Manufacturing·Avícola Quintero

From transcribing every order five times to a plant that records it once.

How Avícola Quintero is replacing WhatsApp, the dispatch sheet, the delivery note and the weighing table with a system that follows each order from the moment it comes in until it leaves on the truck.

Client
Avícola Quintero
Industry
Manufacturing
Size
200 to 550 delivery notes a day · 25,000 to 75,000 chickens a day · 6 roles touching the same order
Counterpart
Cristian Quintero · IT
Duration
4 days to Demo Day · live operation from October 2026
Service type
Custom software: management control system order → picking → weighing → delivery note, with a scale tablet, end-of-day close and reports
5 → 1transcriptions per order
550delivery notes on a peak-season day
4 daysfrom kickoff to Demo Day with the client

The problem

Avícola Quintero is a family business that buys 25,000 to 75,000 chickens a day, sends them to slaughter and, at its plant, sorts them by weight, cuts, packs and ships them to rotisserie restaurants, restaurants and distributors. It moves close to USD 2.5 million a month (COP 10 billion) on margins of USD 0.08 to 0.10 per kilo: it lives on volume and on not giving away grams.

Every order came in over WhatsApp before 7 a.m., and from there someone copied it into the dispatch sheet in Excel, then onto the delivery note, then into the weighing table and finally onto the final dispatch sheet. The same information was transcribed at least five times with nothing guaranteeing the five versions matched.

The cost was not in the hours. It was in the product:

  • One day 6,000 chickens went missing and nobody could say at which stage they were lost.
  • A delivery note said 2,000 chickens and 3,000 went out: over-shipment that was never invoiced.
  • The founder tried to control overweight with his own Excel and could not keep it up: there was no device at the weighing station.
  • Receivables were reconciled by hand every night, starting at 9 p.m.

What we built

We built a management control system where the order is entered once and every stage inherits it. The administrator publishes the day's prices per channel (wholesale, rotisserie, restaurant, retail) and they are frozen; each order takes its customer's price and moves through four states: received, picked, weighed, shipped.

  1. 01
    Line-item orders: average weight per unit converts units to kilos, price inherited from the channel or set per customer, edits allowed until weighing and additions after, with a history of who changed what and when.
  2. 02
    Picking and crate tags: per-order cards with the plant total to pick per product, and crate tags printed six per sheet from what is already recorded, not retyped.
  3. 03
    Scale on a tablet: weighing per crate with the tare deducted, comparison against the order with a configurable tolerance, no-shrink products exempt and automatic advance to the next order. The production role never sees prices.
  4. 04
    The plant cameras, wired to the order: we connected the plant's camera system to measure how long each person takes at sorting, cutting, packing and weighing; that data exposes the constraints and bottlenecks of the dispatch.
  5. 05
    Delivery note and end-of-day close: the delivery note is issued on the scale's kilos with its own sequence number and in two copies; the close makes the day read-only and exports the Excel of prices, dispatch sheet, line items and weighings.

All the business logic lives in the database: the browser cannot write an order, only call the functions that enforce the rules. Data is protected by role and every change lands in an event log. The system is deployed to production with 52 integration tests; the real accounts, the customer base and the first price list load in October 2026, when live operation starts.

The result

01

The order is captured once

From WhatsApp into the system, one time. Dispatch sheet, picking, delivery note and weighing are generated from that record, so they can no longer contradict each other.

02

Weighing is controlled where it happens

The scale tablet shows what was weighed against what was ordered and flags anything over tolerance. It is the control the founder could not sustain in Excel because he had no device on the plant floor.

03

What ships is what the delivery note says

The delivery note is issued on the scale's actual kilos, not on the quantity ordered. It closes the door on un-invoiced over-shipment.

04

Bottlenecks in plain sight

With camera timings per person and per stage, the founder sees where dispatch gets stuck each morning instead of guessing it from the office.

05

A Demo Day, not a PDF

Four days after kickoff, the client tested the system with his own orders and asked for ten adjustments; they were built the same day. Shrink and time figures will be measured once live operation starts.

The Forabi Method

  1. 01
    Free consultation

    Two discovery meetings and three recordings of the team walking through the manual process. We measured volume, roles and where product was getting lost.

  2. 02
    Design

    Order → picking → weighing → delivery note model, with the tolerance, tare and daily-price rules approved by the founder.

  3. 03
    Build

    Business rules as database functions, admin and production roles, scale tablet, printable delivery note, end-of-day close and reports.

  4. 04
    Demo Day

    The client ran the system with the dispatch lead and asked for the real-life adjustments: edits until weighing, crates with tare, manual price, two copies of the delivery note. They were built that same day.

  5. 05
    Adoption

    Testing with the plant team from October 1 and live operation from October 8; ongoing support after that.

Is your operation growing faster than your tools can measure it?

Book your free consultation
Avícola Quintero · Case study · forabi.ai