Due to a non-disclosure agreement, information on this project and outcome is limited.

AB InBev

Owning the seller experience
of a delivery platform across
Latin America and South Africa

  • SHIPPED
  • ORDER MANAGEMENT
  • B2B
  • WEB

TL;DR

TaDá Delivery is AB InBev’s global direct-to-consumer delivery platform: one unified experience across Latin America and South Africa. I joined during the MVP pilot, before launch, and owned the UX of the B2B tool where store managers receive D2C orders and run daily operations. I supported the launch with the existing designs, then we shipped the next version in 3 markets. New components from this work were later adopted into the company’s global design system.

ROLE
UX Designer consultant, joined during the MVP pilot and owned the seller platform experience.
TEAM
4 Product Managers
1 Lead Service Designer
1 Sr. Product Designer (consumer-side focus, supporting)
5 UX Designers (working as a design studio)
2 Lead Engineers
+10 Software Engineers
TIMELINE
Aug 2022 → Feb 2023
COMPANY
AB InBev, BEES
BRAND
TaDá Delivery
TOOLS
Figma, Sketch, Jira, Confluence, Miro & Slack

Before

Abstract illustration: two SaaS tools, a different brand per country.

Two SaaS tools and a different
brand in each of 13 countries

After

TaDá Delivery Sellers homepage with daily KPIs and an orders board
TaDá Delivery Sellers homepage: daily KPIs and orders grouped by stage.

One brand and one platform across
Latin America and South Africa

Orders overview

Daily KPIs and every order stage in one view, from new to delivering. The kanban board and urgency timers replaced the unclear steps and status mismatches store managers worked around before.

Orders homepage with daily KPI cards and a New / Packing / Delivering kanban board
Orders homepage: daily KPIs and a kanban board with urgency indicators

Order management

One screen, one decision. Store managers see the full order — products, pricing, customer details, and delivery notes — and act without switching between tools.

Order detail screen with product list, pricing, customer information and delivery assignment
Order detail: product list, pricing, customer info, and accept/reject buttons
Products card with stock reminder, price breakdown and accept or reject buttons
Products card: stock reminder, price breakdown, and accept or reject actions

Order breakdown

Before accepting, store managers get a stock reminder, each item with its quantity, and a Cold tag where it applies. The price walks from subtotal to total (discount, taxes, delivery fee), so the decision is made with the full numbers.

Order history

Order history sits collapsed by default. A usability test showed that an open history made product information harder to reach. As a store manager put it in the follow-up interview: “It makes it harder to access the product information about the order.”

Expanded order history timeline from order placed to delivered
Expanded order history: five steps, from order placed to delivered
Delivery address card with a dropdown to assign a delivery driver
Dropdown to assign a delivery driver, next to the delivery address

Selecting delivery driver

When the order is ready, the store manager picks a delivery driver from a dropdown and assigns it. The delivery address and customer location sit right above, so the handoff happens in one place.

Privacy by design

Store managers see what they need to hand over the order: a shortened name, a masked ID and phone number, and the customer’s notes. The rest of the personal data stays out of view.

Customer information card with a shortened name and masked ID and phone number
Delivered order: name shortened, ID and phone number masked

Financial

Where store managers see what they sold, what was deposited
and what is still to receive, by period.

Financial screen with sales, credit, amount to receive, refund breakdown and day-end report
Sales, deposits, refund breakdown, and day-end report
  • Refund by type

    Payments in the BEES ecosystem are complex, so I broke the refund down by type. Store managers see what kind of refund they are getting, not just a total.

    Refund card split into coupons, delivery fee, campaigns and markup

    Refund total split into coupons, delivery fee, campaigns and markup

  • Report by categories

    A closing report has to fit each store’s routine. I let store managers
    choose how it’s grouped.

    Day-end closing report with options per order, per delivery driver or per shift

    Day-end report: choose per order, per driver or per shift, then print

  • Metrics at hand

    I kept the metrics clean and informative: sales with the change from yesterday, total orders, and an eye icon to hide the value. Built for better-informed decisions.

