How does the EU Taxonomy connect to double materiality?

The EU Taxonomy (Regulation (EU) 2020/852) and CSRD double materiality are related but separate frameworks with different purposes. Understanding the connection — and the distinctions — is essential for compliance teams to avoid duplicating work and to ensure disclosures are consistent across both frameworks.

The Taxonomy defines whether specific economic activities are environmentally sustainable, based on six objectives: climate change mitigation, climate change adaptation, sustainable use and protection of water and marine resources, transition to a circular economy, pollution prevention and control, and protection and restoration of biodiversity and ecosystems. An activity qualifies as Taxonomy-aligned if it makes a substantial contribution to at least one objective, does no significant harm to any other (the DNSH principle), and is carried out in compliance with minimum social safeguards.

The connection to double materiality works in two directions. First, if environmental topics related to the Taxonomy's six objectives are identified as material in the DMA (through either the impact or financial lens), the corresponding topical ESRS standards — ESRS E1 through E5 — become applicable, and those standards contain specific disclosure requirements that align with and reinforce Taxonomy KPI disclosures under Art. 8 of the Taxonomy Regulation. Second, for companies with significant Taxonomy-eligible activities, the financial materiality assessment of physical and transition climate risks under ESRS E1 will typically overlap substantially with Taxonomy alignment analysis.

Practically: if your DMA concludes that climate change (E1) is material — as it will for most companies in energy-intensive sectors, manufacturing, real estate, agriculture, and transport — then you will need to report under ESRS E1, which requires disclosure of scope 1, 2, and 3 greenhouse gas emissions, transition plans, climate-related risks and opportunities, and Taxonomy-eligible and aligned revenue, capex, and opex percentages. Failing to complete the DMA first means you cannot determine whether these disclosure requirements apply.

Which ESRS standards become applicable once topics are deemed material?

The DMA outcome is the switch that activates specific topical ESRS standards. Under ESRS 1 paragraph 29, a company only applies a topical standard when the sustainability matter it covers is material to that company. This means the DMA is not merely a procedural step — it is the document that legally justifies which disclosures appear in the sustainability statement and which are omitted.

The twelve topical ESRS standards adopted by the Commission in Delegated Regulation (EU) 2023/2772 cover five environmental, four social, and one governance topic, plus two cross-cutting standards (ESRS 1 and ESRS 2 are mandatory for all in-scope companies regardless of DMA outcome):

• ESRS E1 — Climate change: Covers transition and physical climate risks, GHG emissions (scope 1, 2, 3), energy consumption, and EU Taxonomy alignment for climate objectives. Practically material for almost all large companies.
• ESRS E2 — Pollution: Air, water, soil, substances of concern. Highly material for chemicals, manufacturing, agriculture, and waste management sectors.
• ESRS E3 — Water and marine resources: Water withdrawal, consumption, and discharge. Material where operations are water-intensive or located in water-stressed areas.
• ESRS E4 — Biodiversity and ecosystems: Land use change, species impacts, ecosystem services. Currently the most technically challenging standard to implement due to limited standardised data.
• ESRS E5 — Resource use and circular economy: Material inflows and outflows, waste, product end-of-life. Particularly relevant for consumer goods, packaging, electronics, and construction.
• ESRS S1 — Own workforce: Working conditions, equal treatment, other work-related rights. The default assumption for any company with significant headcount is that this will be material.
• ESRS S2 — Workers in the value chain: Working conditions and human rights of workers upstream (suppliers) and downstream (distributors). Critical for companies with complex supply chains.
• ESRS S3 — Affected communities: Local communities' social and economic rights, indigenous peoples' rights, security-related impacts. Material for companies with significant local footprint.
• ESRS S4 — Consumers and end-users: Product and service safety, data privacy, responsible marketing. Material for B2C companies, financial services, and healthcare.
• ESRS G1 — Business conduct: Anti-corruption, political engagement, supplier relationships, tax practices. Mandatory minimum disclosures apply to all in-scope companies under ESRS 2 GOV-1, making G1 practically near-universal.

ESRS 1 (General Requirements) and ESRS 2 (General Disclosures) are not subject to materiality filtering — every in-scope company must apply them in full. ESRS 2 requires disclosure of the DMA process (IRO-1), the results of the DMA (SBM-3), and the governance, strategy, and management of material sustainability matters (GOV-1 through GOV-5).

