Case study Sustainability Engineering

ESG Reporting Automation: From Quarterly Scramble to Auditable Continuous Disclosure

Replaced a quarterly manual ESG data collection process with an automated pipeline ingesting energy, fleet, and supplier data. Reduced reporting preparation time by 84% and produced the company's first assurance-ready GRI and CSRD-aligned disclosure.

Explore the project
The project at a glance
Client
Mid-Market Logistics and Facilities Group
Industry
Sustainability / ESG
Timeline
14 weeks
Project date
Q4 2025
Team
5 Specialists (2 Data Engineers, 1 Full-Stack, 1 Analytics, 1 ESG Consultant)
84%Reduction in reporting prep time
5 daysQuarterly reporting preparation
68%Scope 3 supplier data coverage
0.4%Data extraction error rate
ZeroRestatements due to data errors
CleanBig 4 assurance opinion

Reported project results · Specific to the original project scope. See results and context below.

The brief

Project overview

Data Pipeline Engineering and ESG Reporting Automation

This account is restored from Actinode’s original project portfolio. Results describe that project and are not a guarantee for another engagement. “Ready” or compliance-related terminology describes the stated project scope, not an independent certification.

Business context

A mid-market logistics and facilities management group with 1,200 employees and operations across three EU countries faced rapidly escalating ESG reporting obligations under CSRD. Their existing process was entirely manual and unsustainable at the required frequency and quality:

  • Manual aggregation: Sustainability team spent 6-8 weeks per year collecting energy bills, fleet records, and waste data from 12 operational sites into a master spreadsheet
  • No data lineage: Final carbon figures could not be traced back to source; auditors and investors asked questions the team could not answer with evidence
  • Calculation errors: Emission factor version management was informal; different sites had used different factor tables in the same reporting year
  • Scope 3 blind spots: Supply chain emissions were estimated rather than calculated, a growing exposure as CSRD Scope 3 obligations take effect
  • Delayed decisions: Sustainability data was only available annually; the operations team could not act on energy or fleet efficiency insights during the year

Requirements and constraints

The group required a reporting infrastructure that would:

  • Automate ingestion of energy, fleet, and waste data from 12 sites with minimal manual effort
  • Produce traceable, auditable carbon calculations linked to source documents and emission factor versions
  • Generate GRI 305 (emissions) and ESRS E1 (CSRD) aligned outputs for external disclosure
  • Enable quarterly interim reporting rather than annual-only
  • Provide operational dashboards for site managers to track energy and fleet efficiency in near-real-time
  • Survive a third-party limited assurance review within 12 months of go-live
  • Integrate supplier data collection for Scope 3 Category 1 (purchased goods) and Category 6 (business travel)

Original challenge

The core challenge was not technical; it was data quality and process design. The existing data was scattered across 12 sites in incompatible formats:

Data Collection Reality at Project Start:

  • Energy: 8 sites provided monthly PDF utility bills; 4 sites had smart meters with API access
  • Fleet: 3 separate telematics providers with different API schemas and fuel type coverage
  • Waste: Paper manifests from 4 waste contractors, only partially digitised
  • Business travel: Expense reports in two different expense management systems with no common schema
  • Suppliers: No structured emissions data; suppliers provided unformatted email responses or nothing

Calculation Complexity:

  • Scope 2 required market-based calculations using supplier-specific emission factors alongside location-based DEFRA factors: two parallel calculation paths
  • Fleet data included electric vehicles (requiring grid electricity emission factors), diesel, and petrol: different calculation paths per fuel type
  • Multi-country operations meant different grid emission factors for UK, Netherlands, and Poland operations

Assurance Requirement:

The client's investors had indicated a preference for limited assurance (rather than management assertion) on the FY2025 disclosure. This required audit-grade evidence: not just correct numbers, but a traceable, reproducible trail from source to disclosure.

Solution and contribution

We designed and built a four-layer ESG data pipeline over 14 weeks:

Layer 1: Ingestion (Weeks 1-4)

  • Connected to 4 smart meter APIs (Siemens Desigo, grid supplier API) for automated monthly energy reads
  • Built a PDF extraction pipeline (AWS Textract + validation layer) to parse structured energy bill data from the remaining 8 sites; flagged extraction failures for manual review
  • Integrated 3 fleet telematics APIs (normalised to a common vehicle-journey schema) with fuel type and distance as canonical fields
  • Deployed a supplier data portal: a structured web form where suppliers submit activity data (spend, weight, distance) tagged by Scope 3 category; all submissions ingested automatically
  • All ingested data stored in append-only S3 with Object Lock and source metadata (provider, ingestion timestamp, file hash)

Layer 2: Transformation (Weeks 3-8)

  • All calculation logic implemented as versioned dbt models committed to a private Git repository
  • Emission factor tables stored as version-controlled seed tables, tagged by source (DEFRA, IEA, IPCC) and effective year
  • Every transformation step explicitly references the emission factor version it uses, enabling reproducibility across reporting years
  • Scope 1 (fleet and site combustion), Scope 2 (market-based and location-based, parallel paths), and Scope 3 Category 1 and 6 calculation models
  • Data quality checks at each layer: completeness (expected sites and months), plausibility (outlier detection by site and fuel type), and schema validation

Layer 3: Reporting Period Management (Weeks 7-10)

  • Period-close process creates an immutable, named snapshot of all calculation outputs at the close of each quarter
  • Restatement workflow: any correction to historical data creates a new named version with documented reason; the original snapshot is never overwritten
  • Disclosure mapping table linking each canonical metric to its GRI 305, ESRS E1, and internal KPI definitions

Layer 4: Outputs (Weeks 9-14)

  • Executive sustainability dashboard (Power BI) showing YTD performance vs targets by scope, site, and category
  • Site manager dashboards showing monthly energy intensity (kWh per m²), fleet km per litre, and waste diversion rate
  • Automated quarterly disclosure package: GRI and CSRD indicator table, methodology note, and data quality statement, generated from the period-close snapshot
  • Audit evidence export: for any disclosed metric, a single export produces the source records, transformation version, emission factor version, and calculation trace

Implementation

How we delivered it

Smart Meter and Utility API Integration

Connected to four smart meter management platforms via REST APIs, pulling daily meter reads and aggregating to monthly totals per site and metering point. For utility PDF bills, deployed AWS Textract with a custom post-processing layer that validates extracted kWh values against expected ranges before accepting them into the clean layer. Error rate on PDF extraction dropped from manual error rate of 3.2% to 0.4% after two months of model tuning.