    Compact sales card with change from yesterday Sales card with total orders

    Flexible sales card: total orders and performance vs. the day before

Performance

Store managers see how the store is doing against its goals: preparation and delivery times, rating, acceptance, availability, portfolio and tracking. Green is above the goal, red is below, and a trend line shows the period.

Performance metrics screen with eight KPI cards and trend lines
Performance metrics: eight KPIs against goals, with trend lines
  • Flexible, deeper analysis

    Comparing periods is a must-have for retail, so I let store managers set a custom date range and narrow down exactly the data they are hunting for.

    Custom date range picker with two calendars

    Custom range: two calendars, start and end dates, reset and save

  • Quick filters

    Not every question needs a custom range. Presets get store managers to the answer faster, whatever they are trying to understand.

    Preset period menu: last 7, 30, 60 and 90 days, last year and all

    Presets: last 7, 30, 60 and 90 days, last year, and all

Real-time tracker

A map shows the store, its delivery drivers and the customers waiting, so store managers can follow deliveries as they happen. A notice reminds them that customer data stays inside the company.

Map screen showing the store, delivery drivers and customers, with a data privacy notice
Map view: store, drivers and customers, with a data privacy reminder

Easily check between statuses

The map has a status filter with the same 3 stages as the orders board. Store managers can see where the store, the delivery drivers and the customers are, then plan the next deliveries without leaving the map.

Map status filter open with New, Packing and Delivering
Status filter on the map: New, Packing and Delivering

Account

Store managers manage their own account and their store in one place: personal details, store information, weekly schedule, and, from the side menu, payments, password, security and notifications. Every section has its own Edit button.

A trustworthy, scalable experience

Three kinds of numbers, each with its date: what we measured with store managers while I was on the project, what TaDá Delivery’s head of design reported after, and public AB InBev figures on where the platform went.

Available in 13 countries

  • Mexico
  • Colombia
  • Peru
  • Panama
  • Ecuador
  • Argentina
  • Dominican Republic
  • Honduras
  • El Salvador
  • Bolivia
  • Paraguay
  • Uruguay
  • South Africa

Platform launched in 13 countries in Latin America and South Africa.

  • 300K+Orders per month

    On average, across 60+ major cities (Q2 2023).

    Source
  • 1M+Consumers reached

    By TaDá Delivery in 2024.

    Source
  • +21%Orders

    Growth in 2024 compared with 2023.

    Source

Public figures from AB InBev reports, 2023–2024.

2 store types,
1 platform.

AB InBev-owned stores and external partners, managed by store managers and admins.

  • ~50%Less task time

    About half the time on the main tasks, compared with the two previous SaaS tools.

    Reported by TaDá Delivery’s head of design (June 2023).

  • 4.8/5Average rating

    Clarity and trust, rated by 30 store managers.

    Usability tests (Nov 2022 – Jan 2023).

  • An MVP is a hypothesis,
    so test it like one

    I prepared the test: script, tasks and hypotheses. The team ran it in production, in Mexico City. It gave me the highest fidelity possible and evidence to confirm or refute my hypotheses.

  • Make the problems visible,
    and they get prioritized

    Two 1-hour sessions with every area of the product produced a shared list of 26 inconsistencies. It went into the backlog, and 22 shipped.

  • A solution needs a
    story to get approved

    I used to go straight to the point and skip the context. With two design stakeholders, one with a B2B lens and one with a D2C lens, I learned to prepare the story for each.

Unifying the leading Latin America credit & account limit management, now 70% faster

Customer support teams had to chase 150+ credit limit fields across a mainframe, screen by screen, while handling live...

Read case

Unlocking Australia’s complex geography: from blocked postcodes to 20K+ orders

Navigating Australia’s complex geography, unlocking inaccessible customers and reducing operational waste.

Read case