Serbian exporters preparing for the definitive CBAM period do not need to begin with a multi-year digital transformation. They need to stop losing evidence. With the first annual CBAM declaration for 2026 imports due by 30 September 2027, the most urgent task in late 2026 is to secure the meter histories, electricity contracts, production records, precursor information and calculation assumptions that will be difficult or impossible to recreate later.
Why start with one product family
A 90-day programme should be organised around one outcome: by the end of the period, the factory should be able to reproduce a controlled emissions calculation for a selected product family and trace every material figure back to an identifiable source. That pilot is more valuable than a broad dashboard covering the entire site without clear methodology or ownership. Once the control model works on one route, it can be scaled.
The 90-day plan
Days 1–30: scope and evidence preservation
Days 1 to 30 should define scope and preserve evidence. The company maps exported products to CN codes and producing installations, documents production routes and identifies which emissions components and precursors are relevant. Engineering inventories grid meters, on-site generation, fuel measurements, process submeters and data-retention periods. Procurement collects electricity and material contracts, while sales lists the emissions information already promised to major EU customers.
The same first month should establish a gap register. Typical weaknesses include missing hourly electricity files, unclear generator identity, precursor data held only by a trader, production totals that do not reconcile with warehouse output, uncertain attribute ownership or manual corrections with no approval trail. Each gap should have an owner, deadline and treatment. The objective is not to eliminate every uncertainty immediately but to prevent undocumented assumptions from entering the calculation.
Days 31–60: control interfaces
Days 31 to 60 should build the control interfaces. Meter exports are standardised with consistent identifiers, units and timestamps. Production quantities are reconciled with inventory and scrap. A material genealogy links precursor receipts to product families. Electricity procurement is tested against the intended claim, including generator records, interval data, substitution provisions and origin attributes. Where contracts do not provide necessary evidence rights, amendments or supplier requests should begin before year-end.
This phase should also separate methodologies. The factory establishes distinct calculation templates for CBAM, corporate greenhouse-gas reporting and any product-footprint method used with customers. They can draw from the same source records but must preserve different boundaries and factors. A reconciliation note should explain why outputs differ. That protects the sales team from issuing a corporate Scope 2 result as if it were automatically a CBAM embedded-emissions value.
Days 61–90: a full reporting cycle
Days 61 to 90 should run a full monthly reporting cycle on the selected product family. The team freezes source data, performs energy and material balances, applies the calculation, records manual adjustments and prepares an evidence index. A reviewer who did not build the calculation should be able to follow the trail from reported emissions back to meters, contracts, production records and precursor files. Findings are logged and assigned before the next cycle.
Management should receive a compact dashboard rather than a technical archive. Useful indicators include electricity per tonne, contracted and separately qualified renewable share, precursor coverage by actual verified information, reliance on defaults or estimates, unreconciled balances and open evidence requests. Those KPIs show whether the factory’s verification position is improving and where additional investment or supplier engagement is needed.
What comes after the pilot
Software decisions can follow the pilot rather than precede it. Some plants may need new meters, automated interfaces or a dedicated MRV platform; others can initially use controlled exports from existing energy, ERP and production systems. Buying software before defining the production boundary can digitise the wrong process. The 90-day exercise creates the specification for any later system procurement because it shows which data actually matter and where control failures occur.
The pilot should finish with a management decision on capital and supplier priorities. If the largest uncertainty is missing precursor data, procurement action may create more value than new submetering. If electricity allocation is weak, an additional meter or data interface may be justified. If the calculation is reliable but the customer has not agreed how it will use the result, sales and legal teams may be the next priority. The 90-day programme therefore produces an investment roadmap rather than simply another compliance report.
A second output should be a responsibility matrix covering production, energy, procurement, finance, IT, sustainability and sales. Every recurring dataset needs an owner and deadline, every manual correction needs approval, and every customer release needs a controlled version. Clear ownership is especially important after the pilot because MRV usually fails in ordinary months, not during a well-staffed implementation workshop.
The result should be a verification-ready operating routine, not a finished compliance project. Serbian factories will continue to refine contracts, obtain better precursor data and expand coverage across product families. The key achievement is that evidence preservation, reconciliation and review become monthly disciplines. Historical records that were never captured cannot be recreated by a consultant in 2027. The most valuable action now is therefore to make the 2026 production story reproducible while the data still exist.
Elevated by CBAM.Clarion.Engineer