Understanding GS1 Application Identifiers in a Digital Product Passport

Published on dppkit.io/blog

When you open a Digital Product Passport for the first time, one of the first things you'll encounter is a product identifier — a structured string of numbers that tells verifying systems exactly what product this credential describes and at what level of granularity. Those identifiers are built on GS1 Application Identifiers (AIs), and understanding how they work is foundational to implementing DPPs that actually interoperate with the rest of the UNTP ecosystem.

This post breaks down the three AI codes you'll encounter most in a DPP, what each one means, and why the distinction between model, batch, and serial level matters more than it might initially seem.

What Is a GS1 Application Identifier?

GS1 Application Identifiers are standardized numeric prefixes that tell a parsing system how to interpret the data that follows. They're the same technology behind the barcodes on retail products — when a scanner reads (01)09521234543213, the 01 prefix tells it "what follows is a GTIN (Global Trade Item Number)."

In the context of UNTP's Digital Product Passport, AIs serve the same purpose: they give the product identifier a precise, machine-readable meaning that any UNTP-compatible system can interpret without custom mapping. This is exactly the kind of standardization that makes credentials portable across buyers, platforms, and registries.

Three AIs are relevant to most DPP implementations right now:

AI (01) — The GTIN: Model-Level Identification

What it is: The Global Trade Item Number. A 14-digit identifier that uniquely identifies a product model — a specific product type as defined and sold by a brand.

What "model level" means: A GTIN identifies the product design, not a specific physical item. Every unit of a particular product coming off the line shares the same GTIN. Think of it as the identifier for the recipe, not the batch or the individual cookie.

In a DPP: A model-level DPP describes the properties that are consistent across all units of that product — materials composition, design-phase carbon footprint, regulatory classifications, recyclability. For the EU Battery Regulation, this covers things like battery chemistry, energy density specifications, and declared performance metrics.

GS1 identifier format: In a DPP Kit resolver URL, a model-level DPP looks like this:

https://dppkit.io/resolver/api/1.0.0/gs1/01/09641084000022

The 01 segment is the Application Identifier — it tells any UNTP-compatible system that 09641084000022 is a GTIN identifying a product model. Scan the URL or resolve it programmatically and you get the credential for that product.

When to use it: Model-level DPPs are the right starting point for most manufacturers. A single credential covers your entire product line for that SKU. For compliance purposes, this is where regulators want to see material declarations, design data, and lifecycle information.

DPP Kit roadmap note: Model-level DPP issuance via AI (01) is the focus of our current pilot phase. When you issue a DPP in DPP Kit today, you're creating a Tier 2-conformant model-level credential with automatic validation. It's the highest-leverage place to start — one credential covers every unit of that product.

AI (10) — The Batch/Lot Number: Production-Level Identification

What it is: The Batch or Lot Number. Used in combination with a GTIN (AI 01), it identifies a specific production run of a product.

What "batch level" means: A batch-level identifier narrows the credential from "all units of this product ever made" to "all units from this specific production run." Two products with the same GTIN can have different batch numbers — and therefore different batch-level credentials — if their manufacturing conditions, material sources, or test results differ.

In a DPP: Batch-level credentials capture production-specific data that varies run to run — actual measured carbon footprint (vs. declared), specific material sourcing for that run, production date, quality assurance results, and traceability events tied to that batch. For the EU Battery Regulation, this is where actual battery cell performance test data and carbon footprint calculations live.

GS1 identifier format: The batch number extends the model-level URL — the 10 segment is the AI, followed by the batch value:

https://dppkit.io/resolver/api/1.0.0/gs1/01/09641084000022/10/BATCH2024001

When to use it: When you need to assert something specific to a production run — particularly relevant for batteries (different cathode material sourcing between batches), food (harvest-season traceability), and any product where actual measured values differ from design values.

DPP Kit roadmap note: Batch-level DPPs can be created manually in DPP Kit during the pilot phase. Automated batch-level issuance — triggered directly from production events or ERP exports — is on the roadmap for the UNTP 1.0 release in July.

AI (21) — The Serial Number: Item-Level Identification

What it is: The Serial Number. Used in combination with a GTIN (AI 01), it identifies a single, specific physical item.

What "serial level" means: This is the most granular level of identification — one credential, one physical product. Every unit gets its own unique identifier and its own credential chain.

In a DPP: Serial-level credentials unlock use cases that model and batch credentials can't — end-of-life tracking, ownership transfer, repair history, and resale authentication. They're the foundation for circular economy business models where the value of a product extends beyond the original sale. A battery's state of health at resale, for example, can only be meaningfully communicated via a serial-level credential.

GS1 identifier format: Serial follows the same pattern — 21 is the AI, followed by the unique serial number for that physical unit:

https://dppkit.io/resolver/api/1.0.0/gs1/01/09641084000022/21/SN0000142

When to use it: When individual unit tracking matters — high-value products, regulated equipment, products with resale markets, or any product where the usage history of a specific unit affects its remaining value or regulatory status.

DPP Kit roadmap note: Serial-level DPPs can also be created manually in DPP Kit during pilot. Like batch-level automation, serial-level automated issuance is planned for the UNTP 1.0 release in July. This is particularly relevant for battery manufacturers anticipating end-of-life and second-life credential requirements.

How the Levels Relate to Each Other

These aren't competing approaches — they're a hierarchy, and production DPP implementations often use all three.

Model Level (AI 01)
└── Batch Level (AI 01 + AI 10)
    └── Serial Level (AI 01 + AI 21)

A manufacturer might issue a model-level DPP for material declarations and design data, a batch-level credential for actual measured carbon footprint per production run, and serial-level credentials for high-value units entering markets where resale or repair tracking matters. Verifying systems can traverse this hierarchy — a regulator checking EU Battery Regulation compliance might verify the model-level DPP for chemistry data and the batch-level credential for actual carbon figures in the same verification pass.

Why Getting the Identifier Right Matters

The AI code isn't just metadata — it determines what claims in your credential are legally defensible and what data downstream systems can actually use. Issuing serial-level data in a model-level credential, or vice versa, creates interoperability problems that are hard to untangle later.

It also affects your credential management overhead. Model-level credentials are low volume and high value — one credential covers a product family. Serial-level credentials are high volume and require automated issuance pipelines to be viable at scale. That's exactly why DPP Kit's roadmap phases it this way: get model-level credentials right during pilot, then automate batch and serial issuance before the July 1.0 release when production volumes start to matter.

Getting Started

If you're in the early stages of your UNTP implementation, start with AI (01) model-level DPPs. They give you the highest coverage per credential, cover the majority of current regulatory requirements, and give you the foundation to layer batch and serial data on top as your use cases mature.

If you're already in DPP Kit, you can issue model-level credentials today through the credential editor or AI agent pipeline. Need batch or serial-level credentials before automated issuance ships? You can create them manually — reach out and we can walk you through it.

The standard is moving fast. UNTP 1.0 in July is a significant milestone and brings stability to what has been a moving target during the pilot phase. Now is exactly the right time to be building familiarity with how identifiers work so your implementation is ready when the standard lands.

Questions about product identification or GS1 integration? Get started with DPP Kit — model-level credentials are live and ready to issue.