AI agents for Digital Product Passports and Circular Economy

Brief ID: brief:2026-08-04:dpp-circuloop-agent-wedge
Prepared: 2026-08-04
Frame: A first-pass startup wedge brief on two linked product directions: Digital Product Passports with circularity scoring, and a circular-economy / industrial-symbiosis operating layer. The brief asks whether there is a real buyer, a forcing function, and a fast prototype path.

The call

The stronger near-term wedge is an agentic Digital Product Passport compliance copilot for one regulated vertical, not a broad consumer circularity rating and not a general “Circular Economy OS.”

The reason is simple: regulation creates a deadline. The EU Digital Product Passport is not a vague climate-tech trend. The European Commission describes it as a digital container for product, component, and material information that supports sustainability, circularity, and legal compliance.¹ It is being introduced across product groups, starting with batteries and followed by groups such as textiles, iron and steel, and construction products.¹ That creates a non-optional workflow for companies that sell into Europe.

The second concept, CircuLoop OS, is bigger but harder. Industrial symbiosis means one company’s waste, heat, byproduct, or unused material becomes another company’s input.¹⁴ That is real. But a horizontal marketplace has cold-start, trust, logistics, and data-quality problems. The better route is to start with compliance data capture. Then use that structured product and material data to identify reuse, recycling, resale, repair, or byproduct opportunities.

My lean: build toward “compliance → verified product/material data → circularity use cases.” Do not start with a generic circularity score.

What the two ideas really are

Idea 1: Digital Product Passport + circularity rating

This is a product data layer. A product gets a machine-readable identity. That identity can carry facts about materials, origin, repairability, recycled content, durability, environmental footprint, safety, and end-of-life path.

The consumer-facing version is: make this feel like an energy rating. A buyer can scan a QR code and understand whether a product is circular, repairable, recyclable, or wasteful.

The startup version is different: help manufacturers gather the evidence, map it to regulations, publish the passport, update it over time, and expose role-specific data to regulators, buyers, repairers, recyclers, and customers.

Idea 2: CircuLoop OS for industrial symbiosis

This is a resource-matching and coordination layer. One company has waste or surplus. Another company can use it. The product would collect stream data, classify it, verify it, match it, and help convert a possible match into a transaction.

This is more ambitious. It could matter in chemical parks, plastics, batteries, construction waste, food/agriculture byproducts, heat/steam, water, and packaging. But the hard part is not matching in software. The hard parts are quality, contamination, regulation, liability, logistics, trust, volume consistency, and who pays.

Known / inferred / unknown

Known

• The EU has an official Digital Product Passport initiative under ESPR.¹
• The European Commission says ESPR entered into force on 18 July 2024 and extends ecodesign rules to virtually all physical products, with limited exemptions.²
• The Commission says DPPs are intended for businesses, consumers, repairers, recyclers, and public authorities.
• The Commission says the DPP Registry will store unique identifiers and mandatory registration data.⁴
• Batteries are the first strong forcing case.³ The EU Batteries Regulation entered into force on 17 August 2023. Industry-facing battery-passport vendors state that from 18 February 2027, EV and industrial batteries above 2 kWh placed on the EU market must carry a Digital Battery Passport.¹⁰
• DPP vendors already exist: Circularise, TrusTrace, Kezzler, Avery Dennison / connected-product platforms, Spherity, Minespider, Traceable Digital, Cleo Labs, Makersite, and others.
• Industrial symbiosis platforms already exist in several forms: open waste marketplaces, facilitated synergy-identification tools, social/network platforms, and regional systems.¹⁴

Inferred

• The most fundable wedge is not “circularity rating for consumers.” It is a B2B workflow where compliance is urgent and evidence collection is painful.
• AI agents are useful when the workflow spans messy documents, suppliers, changing rules, and repeated evidence checks.
• The strongest early verticals are likely batteries, textiles/apparel, and construction materials or industrial materials.
• CircuLoop OS becomes more credible after the product has access to structured product/material data, not before.

Unknown

• Which vertical has the strongest real operator access and domain insight.
• Whether there is access to manufacturers, recyclers, industrial parks, or compliance teams.
• Whether the buyer pain is regulation, procurement, resale, recycling, waste monetization, or reporting.
• Whether a 2-week prototype should focus on DPP creation, circularity scoring, waste-stream matching, or compliance gap detection.

Regulatory forcing function

The Digital Product Passport is strong because it is tied to EU market access.

