Extract data from ESG and sustainability reports
A sustainability report is where a company states its environmental, social, and governance performance, and it is read by people who need the numbers rather than the narrative. An ESG analyst building a rating, a procurement team screening a supplier, and an investment data vendor feeding a screen all pull the same metrics from reports built on the major frameworks, the GRI Standards, the European Union's ESRS under the Corporate Sustainability Reporting Directive, and the ISSB IFRS sustainability standards: the greenhouse gas emissions split across Scope 1, Scope 2, and Scope 3, water consumption, waste generated and the share recycled, the renewable energy percentage, the workforce and diversity figures, the health and safety rate, and the board composition and executive pay ratio. A report covering the fiscal year ending 2025-12-31 for a named organization carries dozens of these figures across a document that can run 120 pages.
Numbers hide in tables, prose, and infographics, and the units and scopes have to stay attached. Emissions are stated in metric tons of CO2 equivalent and split by scope, and mixing Scope 2 with Scope 3 overstates a footprint badly, so each figure has to keep its scope. Water is in cubic meters, waste in metric tons, and a recycled percentage is meaningless without the total it applies to. Safety figures are lost-time injury frequency rates per million hours worked, diversity figures are percentages of a defined population, and the executive compensation ratio compares the highest pay to the median. Whether the report is third-party assured, and at what level, changes how much weight an analyst gives the numbers.
Talonic reads the sustainability report and returns the reporting organization and period, the frameworks and assurance status, and the environmental, social, and governance metrics as typed fields with their units and scopes intact, plus the emissions and metric time series as structured tables. A Northwind Robotics Inc. report published 2026-04-30 for the year ending 2025-12-31, stating Scope 1 emissions of 42,000 tCO2e, Scope 2 of 18,500 tCO2e, Scope 3 of 210,000 tCO2e, a 62 percent renewable energy share, and a lost-time injury rate of 1.8, assured to a limited level by Ashford & Vance LLP, loads into a rating model with each metric tied to its unit rather than transcribed from a page. Verification is out of scope: the report is structured as stated, not assured.
What gets extracted from ESG and sustainability reports
How extraction works for ESG and sustainability reports
Sustainability reports are published as designed PDFs, and because no single layout governs them, the same metric appears in a table on one report, an infographic on another, and a sentence on a third. Talonic classifies the report and maps it to the sustainability schema in the Field Registry, which separates the reporting organization and period from the environmental, social, and governance metric blocks and the supporting time-series tables. Greenhouse gas emissions are kept split across Scope 1, Scope 2, and Scope 3 in metric tons of CO2 equivalent, water is kept in cubic meters and waste in metric tons, and each metric carries its explicit unit so a recycled percentage stays bound to the total it applies to. The reporting frameworks, GRI, ESRS, and the ISSB IFRS S1 and S2 standards, and the assurance provider and level are captured as their own fields, and the emissions and resource-consumption tables return as structured rows by year and scope. Analysts and data teams verify an emissions figure or a safety rate against the source report through the confidence score and pixel-region pointer each value carries under DIN SPEC 91491. Assurance remains the job of an independent provider; the extraction does not verify, recompute, or assure the figures.
Sample extraction
An annual sustainability report for a mid-cap manufacturer
{
"document_number": "ESG-2025-NWRB",
"reporting_organization": "Northwind Robotics Inc.",
"reporting_period_start": "2025-01-01",
"reporting_period_end": "2025-12-31",
"publication_date": "2026-04-30",
"reporting_standards": [
"GRI",
"ESRS",
"IFRS S2"
],
"assurance.third_party_verified": true,
"assurance.verification_provider": "Ashford & Vance LLP",
"assurance.assurance_level": "limited",
"environmental.greenhouse_gas_emissions_scope1": 42000,
"environmental.greenhouse_gas_emissions_scope2": 18500,
"environmental.greenhouse_gas_emissions_scope3": 210000,
"environmental.water_consumption": 340000,
"environmental.waste_recycled_percentage": 71,
"environmental.renewable_energy_percentage": 62,
"social.employee_count": 4200,
"social.diversity.women_in_management_percentage": 38,
"social.health_safety_lost_time_rate": 1.8,
"governance.board_size": 9,
"governance.board_diversity.independent_members": 7,
"governance.executive_compensation_ratio": 84,
"greenhouse_gas_emissions": [
{
"year": 2025,
"scope": "Scope 1",
"category": "Stationary combustion",
"emissions_mtco2e": 42000
},
{
"year": 2025,
"scope": "Scope 2",
"category": "Purchased electricity",
"emissions_mtco2e": 18500
}
]
}Frequently asked
Does it keep each emissions scope separate?
Yes. Scope 1, Scope 2, and Scope 3 emissions are each returned in metric tons of CO2 equivalent and kept apart, because merging Scope 2 and Scope 3 overstates a footprint, and a scope-tagged time series lets year-over-year figures compare cleanly.
How does it handle units across metrics?
Every metric carries its explicit unit, water in cubic meters, waste in metric tons, energy as a percentage, so a recycled-waste percentage stays bound to the total it applies to and a figure is never read without its unit.
Does it capture the reporting framework and assurance?
The frameworks used, GRI, ESRS, and the ISSB standards, and the assurance provider and level are captured as their own fields, because whether a report is third-party assured, and at what level, changes how an analyst weights the numbers.
Does it verify or assure the figures?
No. Talonic structures what the report states and links each figure to its source region. Verifying or assuring the numbers is the role of an independent assurance provider, not the extraction.
What fields are extracted from ESG and sustainability reports?
Talonic returns ESG and sustainability reports as schema-validated, typed fields. Common fields include Reporting Organization, Reporting Period, Reporting Standards, Assurance, and more, each normalized (dates to ISO 8601, amounts as numbers) and mapped to a stable key so the output shape stays the same across layouts.
How accurate is extraction from ESG and sustainability reports, and how is confidence reported?
Every extracted cell carries a confidence score from 0.0 to 1.0 and a provenance pointer back to the source page and region, so low-confidence values can be reviewed against the original before the data is trusted downstream. There is no single accuracy number: confidence is reported per field so you can gate on it.
Can I use ESG and sustainability reports extraction in production?
Yes. The same engine behind this guide is available as a production REST API and Node SDK with sync, async, and streaming modes, schema versioning, signed webhooks, and EU-resident processing. Start free with an API key, then scale on usage-based pricing.
What does it cost to extract data from ESG and sustainability reports?
There is a free tier for prototyping and agent evaluation with no credit card. Paid usage is credit-based at 1,000 credits per euro: page ingestion is 100 credits per page and registry-resolved queries are free. See talonic.com/pricing for current rates.
Ready to extract from your own ESG and sustainability reports?
- GDPR
- HIPAA
- ISO 27001 / 42001 aligned
- DIN SPEC 91491 co-author
- EU-resident infrastructure
Author note
Reviewed by Talonic engineering · last reviewed 2026-07-07