EPR — the right fee per SKU and country, not the generic-factor one.

The problem

“We pay EPR with a generic factor because calculating SKU by SKU is impossible.”

EPR is honest in its logic: if you place a product on the market, you finance its end-of-life management. But the implementation is a minefield, because the fee depends on three variables that few organisations cross properly:

1. The packaging BOM per SKU — weight per material (PET, PE, glass, cardboard, aluminium, composites), inks, labels that are not recycled, mechanical separability. This lives in the PLM, in a packaging spreadsheet, or in the head of the R&D manager.
2. The fee matrix of the country's producer responsibility organisation (PRO) — Ecoembes/SCRAP in Spain, Citeo in France, Der Grüne Punkt in Germany, CONAI in Italy. It changes every year, and so do the eco-modulations (bonus for recyclable, surcharge for problematic).
3. The volumes placed on the market per SKU and per country, which come from the ERP — but they are usually aggregated monthly, without the breakdown EPR asks for.

Multiply all of that by fifty active SKUs in ten EU markets and you understand why almost everybody ends up paying with a generic factor. The generic factor is safe against penalties, but it almost always means overpaying, and it forfeits the green bonuses that OEMs and retailers are starting to offer to the supplier who delivers real traceability.

How it fits the IRIS system

iLEAN does not add an EPR system — it crosses the two you already have, for real.

EPR reporting does not fail for lack of packaging software. It fails because the real packaging BOM and the fee matrix per country never meet in time. iLEAN acts as the filler that crosses both, without the packaging manager having to maintain a hundred-column spreadsheet.

Tracer anchors the real BOM to the SKU. Connect captures the changes — from the supplier, from the PRO, from the PLM. The agent applies the matrix country by country and prepares the quarterly report. The person signs.

The three iLEAN pieces applied to EPR:

• Tracer — anchors the exact packaging composition to each SKU: weight per material, inks, labels, separability. If the recycled PE supplier changes, the data is updated per batch; the SKU does not stay stuck with last year's BOM.
• Connect — captures changes where they happen: the supplier email with the recycled material certificate, the PLM modification when R&D redesigns the packaging, the PRO notice when it updates its tariff, the message from the local importer when a category reopens. Everything enters the system at second zero.
• Agent — keeps the matrix of fees and eco-modulators per EU country alive, applies it to the real weight per material of each SKU, calculates the quarterly fee per country, prepares the report in each PRO's format (Ecoembes, Citeo, Der Grüne Punkt, CONAI, etc.) and leaves it ready for signature. It detects significant deltas versus the previous quarter and alerts the packaging manager through their own channel.

And all of it under the three security rings: the calculation happens in ring 3, the cross-validation (does the BOM match the real batch weight? does the new PRO fee apply from the date it said?) happens in ring 2, and the signed report lives in ring 1. Never the other way round.

See the full IRIS architecture →

Before and after

EPR by hand vs. EPR crossed with iLEAN

Aspect	EPR with generic factor + spreadsheet	With iLEAN Tracer + Connect + Agent
Packaging composition	Last year's average per category	Real BOM per SKU, updated per batch
Country fee matrix	PRO PDF downloaded at the start of the year	Live, captured by Connect when it changes
Eco-modulation	Applied only where the manager remembers	Applied automatically per country and material
Declared fee	Conservative (overpays) or generic	Fair per SKU and country, with evidence
PRO audit	Rebuild the quarter's calculations	Dossier per SKU with full traceability
OEM green bonus	Lost — there is no data to justify it	Available, with auditable evidence

Impact estimate

Impact estimate for your plant — to be validated with your numbers.

The block below is an estimate to be validated with the concrete data of your operation. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• Average producer with 50-300 active SKUs in 5-15 EU markets, today declaring EPR with a generic factor or with an outdated packaging BOM.
• Tracer + Connect + Agent pilot on one packaging category and two countries (typically: Spain + France). First expected value in a few weeks.
• Indicative payback between 4 and 9 months, depending on the volume of EPR fees paid with a generic factor and on the eco-modulation available across your SKU matrix.
• Reduction of the packaging manager's manual hours on quarterly reporting and chasing supplier data: ≥ 30 %, with room for more in subsequent years.
• The hard lever combines direct savings (eco-modulation properly applied lowers the real fee) + protection against penalties in a PRO audit + access to the green bonus that OEMs and retailers increasingly offer.

And the packaging manager's reasonable doubt

“What if the agent gets the eco-modulation wrong and that becomes a problem with the PRO?” — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely recontextualises a piece of data (read the PRO tariff, apply the weight per material, calculate the fee), the best models brought error below 1.5 % ^[1]. And even so, the report does not go out on its own: the agent prepares it, the responsible person reviews and signs it. The three rings are there for exactly this.

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

Keep reading: EUDR deforestation · Conflict minerals 3TG · CBAM carbon border