Agave intake traced by photo

The problem

The foundation of all traceability is typed with a driver waiting.

Agave is 60-70% of variable cost and it is the only input whose origin traceability is both mandatory and verifiable by a third party. Intake is therefore where a badly entered figure costs the most: it hits batch cost, raw material inventory and the certification file. It is also the point under the most time pressure, because there is a truck and a driver waiting. So one of two things happens: either the carrier waits 20-45 minutes while somebody types — and that wait gets billed — or it gets signed off in a hurry and origin traceability goes soft. The second option resurfaces months later, when the certifier asks which field the agave in a given batch came from and the paperwork does not add up. By then there is no way to reconstruct it.

• Agave is 60-70% of variable cost and the only input whose origin traceability is both mandatory and verifiable by a third party. Intake is where a badly entered figure costs the most: batch cost, raw material inventory and the certification file.
• The truck arrives with the piñas; the receiving manager has to weigh, check the delivery note, confirm the field is registered with the certifying body, sign, and then enter kilos, price, field and folio line by line into the ERP.
• One of two things happens. Either the carrier waits 20-45 minutes while somebody types, and that wait gets billed, or it gets signed off in a hurry and origin traceability goes soft.
• The second option resurfaces months later, when the certifier asks which field the agave in a given batch came from and the paperwork does not add up.

How it fits the IRIS system

Connect in photo-to-ERP mode — ticket, delivery note and folios, no supplier changes.

Connect in photo-to-ERP mode. The flow:

The photographic evidence stays attached to the intake for good. That is what turns the question the certifier asks months later — which field did this batch come from — into a query instead of a hand reconstruction. And from then on every autoclave load knows which agave it was made from, a thread that runs all the way to the bottle. The driver leaves in minutes and the manager signs on what he saw, not on what he had time to type.

• Photo of the weigh ticket (tare, gross, net) and of the supplier's delivery note.
• Photo of the field folios and plantation registry codes, and of the load identification labels, printed or handwritten.
• The system extracts kilos, supplier, field, folio, agreed price and harvest date, and summarizes on screen.
• The manager validates with one tap.
• Connect posts to the ERP through the API and keeps the photographic evidence attached to the intake for good.

From then on, every autoclave load knows which agave it was made from, and that thread runs all the way to the bottle (case 12).

See the full IRIS architecture →

Before and after

Today's intake versus the captured intake

Aspect	Today	With iLEAN Connect
Typing per truck	20-45 minutes	One validation tap
Carrier waiting time	Billed	Eliminated
Field folios and registry codes	Checked in a hurry	Extracted and cross-checked
Which field a batch came from	Rebuilt at the certifier's visit	A query, with photos
Evidence of the load	A signed ticket in a folder	Photos attached to the intake
Intake and purchasing licenses	The whole weighbridge crew	One, the manager who validates

From 20-45 minutes of typing per truck to one validation tap. From origin traceability rebuilt by hand at the certifier's visit to a file available per intake and per batch, with photos.

Impact estimate

Impact estimate — to be validated with your numbers.

The block below is an estimate to be validated against your plant's actual data. We put it forward so the committee has an order of magnitude; we refine it during the assessment.

• 💰 CFO angle: direct saving on ERP intake/purchasing module licenses, typically −50% to −70%, because only the manager who validates needs one, not the whole weighbridge and warehouse crew. Recurring.
• Estimated payback 4-8 months, depending on deliveries per week in season.
• Carrier waiting time eliminated, truck after truck, which in season is several trucks a day.
• And origin traceability that is complete by construction, which is what holds up when the certifier asks.

Direct saving on ERP intake/purchasing module licenses — typically −50% to −70%, because only the manager who validates needs a license, not the whole weighbridge and warehouse crew — plus carrier waiting time eliminated. Estimated payback 4-8 months depending on deliveries per week in season; the license saving is recurring. *Estimate to be validated.*

And the fair question from the production manager

"Every supplier's delivery note and every field folio look different — does it read them?" — yes, and without asking any supplier to change anything: the model extracts the structure of the ticket, the note and the folio without a per-supplier template, printed or handwritten. It is an anchored task, where the best models drop below 1.5% error [1], and the receiving manager validates on screen before the intake posts to the ERP. A folio that is not on the registered list is flagged before he signs, not discovered at the certifier's visit.

^[1] OpenAI paper "Why Language Models Hallucinate", 2025 — on the reliability of AI in anchored tasks.