The European Commission’s DPP page says:

“The Digital Product Passport (DPP) is a digital container for products, components, and materials, which will store relevant information to support products’ sustainability, promote their circularity and strengthen and facilitate legal compliance.”

Source / provenance: European Commission, Digital Product Passport page, official text read.
Why this matters: This defines the DPP as infrastructure for compliance and circularity, not a consumer marketing badge.

The same page says the DPP will help:

“businesses, consumers, repairers and recyclers, and public authorities access reliable product information in a more transparent and standardised digital way.”

Why this matters: The user roles are wider than the buyer at checkout. Repairers, recyclers, and authorities matter because they create B2B workflow demand.

The ESPR page says:

“The Ecodesign for Sustainable Products Regulation (ESPR), which entered into force on 18 July 2024, is the cornerstone of the Commission’s approach to more environmentally sustainable and circular products.”

It also says ESPR extends the old Ecodesign Directive:

“to cover virtually all physical products,” with limited exemptions such as food/feed and medicinal products.

Source / provenance: European Commission ESPR page, official text read.
Why this matters: The market is not one narrow category. But the delegated acts roll out by product group, so the startup must choose a wedge instead of trying to serve all products.

The European Commission DPP page describes the Registry:

“The DPP Registry is the EU database supporting the Digital Product Passport system. It will be established at EU level and will store the unique identifiers for each product, alongside mandatory registration data as specified in the applicable legal act.”

Why this matters: This creates a product architecture constraint. A vendor probably does not replace the EU Registry. A vendor helps the company host the passport, manage product data, produce identifiers, and keep evidence in shape.

Market map

1. DPP / traceability platforms

These companies help brands and manufacturers create and manage product passports or traceability records.

Examples from the source run:

• Circularise — supply-chain traceability, DPPs, mass balance, chain-of-custody, and data sharing.
• TrusTrace — fashion supply-chain traceability and DPP compliance.
• Kezzler — connected products, serialization, traceability, DPP implementations.
• Spherity — verifiable DPPs, credentials, battery and product data, AI-accessible DPP interfaces.
• Minespider — battery passports, supply-chain traceability, Catena-X onboarding.
• Traceable Digital — AI-native DPP compliance engine aimed at manufacturers.
• Avery Dennison / connected product players — QR/RFID/labeling and connected identity infrastructure.

Read: crowded, but not closed. The buyer problem is still early and messy. A new entrant needs a vertical wedge, not a generic “DPP platform.”

2. Product compliance AI

This category is broader than DPP. It maps products to rules across countries and product categories.

Examples:

• Cleo Labs markets product compliance for global brands across CSRD, AGEC, GPSR, DPP, REACH, and other regulatory surfaces.
• Search results and X/Grok radar surfaced Cleo-like multi-agent compliance positioning.

Read: this is a serious adjacent category. It suggests the wedge may be “product compliance agent” with DPP as the concrete first obligation.

3. PLM / LCA / sustainability data platforms

These tools live nearer product design, procurement, bill of materials, life-cycle assessment, supplier data, and product carbon footprints.

Examples:

• Makersite — AI-powered product lifecycle intelligence.
• One Click LCA — strong in construction/product environmental data.
• Stibo Systems / Informatica / enterprise data platforms — master data and compliance plumbing.

Read: enterprise incumbents can win if the buyer is already embedded in PLM/ERP. A startup must win where speed, vertical focus, or messy evidence extraction beats broad enterprise suite depth.

4. Industrial symbiosis / waste marketplace tools

These are the CircuLoop OS neighborhood.

Examples:

• Sfridoo — marketplace for waste materials, byproducts, secondary raw materials, and surplus resources.
• Academic reviews describe categories such as open online waste markets, facilitated synergy-identification systems, industry-sector identification systems, social network platforms, knowledge repositories, and regional identification systems.

Read: the software shape exists. The hard question is not “can you build a matching UI?” It is whether you can get reliable stream data, trust, quality specs, logistics, and transaction liquidity.

X/Grok radar

X/Grok was useful here.²¹ It surfaced a live market narrative:

• DPP is moving from education to implementation.
• The DPP Registry, unique identifiers, and product-group deadlines are getting attention.
• Battery passports are the most concrete early forcing case.
• AI-native DPP vendors are already using “upload reports → extract fields → map to regulation → publish passport” messaging.
• Web3/trust-layer projects such as OriginTrail and IOTA are active in the public conversation, but those claims need source-of-record verification before they matter.
• Industrial symbiosis is discussed as a data + marketplace + project-development problem, not only as a circular-economy concept.

