Sobrasada IGP Mallorca — you do not touch the curing curve, you watch it better.
The curing of Sobrasada IGP Mallorca depends on a traditional curve of temperature, humidity and ventilation that the master adjusts by eye. Today that reading is aggregated per room and the master's eye is never digitized. iLEAN Edge adds AI vision and per-rack sensors, cross-references them with the master's voice and alerts before the defect is visible — without touching the recipe. The person signs.
The room measures averages — the piece lives on a rack.
Curing Sobrasada IGP Mallorca is one of those traditions that works because somebody has spent decades sharpening their eye. The master curer walks through several times a day, looks at the color of the casing, feels the string, decides whether to open a window or bring the humidity up a point. The room thermo-hygrometer gives an aggregated reading that does not see what the master sees: the corner of the cellar where less air moves, the rack that catches filtered sunlight, the pocket of humidity that forms when fresh product comes in.
The gaps in the classic model:
- The master's eye is never digitized. The day they retire, half the cellar goes with them. And the next master takes years to reach the same level.
- The thermo-hygrometer averages and hides micro-pockets with different conditions. The affected piece is spotted once it is already affected.
- The incident is discovered late. The window left open too long at shift change, the night the extractor did not start — all of that shows up in the finished piece, already as waste or already as a defect in appearance.
- The SCADA only controls the room set point, not the reality rack by rack.
The system works 95% of the time — and the remaining 5% is the difference between a season with waste under control and a season that eats the margin. Nobody wants to touch the recipe. But everybody wants to watch better.
iLEAN Edge does not touch the set point — it seals the crack between the room and the piece.
The problem with traditional curing is not the recipe: the recipe is what makes the product. The problem is that the reading of reality is aggregated per room while the product lives per rack, and the master's knowledge is never captured as a permanent capability. iLEAN acts as the putty that fills that gap without asking you to change anything about the tradition.
Edge sees the rack and the color of the piece. Connect captures the master's voice and the SCADA wherever it lives. The agent cross-references with the expected curve and alerts earlier — without touching the recipe. The person signs — never the other way round.
The iLEAN pieces applied to Sobrasada IGP Mallorca curing:
- Edge — a machine vision terminal (CNN) over specific racks + distributed temperature, humidity and CO2 sensors. It reads casing color, growth of the desired mold, surface appearance. It detects local microclimates the central thermo-hygrometer cannot see. It works with no network — if the cellar loses WiFi, it keeps reading.
- Connect — captures the master's voice as they walk the cellar (“this rack is running behind, bring the humidity up a point”), the SCADA set point, the shift-change reports, the incident an operator jots down in a notebook. And it captures what comes from outside: the day's weather report (relevant in Mallorca, with its winds), an email from a customer with a special order.
- Agent — cross-references the per-rack reading, the master's voice and the SCADA with the expected curve for that curing stage. If it detects a local deviation (a humidity micro-pocket, a rack with insufficient ventilation, a window left open) it alerts the master through whichever channel they use, before the defect is visible. And it keeps learning the master's criteria as a permanent capability.
Traditional cellar with a thermo-hygrometer vs. cellar with iLEAN Edge
| Aspect | Traditional cellar | With iLEAN Edge |
|---|---|---|
| Granularity of the reading | Aggregated per room | Per rack + distributed sensor |
| Local microclimate detection | The master senses it on the walk | Edge measures, the agent cross-references |
| The master's knowledge | In their head | Captured by Connect, linked to the data |
| Window / extractor incident | Discovered in the piece | Detected right then, alert to the master |
| Traditional recipe | Untouched | Untouched — you just watch better |
| Evidence for the IGP / retailer | Manual spreadsheet | Automatic dossier per batch and per cellar |
Impact estimate for your plant — to be validated with your numbers.
The block below is an estimate to be validated with the specific data of your plant. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.
- Traditional or semi-climate-controlled Sobrasada IGP Mallorca cellar, x cellars, a master curer with veteran judgment, waste from irregular curing recorded on a spreadsheet.
- Edge pilot in one cellar (camera over the critical racks + distributed sensors) + Connect capturing the master's voice and the SCADA + an agent cross-referencing against the expected curve. Without touching the recipe. First value within a few weeks.
