- Overview of this case study

Operation Management System Transformation

A summary for improvement for this supply chain OMS, which develops the Order management system into the Operation management system.

Back to home

Solo Found designer for end-to-end workflows

Independently researched &Co-design workshops with cross-functional teams Analysis Insights & Information Architecture Mapping service journey map and Shaping solutions and transfer Scoping, Design & Iterations Testing & delivery new features

The central product powering Collectiv Food's operations

I work closely with 2 product owner, 2 Business Analyst and directly with the FE &BE development team

Impacts

B2B products help businesses manage information and optimise their operations through technology, reducing costs and improving efficiency.

By removing inefficiencies and moving towards automation, this system now supports over 250 orders and 23,000 kg of goods per week, and contributes to more than 5% month‑on‑month growth in sales.

- Define

Key Problem Space

For the Planning team

Need a clear, shared delivery schedule to monitor regular order placement.

For the Buyer team

Struggle to identify the best options among hundreds of products.

Inaccurate cost data makes margin calculation unreliable.

For the Operations team

No dedicated tool to manage innovative last‑mile delivery models.

For the Account Management team

Too much repetitive and tedious update work.

For the Finance team

Need to manually enter tens of thousands of invoices every week.

- Define

The Goal

Transforming manual supply chain processes into a semi-automated supply chain workflow

- Research

Methods

Observation

Did a deep dive into each department to understand the specific challenges in their daily workflow, and identified opportunities to integrate our system with different third‑party providers.

Ideas library

Set up an open channel for anyone to surface the issues they face, and collected business needs and challenges from cross‑functional teams.

Interviews

Ran user interviews across teams to dig deeper into their concerns, behaviours, and decision‑making patterns.

Seamless workflows between the customer and internal department are the sustainable way.

So we chose to build our first & foundational AI agent to empower users with automated data analysis and reporting.

Business dashboard showing revenue, deals, customer list, growth chart, and new activity.

Visit our clients to learn the third-party order system they are using

So we chose to build our first & foundational AI agent to empower users with automated data analysis and reporting.

Hear teammate needs, weekly meetings with various teams to find what makes them struggle in the daily work process

So we chose to build our first & foundational AI agent to empower users with automated data analysis and reporting.

Shadowing how last-mile delivery partners adapt our innovative delivery modal

So we chose to build our first & foundational AI agent to empower users with automated data analysis and reporting.

- Research

Scoping: Identifying the root cause in a complex journey

The Purpose?

From Service Journey to Service Blueprint

The challenge?

The entire supply chain cooperation process is complexBusiness processes are not clear and not visualized

Users have to repeat many of the same actions every day
Some data is not stored in OMS

- Research

Developing: Simplifying complexity

For how to achieve large functions, I start by splitting them down into small design demands.

That target each design demand. Conduct a second round of research in more detail and plan specific UX task.

- Devloping

Design Principles

Ease of Operation

Design interactions that feel predictable and help users anticipate what happens next.

Clarity → Information hierarchy

Use strong visual hierarchy so users instantly see what matters most on each screen.

Easy to learn, fast to use

Make flows intuitive to pick up and optimise layouts to improve browsing and task efficiency.

- Devloping

Aesthetic Identity- Design System

Collectiv Food uses Ant Design as its design system, which saves the cost and effort of building one from scratch, so I can focus on developing OMS features.

Ant Design is built for enterprise products and provides stable, reusable components that support fast iteration in complex scenarios

- Delivery

The Final Solution

Key Features

Feature Improvements-Future change in cost

The farmer (producer) changes the costs from time to time. The buying team needs to input the new cost on the exact date. But buying team may have received the new price a week ago, or maybe 3 weeks ago. And these cost changes are received on a daily basis

Deployment 1.0: The user (the Buying team) can add a new cost price, along with a start date. And the old cost will be automatically added to the expiration date

Deployment 2.0: Users (the Selling team) can see all the price changes in the pricing tool. Plus advanced search, the selling team will be able to provide customers with more accurate prices and also ensure that we get a reasonable margin

Happy pass flow - Agreed delivery schedule

Every restaurant should have a delivery schedule (slots) signed with Collectiv Food, and we serve thousands of restaurant locations. We need to keep track of each delivery schedule and continuously monitor whether each order matches the restaurant's delivery slot. The most important thing is that users (the Planning team) use these delivery schedules to plan the delivery routes daily based

Deployment 3.0: Once we onboard a new customer, the account manager can add the delivery schedule for the customer. The process is just simple as adding a meeting in Google calendar.

Deployment 4.0: Restaurants are placing orders every day through various third-party systems or mail, etc. These orders are integrated into our order system and appear on this schedule calendar. Users (the Planning & Operation teams) can easily see if the order matches the agreed delivery schedule, and take action when an Exception order occurs

PS: Ant design doesn't have a calendar component, so I found a Calendar UI library FE resource to be the structure and combined it with small components in the Ant design system. FE engineers and I worked cohesively to fulfill this feature.

Happy pass flow - Bulk update tool

We serve many chain restaurants, which means that one customer could have many restaurants. So when we need to update the price or transaction status of a product. The account manager needs to manually click over 100 times to update all locations, it is very boring repetitive work and can't go wrong

Deployment 3.0: The user (the Account manager team) can bulk update product sell prices for all addresses under the same customer, this can be done easily in one modal two steps.

Deployment 4.0: Users (The account management team) can import all the approved products from one location to another new location, this way, when a customer opens a new Restaurant, the account manager doesn't need to add the approved product one by one to the new Restaurant.

Result 🌱

OMS is being continuously improved, so users no longer need to repeat time‑consuming manual tasks.

For example, an account manager can now update prices in a few minutes instead of several hours, and OMS has become a critical part of Collectiv Food’s daily operations

Over 200 drops per week, a total of around 20000kg can be managed by 5 operation team member

Over 600 locations, 80 accounts can be managed by 5 account managers

Around 200 invoices per week can be managed by 3 finance team members

There are over 3000 products on record and only 12% of our new businesses fail because the buyer team can't find the right product.

See next…

“If an AI Marketing Team Sat Beside You”

Concept→1 case study for AI product

Bagel pay

Digital product all-in-one payment solution