Important limit: X/Grok is radar, not proof. I used it to find players and claims. I treated official EU pages, company pages, and accessible specialist sources as the grounding lane.

Where AI agents create real leverage

AI agents are useful if they reduce repeated cross-source coordination. They are weak if they only add a chat layer to a normal database.

Strong agent jobs

1. Supplier document ingestion
   Input: PDFs, certificates, test reports, spreadsheets, declarations, emails.
   Agent job: extract fields, map evidence to a DPP schema, flag missing data, ask suppliers for exactly what is missing.

2. Regulation-to-product mapping
   Input: product category, material, geography, SKU, supplier chain.
   Agent job: map the product to ESPR, battery, textile, REACH, PFAS, CSRD, or other obligations.

3. Evidence gap checker
   Input: draft DPP and evidence folder.
   Agent job: say which claims lack proof, which fields are stale, and which evidence is source-bound.

4. Passport update monitor
   Input: changing supplier data, lifecycle data, repair/resale/recycling events.
   Agent job: maintain a living passport instead of a static QR page.

5. Circularity opportunity scout
   Input: structured material data, product category, location, volumes, recyclability, take-back path.
   Agent job: identify possible resale, repair, recycling, remanufacturing, or industrial-symbiosis opportunities.

6. Waste-stream match packet
   Input: byproduct description, quality specs, contamination limits, volume, location, legal constraints.
   Agent job: produce a match list and a transaction-readiness packet.

Weak agent jobs

• “Ask a chatbot if this product is sustainable.”
• “Generate a circularity score from thin data.”
• “Make a marketplace with AI matching” before there is trustworthy supply and demand data.
• “Blockchain DPP” as the main pitch before the buyer feels a compliance workflow pain.

Wedge decision matrix

Top verticals to test first

1. Batteries

Why it ranks first: the deadline and buyer pain are clearest.¹⁰ Battery passports are already framed as a market-access condition. Dynamic data also matters: battery state of health, lifecycle performance, second-life use, carbon footprint, due diligence, recycled content, and value-chain responsibilities.

Best prototype: battery passport evidence copilot. Do not build a polished dashboard. Build a document ingestion and missing-field checker.

Source support: EU Batteries Regulation official page; Minespider 2026 battery passport report; Global Battery Alliance materials; X/Grok radar around battery passport implementation.

Risk: crowded market and high domain burden. A new entrant needs a narrow customer segment, such as smaller suppliers trying to sell into OEM chains.

2. Textiles/apparel

Why it ranks second: textiles are a named DPP priority area.¹¹⁹ The supply chain is fragmented. Brands already care about traceability, chemical safety, resale, repair, and green claims. The data problem is messy enough for agents.

Best prototype: DPP readiness copilot for apparel SMEs. Upload tech packs, supplier certificates, material declarations, and chemical safety docs. Produce a readiness report, missing evidence list, and supplier request packet.

Source support: European Commission DPP page names textiles; European Parliament study describes textile DPP as traceability/circularity/transparency infrastructure; TrusTrace source text says the visible QR/app is only the tip of the iceberg and the real challenge is lifecycle data.

Risk: many startups already sell fashion traceability. The wedge must be “fast compliance/evidence copilot for underserved suppliers,” not another consumer story page.

3. Construction products / industrial materials

Why it ranks third: construction products, iron and steel, and aluminium appear in the DPP/ESPR source trail.¹⁵ Material data, EPDs, LCA, recycled content, and procurement rules already matter.

Best prototype: EPD/DPP mapper. Input a product datasheet and EPD. Output a passport-field map, evidence confidence, and circularity gaps.

Source support: European Commission DPP page names construction products, iron and steel; ESPR page says requirements can cover durability, circularity, recyclability, and environmental footprint; GS1 notes DPPs apply to products, components, and intermediate products placed on the EU market or exported into it.

Risk: sales are slow and existing LCA/product-data tools are strong.

Prototype concepts ranked by speed-to-learning

Prototype 1: DPP evidence gap checker

What it does: upload product documents. The agent extracts DPP-like fields, maps them to a vertical schema, and returns “ready / missing / unsupported / stale.”

Why it teaches fastest: it tests the core pain: whether companies have messy data and will value a tool that turns documents into compliance evidence.

Source basis: EC DPP definition;¹ TrusTrace lifecycle-data challenge;⁹ AWS agentic AI framing around supplier data harmonization and automated compliance checks;¹³ Minespider data-fragmentation note.¹⁰

Build: one vertical, one schema, one upload flow, one report page.

