Vending Central Controller

Kiosoft

Sept-Dec. 2026

Visualizing machine health to drive sales teams

Design: Emily Chan

Design Lead: Alex Rose Henry

Dev Lead: Niko Fu

The Problem

A large learning curve for stakeholders

Created by convoluted flows, buried buttons, and unclear copy

Operational costs directly impact revenue

Longer service visits = fewer machines selling

Longer service visits = fewer machines selling

+Unappealing interface = marketing + pitch difficulty

+Unappealing interface = marketing + pitch difficulty

+High learning curve = higher support cost

+High learning curve = higher support cost

Users of the Central Controller:

Operators

  • Restock machines + diagnose errors

  • Configure products + customer-facing content

Sales Team

  • pitch vending hardware + software package

  • Onboarding + tutorials for new clients

When I joined, feature specifications were mostly complete.

My responsibility was translating them into interfaces, and bringing up any gaps. 

A design system for large tablets

I started with creating a design system in Kiosoft's design language for their upgraded 19'' tablets

Getting used to the scale and sketching out touchpoints based on ADA

Reviewing international kiosk competitors (touchscreen vending is more common in Asia!)

Discussing device lag and navigation considerations

A foundation for Kiosoft's design of large tablet interfaces

Next, use the components to consolidate user flows

The amount of tabs + buttons is overwhelming.

Currently subpages edit individual information, yet they overlap.

stock

Price

Image, stock, price

Organizing tasks by frequency and urgency

What I found:

Action Type

Image/Price edits

Frequency

Monthly

Needs

Consolidate

Action Type

  1. Stock Refills

Stock Refills ‎

  1. Image/Price edits

Frequency

Daily

Monthly

Needs

Separate, fast workflow

Consolidate

I cut down on subpages by consolidating monthly updates in one page, and daily updates in another.

Next, the buttons.

Revealing the hidden rules to refill stock

A consistent mental model is needed.

Creating clear progression and copy

  1. Remove hidden interactions

⤷ Force users to select a slot or tray before an action

⤷ Force users to select a slot or tray before an action

  1. Consolidating "load tray" and "fill"

⤷ Introducing the term "autofill", which applies to any selection

  1. Consolidating "batch revise" and *long press*

⤷ Introducing "revise" to always open the same modal

Input from operators

Original

2 Steps

I found it strange "batch revision" was required to enter "edit" mode

My Ideal

1 Step

Since stock refills happened so often, I thought it'd be better to remove the preliminary step and change the copy to "edit'

Final

2 Step based on competitors

Operators service multiple vending machines from different companies daily. They preferred the 2 step out of muscle memory, with "revise" being the common language in competitor software. "Edit" replaced "batch revise" as the entry to edit mode.

The impact of the new mental model

A reusable slot/tray selection flow that was able to be reused for:

  • Phase 1: Stock refill, image/price updates, slot capacity edits, slot testing

  • Future phases: slot splitting/merging, machine calibration

Final Designs

Visualizing machine health with Central Controller

Workflows made for day-to-day, and month-to-month tasks

Diagnose connection problems for new vending technology

Maintain company partnerships with ads

Reflection

Hybrid design is a bridge

Sometimes we think something is intuitive but for legacy systems the old fashioned way becomes the industry standard so design with that in mind and always clarify.

Working with devs feedback helped me break free from personal bias. Staying flexible made my work realistically helpful. 

Curious and curiouser.

Curious and curiouser.