How can AI tools reduce the time and cost of double materiality data collection?

A first-time DMA for a wave 2 company typically involves navigating four to six separate EFRAG documents, the Commission Delegated Regulation (EU) 2023/2772 itself (running to over 380 pages in the Official Journal), sector-specific guidance, and internal documents ranging from supplier contracts to risk registers and board minutes. The challenge is not that this information does not exist — it is that locating the precise paragraph, datapoint number, or phase-in provision that applies to a specific situation takes hours of expert reading time.

AI document Q&A tools change the economics of this work. Rather than a sustainability consultant billing hours to search ESRS 1 for the definition of "irremediability" or to confirm whether the ESRS S1-7 phase-in applies to wave 2 companies, a compliance officer can query an AI tool loaded with the full ESRS corpus and get a cited, accurate answer in seconds — with the relevant paragraph referenced so the result can be independently verified.

The concrete use cases for AI in the DMA process include:

• Rapid ESRS mapping: Given a topic identified in the DMA long list (e.g. "deforestation in palm oil supply chain"), an AI tool can immediately identify which ESRS standard applies (E4 and S2), which specific disclosure requirements are triggered, and what the mandatory vs. voluntary distinction is within that standard.
• Phase-in tracking: CSRD contains multiple phase-in provisions for specific datapoints and for smaller companies. An AI tool can maintain a parsed, queryable version of all phase-in rules and answer company-specific questions ("Does the phase-in for ESRS E4 biodiversity apply to us as a wave 2 non-financial company?") without manual page-scanning.
• Stakeholder engagement documentation: AI can assist in analysing large volumes of qualitative stakeholder feedback and mapping responses to specific ESRS sub-topics, speeding up the process of connecting engagement outputs to materiality scoring.
• DMA record generation: AI tools can assist in drafting structured DMA records — including the topic-by-topic materiality rationale, omission justifications, and ESRS reference mapping — from working-paper inputs, dramatically reducing the time from assessment completion to auditable documentation.
• Consistency checking across disclosures: Regulatory technical standards and EFRAG guidance require consistency between DMA results, SBM-3 disclosures, and topical ESRS disclosures. AI can flag inconsistencies — for example, where a topic is described as not material in the DMA but a related risk appears in the financial risk section without explanation.

What are the most common double materiality assessment mistakes — and how do you avoid them?

The first wave of CSRD filings in 2025 produced a body of practical experience — shared through EFRAG's implementation workshops, national competent authority guidance, and professional services firm observations — that identifies consistent failure points for companies conducting their first DMA.

Mistake 1: Treating the DMA as a tick-box rather than an evidence-based assessment. Auditors conducting limited assurance under CSRD Art. 26 will request the underlying DMA working papers — the scoring matrices, stakeholder engagement records, and rationale for each topic decision. A DMA that cannot demonstrate how each topic was scored against ESRS 1's criteria will not withstand assurance review. EFRAG's Implementation Guidance explicitly warns against "top-down" DMAs that reverse-engineer results from a pre-decided materiality list.

Mistake 2: Confining the impact assessment to own operations only. ESRS 1 paragraph 22 is unambiguous: impacts must be assessed across the upstream and downstream value chain. For manufacturers, this typically means tier 1 and 2 supplier conditions; for retailers, it means own-brand supply chains; for financial institutions, it means financed emissions and investee companies. Limiting the scope to the legal entity boundary will produce an incomplete DMA that may need to be redone.

Mistake 3: Conflating financial materiality with the traditional accounting concept. Financial materiality under CSRD is broader than IFRS or GAAP materiality — it includes risks and opportunities that could reasonably be expected to affect financial performance, even if they have not yet done so. This forward-looking element requires scenario thinking, not just a review of current financial statements.

Mistake 4: Not documenting omission rationale. Every topic not considered material must be documented with a reason. Auditors will check whether the omission is justified, especially for topics that are material for comparable companies in the same sector. A DMA that omits climate (E1) for a manufacturing company without a detailed, evidence-based rationale will attract significant scrutiny.

Mistake 5: Disconnecting the DMA from data collection planning. The DMA output is only useful if it drives immediate action on data collection. Each material disclosure requirement maps to specific data points — some of which require new data collection processes, supplier surveys, or systems integrations. Companies that complete the DMA without immediately mapping disclosure requirements to data owners miss the practical value of the exercise and risk being unable to populate their sustainability statement with sufficient data.