client’s internal app were misleading, so we introduced a new status system to notify the employee about their Leave Request status.


Unfortunately, it was not possible to reduce the number of statuses or merge some of them. But at least they were not utterly technical.

Admin dashboard

Admin dashboard

Admin dashboard

Another part of backgroud support of the app was creating an Admin dashboard where HR can manage employees’ requests and feedback.

Another part of backgroud support of the app was creating an Admin dashboard where HR can manage employees’ requests and feedback.

Another part of backgroud support of the app was creating an Admin dashboard where HR can manage employees’ requests and feedback.

Screen of the dashboard where HR can manage employees’ feedbacks and requests

Got something on your mind? Let’s talk

Designed by Kateryna Radkovska