Prototype 2: Supplier chase agent

What it does: after identifying missing DPP fields, it drafts precise supplier requests with the exact document or data field needed.

Why it teaches: DPP pain is partly cross-company coordination. If supplier chase is the time sink, this is valuable.

Source basis: DPP requires product data across lifecycle and chain;¹ TrusTrace and Circularise describe supplier and chain-of-custody data as the hidden work.⁸⁹

Build: no marketplace, no scoring. Just: “Here is what we need from Supplier A, B, C.”

Prototype 3: Battery passport mini-copilot

What it does: for a mock EV/industrial battery, map documents to passport fields: chemistry, carbon footprint, recycled content, due diligence, state-of-health placeholder, second-life notes.

Why it teaches: the buyer deadline is clearest in batteries.

Source basis: EU Batteries Regulation;³ Minespider 2026 report;¹⁰ GBA battery passport materials.²⁰

Build: use public sample docs first, then ask source for realistic anonymized examples if they can be obtained legally and anonymized.

Prototype 4: Circularity rating with evidence confidence

What it does: generate a consumer-readable circularity label, but every score must link to evidence. Unsupported claims get low confidence.

Why it teaches: tests the source’s original energy-rating analogy while avoiding greenwashing.

Source basis: EC DPP’s consumer/repairer/recycler role;¹ DPP challenge sources on verification and greenwashing.

Build: start with 10 products in one category. The output is not “score = 82.” It is “repairability: supported; recycled content: unverified; end-of-life path: weak.”

Prototype 5: Industrial symbiosis match packet

What it does: take a structured waste/byproduct stream and produce possible matches, missing specs, legal/logistics blockers, and outreach language.

Why it teaches: tests CircuLoop without building a marketplace.

Source basis: industrial symbiosis review says open e-marketplaces match supply and demand streams; the same review identifies the informational gap as a main barrier.¹⁴

Build: use one material stream and one region. A national marketplace is too broad.

Open questions

1. Which industry is closest to the available founder/domain edge: chemicals, plastics, batteries, textiles, food/ag, construction, packaging, or electronics?
2. Which workflow is visibly broken in real operations: supplier data collection, compliance, waste matching, certifications, audits, procurement, or recycling?
3. When you say “circularity rating,” who reads it: consumers, procurement teams, regulators, repairers, recyclers, or brands?
4. Who pays for the first version?
5. Does the buyer have a legal deadline or just a sustainability aspiration?
6. What data does the company already have today?
7. What data is usually missing?
8. Who owns the documents: supplier, manufacturer, brand, certifier, recycler, or logistics provider?
9. Do target companies have waste/byproduct streams they already struggle to monetize?
10. Can the team get 3 anonymized real examples: one product document packet, one compliance requirement, and one waste/byproduct stream?
11. Is the likely wedge a compliance SaaS, a marketplace, a data standard, or a consultant-assisted product?
12. What would make a customer try a prototype within two weeks?

The skeptic’s case

1. DPP tooling may become an incumbent feature

PLM, ERP, traceability, labeling, and ESG vendors can add DPP modules. If the product is only a data store with a QR code, incumbents win.

Counter: incumbents are weaker at messy, fast, vertical-specific evidence automation for SMEs. That is the wedge.

2. Circularity scores can become greenwashing

A rating is only credible if it is evidence-backed. Unsupported claims create legal, reputational, and buyer-trust risk.

Counter: make the rating a confidence-labeled evidence map, not a single magic score.

3. Industrial symbiosis marketplaces are hard

Marketplaces need liquidity. Industrial byproducts need quality specs, volume consistency, legal handling, transport, and trust. A matching model alone will not solve this.

Counter: start with “match-readiness packets” for a narrow material stream. Do not launch a horizontal marketplace.

4. AI can hallucinate compliance

Compliance mistakes have legal consequences. An agent must cite the exact source, field, and evidence file behind each recommendation.

Counter: use agents as evidence collectors and gap checkers, not autonomous certifiers.

5. The buyer might not be the visionary user

Sustainability teams may like it, but procurement, compliance, product, and legal may control budget. The first product must hit a budget owner’s pain.

Counter: require concrete buyer access before choosing the wedge.

What would change my mind

I would move the recommendation toward CircuLoop OS first if the evidence shows:

• a known industrial park or cluster with active waste/byproduct exchange pain;
• recurring material streams with volume, location, quality specs, and disposal cost;
• at least one buyer who wants those streams;
• a trusted intermediary who can validate specs and logistics;
• a reason existing marketplaces are not enough.

