‹ All projects
B2B · Tax

From weeks
to minutes.

A small team issuing tax invoices one by one for more than 3,000 franchises. The system existed, it just wasn't built for that volume.

60%
Fewer steps in the flow
~3×
Fewer clicks per operation
1.200×
Faster at high volume
86
SUS in usability research
01Context

The business grew. The product stood still.

Large franchises with high exchange and return volumes had to issue hundreds, sometimes thousands, of tax invoices at the end of each seasonal cycle. A small team issued thousands of invoices one by one, with whole days of manual work piled up at the end of the cycle, right when the team was already worn out from the peak.

On top of that, it was a period of tool migration. Franchises compared the new tool with the old one all the time, and any sign of regression weighed against the whole project.

02Problem

Linear execution in an operation that needed scale.

The initial diagnosis pointed to a "slow system". Up close, the problem was different: the tool demanded manual micro-execution for a repetitive task that could be fully grouped. Invoice-by-invoice issuing, no view of the whole batch, errors that meant reprocessing everything again. In some states, per-consumer issuing rules multiplied the rework even further.

Core insight

The system forced linear execution on a task that was, by nature, high-volume and simultaneous.

03Goal

It wasn't about speeding up issuing. It was about changing its logic.

  • Scalability: remove unnecessary repetition without requiring a specific technical profile
  • Centralization: status, errors and pending items for the whole batch in one place
  • Autonomy with safety: each invoice keeps its own tax individuality, avoiding chain failures
04My role

Discovery, definition and delivery of the whole flow.

I owned the full cycle, from operational immersion to engineering handoff. I worked directly with franchises, regional coordinators and the rollout team.

05Process

Understand what they did, and why they did it that way.

An intensive two-week discovery focused on behavior: how they organized orders before starting, where they made the most mistakes, at what point the process "broke".

01

Immersion with real operators

Direct observation of the flow during a full issuing cycle, interviews with franchises and regional coordinators.

02

Mapping workarounds

Every deviation from the official flow was a sign of where the tool didn't match the reality of the operation.

03

Regulatory alignment by state

Understanding the tax constraints, especially in states with specific rules, before designing anything. What was a legal obligation couldn't be simplified.

04

Redesigning the task architecture

It wasn't a visual redesign. It was reorganizing the structure of the whole flow based on the mapped operational reality.

05

Validation before launch

Pilot with real franchises, feedback collected systematically, adjustments before rolling out to everyone.

06Solution

The interface as a management layer, not an execution layer.

I built it around batch issuing logic that kept each invoice's individual autonomy. Tax compliance guaranteed, without sacrificing efficiency at volume: a consolidated view of the batch, smart multi-select, and granular feedback per invoice, without stalling the rest.

Recording of the flow in the productStore · Records · Invoice details · Issuing

From the franchise validation

"It used to take me three days. With this, we delivered before the end of the month." (Franchise operator, validation pilot)

07Impact

The numbers tell part of the story.

The most relevant impact was behavioral. Operators were freed up for other tasks. Overtime stopped being inevitable at peak. The dependence on a few key people, who held the whole operation together, was gone.

08Learnings

What this project teaches about back-office products.

  • Map real behavior, not the documented process: what people actually do matters more than what they're supposed to do
  • Workarounds are product opportunities disguised as hacks
  • In a regulatory context, simplify everything around the legal obligation, not the obligation itself
  • Scale isn't speed. It's removing unnecessary repetition
  • Complex back-office doesn't need a pretty interface. It needs an interface that respects the cognition of whoever operates it

Final insight

If technology doesn't scale with the business, the problem is a product problem, and it goes beyond the interface.

Next casePix Disputes (MED)

Let's talk.

Happy to go into detail in a conversation.

Talk on LinkedIn or igorlimaddg@gmail.com