Last week, EcoActive hosted the webinar “From CSRD Requirements to Assurance-Ready Reporting: How Automation and Governed AI Change the Reporting Cycle.” The session was an opportunity to look at the changes presented under the revised CSRD and ESRS framework and, more importantly, what those changes mean for the way reporting teams manage applicability, materiality, data, evidence, validation, reviews, approvals, and assurance readiness.
Through practical discussion and a live platform demonstration, we explored where traditional reporting processes tend to break down, how automation can reduce manual coordination and late-stage validation, and where governed AI can support the reporting cycle while keeping human review and traceability in place.
Here, we share five key takeaways from the webinar on what reporting teams should consider as they prepare for a more connected and assurance-ready CSRD reporting process.
First, what changed under the revised CSRD and ESRS?
Before looking at the reporting process itself, the webinar outlined the changes presented under the revised framework.
The session covered a narrower scope, a significant reduction in mandatory datapoints, the discontinuation of sector-specific ESRS development, changes affecting value-chain information requests, and the continuation of limited assurance.
The speakers also emphasized that double materiality remains in place. Companies still need to consider both financial and impact materiality, although the session discussed simplifications in how the assessment can be approached.
Among the key changes presented during the webinar:
- The revised scope was described as applying a cumulative threshold of more than 1,000 employees and more than €450 million in net turnover.
- Mandatory datapoints were presented as reducing from 1,052 to approximately 290-325, with voluntary datapoints removed.
- Development of sector-specific ESRS was described as discontinued.
- Limited assurance remains, while the previously planned move to reasonable assurance was described as no longer part of the revised approach discussed during the session.
- A value-chain cap was discussed, limiting the information large in-scope companies can request from smaller counterparties.
- ESRS 2 general disclosures remain mandatory, while topical standards apply where the organization’s materiality assessment identifies the topic as material.
- The session stated that the revised scope and requirements would apply for financial years beginning 1 January 2027, with the first reports published in 2028.
Readers can also refer directly to the EFRAG ESRS Knowledge Hub for the underlying standards and requirements.
The webinar also covered three reporting paths for FY2026: continuing with the existing framework, early adopting the simplified ESRS, or using a hybrid approach with specified reliefs. Whichever route is chosen, the session emphasized that the reporting path applied must be disclosed.
For reporting teams, however, fewer requirements do not remove every source of complexity. That became the focus of the rest of the session.
1. Fewer data points do not remove reporting-process complexity
A reduced disclosure requirement may change how much needs to be reported, but reporting still involves multiple connected activities.
Teams need to understand applicability, determine what is material, identify the relevant disclosures, collect information from across the organization and value chain, manage reviews and approvals, and ensure that the information eventually reported is complete and consistent.
This was reflected in the webinar poll. Participants identified collecting data across teams and suppliers and managing evidence and data lineage among the areas creating effort or risk. The discussion noted that a gap early in the process can reappear later, making these challenges part of the same reporting chain rather than isolated issues.
What reporting teams should consider
Preparing for revised requirements should therefore involve more than reviewing which datapoints remain applicable.
Teams also need to consider how each requirement connects to the people, data, evidence, reviews, and approvals behind it as part of the wider ESG reporting process.
2. Materiality should connect directly to reporting requirements
The webinar highlighted a common challenge in the double materiality assessment process.
Stakeholder engagement may sit in one place, IRO assessments somewhere else, scoring methodology in another document, and board sign-off separately again. At the end of the exercise, teams may have a list of material topics but still need to manually determine which datapoints should be reported against them.
The session demonstrated how the outcome of the DMA can help define the scope of reporting, while material topics remain connected to stakeholder engagement, scoring methodology, policies, actions, targets, and disclosure requirements.
What reporting teams should consider
A double materiality assessment should not end with a standalone list of material topics.
Reporting teams should be able to trace why a topic was considered material and how that conclusion connects to what is subsequently reported.
3. Assurance readiness starts with connected evidence and data lineage
The webinar also focused on the challenge of tracing information back to its source.