I would move toward consumer circularity rating first if the evidence shows:

• a product category where consumers or B2B buyers already care;
• verifiable product data availability;
• a credible distribution channel;
• a buyer willing to pay for the label or conversion lift;
• low greenwashing/regulatory risk.

Next 48 hours

1. Pick one vertical. Default to batteries if there is technical access, textiles if there is brand/supplier access, and construction materials if there is industrial/materials access.
2. Get three real examples. One product, one data packet, one painful workflow.
3. Build a fake-door output. No backend. Produce a sample DPP readiness report from public documents.
4. Test with one domain person. Ask: “Would this save time? What is wrong? Who owns this?”
5. Do not name it yet. CircuLoop is nice, but the first test should validate pain, not branding.

Bottom line

The idea is real enough to deserve a serious run. But the product should not start as a general sustainability agent or a circular-economy marketplace. The most likely wedge is a source-cited compliance and evidence agent for DPP readiness in one vertical. If that works, it can expand into scoring, lifecycle services, and eventually industrial symbiosis.

Source trail

Official / source-of-record

1. European Commission — Digital Product Passport
   Used for the DPP definition, roles, product groups, and Registry language.

2. European Commission — Ecodesign for Sustainable Products Regulation
Used for ESPR entry into force, scope across physical products, and product sustainability requirements.

3. European Commission — Batteries
Used for EU battery-regulation context and circular battery policy.

4. European Commission — DPP Registry
Used for the Registry role as a central EU-level database / identifier layer.

5. GS1 Europe — Digital Product Passport powered by GS1 standards
Used for standards/interoperability framing and global-trade implications.

Competitor / market sources

6. Traceable Digital
AI-native DPP compliance positioning for manufacturers.

7. Cleo Labs and Cleo Labs product compliance
AI-powered product compliance across regulations and countries.

8. Circularise and Circularise DPP platform
Traceability, DPP, chain-of-custody, supplier-data framing.

9. TrusTrace DPP compliance
Fashion/textile DPP compliance and lifecycle-data challenge.

10. Minespider — Digital Battery Passport Implementation Report 2026
Used for battery deadline framing, data fragmentation, and dynamic-data difficulty.

11. Makersite
Product lifecycle intelligence and sustainability data platform category.

12. Spherity — AI-enabled DPPs
Used as a live competitor example for AI-queryable / verifiable DPPs.

13. AWS — Managing sustainability data for DPPs with agentic AI
Used for agentic AI roles: supplier data harmonization, compliance checks, sustainability data fabric.

Industrial symbiosis sources

14. Industrial symbiosis platforms for synergy identification — systematic review, PMC
Used for platform types, e-marketplace language, matching supply/demand streams, and informational-gap barrier.

15. Sfridoo and Sfridoo circular economy platform
Used as a market example for waste/byproduct/secondary-material marketplaces.

16. European Circular Economy Stakeholder Platform — Sfridoo good practice
Used as corroboration that Sfridoo is positioned as a B2B industrial waste/byproduct marketplace.

Pilot / sector sources

17. CIRPASS-2
Used for real-setting DPP pilot deployments.

18. Fraunhofer IZM — CIRPASS-2
Used for pilot value chains: textiles, electronics, tyres, construction products.

19. European Parliament study — Digital product passport for the textile sector
Used for textile DPP as traceability/circularity/transparency infrastructure. Full PDF extraction was not run in this pass; search-visible official metadata was used.

20. Global Battery Alliance — Battery Passport
Used for battery passport category and industry coordination framing.

X/Grok radar

21. X/Grok query: digital product passport AI agents circular economy industrial symbiosis startup DPP ESPR battery passport product data compliance waste stream matching
    Used as liveness radar. It surfaced Traceable Digital, Cleo Labs, WasteTrade, OriginTrail, IOTA, Spherity, CIRPASS-2, and industrial-symbiosis AI/project-development discourse. Claims from X were not treated as proof without other source support.

Run notes

• Research lanes: X/Grok, native web search, official/regulatory, competitor, industrial symbiosis, skeptic/data-quality.
• Key source roles: European Commission official pages, company pages, academic review, X/Grok radar.
• Degraded sources: some PDFs were located but not fully extracted in this pass; they are listed honestly as search-visible or metadata-supported where applicable.
• Manual synthesis: yes. The source-runner was not used; the manual source-layer workflow was used directly.