- Expected reduction in waste from irregular curing above 30% against the baseline measured during the immersion.
- Indicative payback between 4 and 9 months, depending on IGP volume per year and the historical waste percentage per cellar.
- Extra lever (not monetizable in the short term): the master's eye is captured as a permanent capability — the day they retire, their criteria stay inside.
And the master curer's reasonable doubt
“What if the AI gets it wrong and tells me to raise the humidity when it shouldn't?” — hallucination is a problem of free generation, not of anchored tasks. Here the AI only recontextualizes real data (the sensor reading, the image of the rack, the expected curve). In tasks anchored to the source, the best models brought error below 1.5% [1]. And even so, what is critical is never decided alone: the agent alerts, the master decides. Edge does not act on the set point without a human signature. Nothing and nobody touches the traditional recipe without your OK.
[1] OpenAI paper “Why Language Models Hallucinate”, 2025 — on the reliability of AI in anchored tasks.
What people ask about Sobrasada IGP Mallorca curing with AI
What defines Sobrassada de Mallorca IGP and why is the curing curve so critical?
The Sobrassada de Mallorca IGP protects the traditional Mallorcan sausage made with lean pork and pork fat, salt, paprika and spices, stuffed into natural casing and cured under the island's Mediterranean microclimate. The curing curve — temperature, relative humidity and ventilation — is what gives the product its color, its spreadable texture and its characteristic aroma profile. One small deviation (humidity too high, insufficient ventilation) and the piece loses uniformity: it develops case hardening, pockets of separated fat, or growth of unwanted mold. The recipe is traditional; the curve cannot be touched — it can only be watched better.
How is a sobrasada curing cellar controlled today, and where are the gaps?
In traditional cellars, control rests on the master curer's eye: several walks through the cellar a day, a look at the thermo-hygrometer, a decision on whether to open or close a window. Modern cellars have a SCADA with a room set point, but the reading is still aggregated per room — not per rack, not per piece. The gaps: (1) the master's eye is never digitized and is lost the day they retire; (2) the thermo-hygrometer averages the room and hides micro-pockets with different conditions; (3) the incident (a day a window was left open too long) is discovered once the piece is already affected.
How does iLEAN Edge govern the curve without breaking the traditional recipe?
iLEAN Edge does not touch the recipe, does not touch the set point, does not touch the master. What it does is see better: AI vision cameras (CNN) read the color, the growth of the desired mold and the look of the casing on specific racks, not aggregated per room; distributed temperature, humidity and CO2 sensors detect local microclimates the central thermo-hygrometer cannot see; an agent cross-references all of that with the expected curve and alerts the master before the defect is visible — a window left open too long, a rack with less ventilation, a pocket of high humidity. The master decides and signs. The recipe does not change.
How is the master curer's knowledge captured so it does not get lost?
Connect captures the master's voice: when they walk through the cellar and say “this rack is running behind, bring the humidity up a point”, it is recorded and linked to the sensor reading, to the Edge image and to the curing stage. Over time, the agent learns the pattern — what “running behind” means in measurable terms — and leaves it as a permanent capability of the plant. The day the master retires, their criteria stay inside the system. It does not replace them — it gives them a tool so the next master starts out knowing what took this one thirty years to learn.
How much does it cost and when do you see the first saving in a Sobrassada IGP cellar?
The typical pilot covers one Edge with a camera + distributed sensors in one cellar + Connect capturing the master's voice and the SCADA + agents cross-referencing against the expected curve. Moderate initial investment, without touching the recipe or the traditional infrastructure. First value within a few weeks: detection of local microclimates that today only show up in the finished piece. Indicative payback between 4 and 9 months, depending on IGP volume per year and the share of waste from irregular curing. Hard lever: one batch of waste avoided per cellar per season — that pays for the pilot. An estimate to be validated with your numbers.
Tell us your case and in 48h we'll send you the estimated ROI of this AI project for your Sobrassada IGP cellar.
We work on your plant's real data, not ours. Diagnostic with no commitment.
Request estimated ROI in 48h See Iberian