Cervejaria Proibida

Context

Cervejaria Proibida runs B2B distribution in southern Brazil — points of sale, bars, restaurants, events. The commercial model historically required an in-person visit from the rep to every new client, with the commercial agreement (volume, terms, conditions) negotiated face to face. The director’s judgment — what to accept, what to reject, how far the price could go — lived in his head and came down case by case, validating or recalibrating what the rep proposed in the field.

It was a functional model for 30 reps. When expansion required doubling to 70, the bottleneck appeared: the director didn’t have time to validate 70 deals a day, and the mandatory in-person visit became an obstacle — travel cost exploded, time lost in transit became unviable, and growth stalled.

The pain

Three problems piling up:

• Travel cost — an in-person visit per deal, on a route spread across the South. The team spent on average R$ 92/day on fuel and tolls per rep. The weekly bill added up to seven thousand reais in the field alone.
• Director’s time — case-by-case validation by phone/WhatsApp, reconstructing context (“which client is this again?”, “what’s their history with us?”). Each validation took 12-18 minutes.
• Inconsistent judgment across reps — without a tool showing each client’s acceptable ceiling/floor, new reps negotiated either too conservatively (losing deals) or too generously (poor margin).

The solution

We built a digital pre-agreement system with codified judgment:

• Unified client registry — history of volume, average ticket, default risk, seasonal events, strategic importance. Visible to the rep before the conversation.
• Parameterized negotiation range — each client got a ceiling/floor calculated from rules (size, history, seasonality). The rep saw on the spot whether the proposal fit or required escalation.
• Asynchronous approval flow — outside the range, it triggers approval to the director with pre-assembled context (client, history, reason for the exception). The director decides in 2 minutes, not 15.
• Agreement formalized remotely — when within range, it closes without involving the director; when approved by exception, likewise. Digital signature.

Rollout ~4 weeks, with sequential onboarding of each region.

The result

Measured after 6 months on the new model:

• R$ 6,450 / week saved on travel (in-person visits cut by ~70% — kept only for new clients and major-account renewals).
• The team scaled from 30 → 70 reps without a proportional increase in the director’s load — his validation time dropped from 4-5h/day to ~45 min/day.
• Average margin stabilized — new reps started closing within the target range, without the swing of “negotiating by instinct.”

How it would run today on F7 KORE

This project was delivered by the team behind F7 KORE, with business rules codified in a structured workflow. On today’s platform, the same process would gain three things:

• The proposal ready before the conversation. The AI agent builds the proposal within the calculated range, with the customer’s history from the ERP.
• Exceptions with context. Outside the range, the approval reaches the director in Teams or Outlook with customer, history and reason already gathered.
• Margin tracked in real time. Leadership asks how the week’s deals are going and gets numbers and context, with the lineage of every data point.