Fleet Telematics Normalisation

Wrote adapter layers for three telematics provider APIs (different schemas, authentication methods, and data update frequencies). Normalised all journey records to a canonical schema: vehicle ID, fuel type, distance (km), fuel consumed (litres or kWh for EVs), date. This normalisation layer made it possible to run a single Scope 1 fleet calculation model regardless of which telematics provider reported the data.

Supplier Data Portal

Built a lightweight supplier intake portal where supply chain partners submit activity data by Scope 3 category. Submissions are validated against plausibility rules (e.g. spend-to-weight ratios for Category 1) before ingestion. Non-conforming submissions trigger an automated request for clarification. Supplier coverage reached 68% of spend-weighted Category 1 emissions within one quarter of launch.

Versioned dbt Calculation Layer

All 47 emission calculation models implemented as dbt transformations, with each model referencing the emission factor seed table by explicit version key. Running the same source data with different factor year versions produces labelled, distinguishable outputs, making restating for factor updates a controlled, documented process rather than a manual recalculation.

Assurance-Ready Period Close Process

Designed a period-close workflow: at the end of each quarter, a designated sustainability manager runs the close process, which snapshots all calculation outputs into an immutable Parquet dataset in S3 with Object Lock. The snapshot carries a manifest of all source file hashes, dbt model versions, and emission factor versions used. This manifest is the primary exhibit for assurance reviewers.

Limited Assurance Preparation

Engaged with the client's chosen assurance provider (a Big 4 firm) six weeks before the first assurance review. Provided the assurance team with access to the audit evidence export tool, which produced a traceable record for any sampled metric on demand. The review completed in 3 weeks with 4 minor findings, all resolved within one week. The client received a clean limited assurance opinion on their FY2025 GRI 305 and ESRS E1 disclosures.

Technical capabilities

Automated Multi-Source Data Ingestion (APIs + PDF Extraction)

Versioned dbt Calculation Models with Explicit Emission Factor References

Append-Only Raw Storage with Audit-Grade Retention

Immutable Period-Close Snapshots for Assurance Evidence

GRI 305 and CSRD ESRS E1 Aligned Disclosure Outputs

Supplier Data Portal for Scope 3 Category 1 Collection

Site and Executive Sustainability Dashboards

On-Demand Audit Evidence Export per Metric

The outcome

Results & business impact

  • Reporting preparation time reduced from 6-8 weeks to 5 days, an 84% reduction

  • First assurance-ready GRI and CSRD-aligned disclosure produced for FY2025

  • Clean limited assurance opinion from Big 4 reviewer on the first submission

  • Quarterly reporting cadence established from Q1 2026 onwards

  • 68% Scope 3 Category 1 supplier data coverage achieved within one quarter of portal launch

  • PDF extraction error rate reduced from 3.2% (manual) to 0.4% (automated with validation layer)

  • Zero restatements required in FY2025 disclosure due to data errors

Interpret figures with the original scope and baseline above. Where the account does not specify a measurement window, sample or denominator, do not infer one; discuss the relevant evidence and comparable workload with us before using a result to estimate your project.

What this means for the business

Regulatory Readiness: CSRD compliance achieved ahead of the mandatory reporting deadline for their company size tier

Investor Confidence: First assured disclosure strengthened credibility with ESG-focused institutional investors

Operational Visibility: Site managers now receive monthly efficiency dashboards, enabling in-year decisions on energy and fleet that were previously impossible

Reduced Staff Burden: Sustainability team's manual data work reduced from 8 weeks per year to under 1 week, freeing capacity for strategy and stakeholder engagement

Supply Chain Insight: Structured supplier data collection revealed that 3 suppliers accounted for 51% of Category 1 emissions, a fact the client had not previously known

Audit Cost Reduction: Traceable, on-demand evidence reduced external assurance time, directly lowering assurance fees in subsequent years

Conclusion

ESG reporting is becoming indistinguishable from financial reporting in its rigour and accountability requirements. This engagement demonstrated that the infrastructure gap between a credible disclosure and a spreadsheet-based estimate can be closed in a single project, and that the operational benefits (real-time site data, supplier transparency, faster reporting cycles) deliver business value beyond just regulatory compliance.

Next steps in the original project account

Next phase includes Scope 3 Category 11 (use of sold products) calculation models for the client's facilities management division, integration with building management systems for real-time energy intensity tracking, and a customer sustainability report portal for B2B clients requiring supply chain emission disclosures.

Technology & expertise

  • AWS (S3, Textract, Lambda)
  • dbt
  • PostgreSQL
  • Power BI
  • Python
  • Terraform
  • REST APIs (Telematics, Smart Meter)
  • Object Lock / Immutable Storage
Explore the relevant industry

Further reading

Related engineering guides

Explore more case studies

Discuss a focused software engagement

Tell us the problem, users and constraints. We can discuss whether discovery, a first release or a system improvement is the right next step.

Discuss your project scope