Sustainability data can come from ERP systems, invoices, spreadsheets, questionnaires, suppliers, and other sources. Information from the value chain may also need to be brought into the reporting process. Before reaching the final report, data may undergo calculations, transformations, standardization, or unit conversions.
When supporting evidence is scattered, teams may later need to reconstruct where a value came from, what happened to it, and who was involved.
The session demonstrated an alternative in which a value remains connected to its source, calculations, supporting evidence, ownership, and change history throughout the reporting process.
What reporting teams should consider
Evidence should remain connected to the information it supports as reporting progresses.
That means that when assurance teams ask where a value came from, reporting teams have visibility into its source, supporting evidence, changes, and ownership without having to rebuild that history at the end.
4. Validation should happen throughout the reporting cycle
Another challenge discussed during the webinar was late-stage validation.
In a manual reporting process, issues such as missing information, calculation errors, inconsistencies between tables and narratives, or differences in the same number appearing in multiple places may only be identified once the report is nearly complete.
During the live demonstration, validation was shown much earlier in the process.
Checks could identify missing data, calculation inconsistencies, linking mismatches, and gaps between what had been reported and what was required as information moved through the reporting environment.
The demonstration also showed how a change to information that had already been updated or approved could be surfaced again to the relevant users, with the status of related tasks, reviews, and approvals remaining visible.
What reporting teams should consider
Validation does not have to be a final checkpoint.
Running checks as information enters and moves through the reporting process gives teams more opportunity to identify issues during preparation.
And when information changes, the affected reporting activity can return to the appropriate workflow instead of depending entirely on manual follow-up.
5. Automation and governed AI have different roles to play
A key part of the webinar focused on how automation and AI should be used within reporting.
The session emphasized starting with standard automation, rather than applying large language models to every activity.
The approach discussed was to use traditional automation first, then techniques such as NLP and machine learning where appropriate, and use LLMs for areas where language interpretation, analysis, and narrative capabilities are useful.
The speakers also highlighted the importance of exercising greater control where numbers and reporting data are involved.
At the same time, AI use must remain governed.
Human review and approval were described as essential. AI-supported activity can remain identifiable, including visibility into which model and model version were used. The session also described governed AI as going beyond tracking who did what and when: it includes defining what AI is allowed to do within the reporting process.
This approach to Agentic AI in disclosure management places AI within the wider reporting workflow rather than treating it as a separate activity.
What reporting teams should consider
The question is not simply where AI can be added.
It is:
Which activities are better handled through deterministic automation, where can AI meaningfully assist, and how can AI-supported activity remain visible, reviewable, and traceable?
The bigger takeaway: keep the reporting cycle connected
The webinar began with changing CSRD and ESRS requirements, but the discussion ultimately came back to how the reporting process itself is managed.
Materiality needs to connect to reporting requirements. Data needs to connect to disclosures. Evidence needs to remain attached to the information it supports. Validation needs to identify issues while reporting is underway. Reviews and approvals need to remain visible as information changes.
The live demonstration also illustrated how data lineage can remain intact as a value moves through different parts of the process. Once a number enters the reporting environment, its identity can remain connected as it appears in statements, narratives, and ultimately the final report.
The Q&A extended this idea to financial and sustainability reporting. The speakers explained that the same number can flow into multiple reports rather than being collected separately for each output — described during the session as “collect once and report multiple times.”
That matters particularly when something changes.
Instead of tracing every consequence manually, reporting teams can retain visibility into the changed information, the areas it affects, the reviews that require attention, and the evidence and history behind it.
The final webinar poll reflected this emphasis: connected data, evidence and disclosures emerged as the leading area participants saw as an opportunity to improve the next reporting cycle, followed by workflow ownership and approvals.
See the reporting cycle in action
During the webinar, we demonstrated how EcoActive supports the reporting process across applicability, materiality, data management, evidence, validation, workflow, review, governed AI, and final report preparation.
These activities come together within EcoActive’s ESG Disclosure Management environment, helping teams maintain the connections between information, decisions, reviews, and final disclosures.
See how the process works in practice
