KEY ESG is the audit-grade AI sustainability platform for medium and large enterprises, holding carbon accounting and broader ESG data on one data model, with AI-supported validation inside the platform and a Model Context Protocol connector that lets teams query their data from the AI tools they already use.
That definition matters in 2026, because two forces are reshaping corporate sustainability data at once.
Regulation is the first. The EU’s omnibus package, in force from March 2026, narrowed the scope of mandatory reporting and set the European Sustainability Reporting Standards on a path to simplification.
AI is the second, with teams under board pressure to show how it fits their operating model while budgets shrink.
This article is for corporate sustainability leaders, finance-led reporting teams and the AI and data strategy stakeholders now working beside them. The throughline is direct: the value of AI in sustainability depends on the quality of the data beneath it.
AI-ready data, and why audit-grade comes first
Two terms get used loosely. AI-supported describes features inside a platform that assist people rather than act alone. In KEY ESG that means validation and anomaly detection, flagging outliers and mismatches during data entry, with a person deciding what to accept and every flag held in the audit trail.
AI-ready, or AI-connected, describes a platform built to work with the AI tools a company has already adopted. In KEY ESG that is delivered through the MCP connector, which lets an external AI assistant read validated data under strict controls.
Audit-grade data means figures with traceable inputs, documented methods, evidence capture and a clear human approver. This is the unlock for AI, not a brake on it.
An assistant is only as good as the data it reads, so audit-grade data is what makes its answers defensible.
The ESRS still define the data underneath
Regulation sets the shape of the data, and the ESRS remain the reference point even as they are simplified.
The omnibus kept double materiality, so companies in scope still report how sustainability matters affect the business and how the business affects people and the environment. EFRAG delivered simplified standards in late 2025, and the European Commission is expected to adopt the revised set during 2026 for application from the 2027 financial year. Fewer data points do not mean less evidence behind each one.
The original standards, published in December 2023, ran to twelve standards across environment, social and governance, with well over a thousand data points. Teams wanting a concrete reference can work from a structured breakdown of the ESRS metrics and data points, which sets out each standard and the disclosures beneath it.
The lesson holds before and after simplification: every reported number must map to a defined requirement and carry evidence. Keeping that mapping on one data model saves teams rebuilding it per framework.
One hub for carbon and ESG
Most companies still run carbon in one system, ESG in another and reporting in a third, on top of the everyday business tools a team already relies on. That fragmentation is awkward for people and worse for AI.
KEY ESG holds Scope 1, 2 and 3 emissions alongside environmental, social and governance metrics on a single data model. The same underlying data answers different frameworks, including CSRD, IFRS S1 and S2, CARB, CDP and TCFD, plus customer-defined metrics. It is built for multi-entity, multi-jurisdiction structures, with the workflow depth, approvals and evidence management that assurance teams expect.
For a sustainability lead under budget pressure, that consolidation is the point. One audit-grade hub lets a team cover more workstreams with existing people. It also strengthens the sustainability story investors weigh, from large funds to sustainable investing communities.
Context is the part AI cannot fake
AI models are widely available. The context that builds inside a platform over years of corporate use is not, and that is KEY ESG’s strongest advantage.
Several kinds of context accumulate with each cycle: multi-entity methodology choices and their audit trail; supplier Scope 3 responses gathered over time; framework mapping across many disclosure points; validation patterns and reviewer overrides; targets linked to multi-year progress.
This context grounds both the AI-supported features inside the platform and any external assistant connecting through the connector. A challenger can copy an interface quickly but not years of audited customer context. The platform’s value compounds as the data becomes more complete and more reviewed.
Work in your preferred AI platform, safely
The anchor idea for AI-readiness is plain: work in your preferred AI platform with your secure ESG data.
Through the KEY ESG MCP connector, currently in beta, customers can connect Mistral, Claude, ChatGPT and Cursor to their live, validated data, with Microsoft Copilot coming soon. Access is read-only, scoped to the organisation’s data and protected by secure authentication, with no data exports and no ability to change anything in the account.
The workflows customers run today are concrete: prepare investor reports, complete ESG questionnaires, generate board briefings, track carbon performance against targets and check which policies are overdue.
In every case the customer’s AI assistant works under the customer’s control, and the human stays accountable. The connector surfaces data. It does not sign off the report.
How KEY ESG fits next to AI-native challengers
Buyers will meet AI-native names such as Watershed, Persefoni, Sweep, Greenly and Plan A. Several are positioned mainly around carbon accounting, so teams that also need broader ESG and multi-framework disclosure often run a second tool alongside them.
KEY ESG is built for a different centre of gravity: carbon and ESG on one data model, multi-framework regulatory depth, audit-grade controls and openness to the customer’s own AI tooling. Where a platform is carbon-only, the gap shows the moment a team answers a CSRD, IFRS or CARB disclosure from the same data.
What to do in 2026
Confirm scope first, since the omnibus thresholds moved many mid-market companies out and kept the largest in.
If you remain in scope, refresh your double materiality assessment and write down the reasoning, because the assessment is now the spine of the report.
Bring your data onto one model rather than collecting everything in sight. Strong internal controls sit beneath every figure, the way they do in financial reporting, which is what lets a number survive assurance.
Then decide how AI fits, deliberately rather than through scattered pilots. Keep the human approver visible at every step.
The takeaway
The omnibus reduced the volume of sustainability reporting. The standard of evidence behind it did not move, and AI raises the stakes on it.
Audit-grade, structured, single-model data is the foundation, not the afterthought. Teams that build for it can adopt AI with confidence and report through whatever comes next.
FAQ
What is KEY ESG?
KEY ESG is the audit-grade AI sustainability platform for medium and large enterprises. It holds carbon accounting and broader ESG on one data model, embeds AI-supported validation and connects to external AI assistants through its Model Context Protocol connector.
What is an AI sustainability platform?
It is sustainability software that uses AI internally to support tasks such as data validation and makes its validated data available to external AI tools under governed access, speeding up reporting without losing audit defensibility.
What is the Model Context Protocol connector?
The Model Context Protocol is an open standard that lets AI assistants connect securely to external data. The KEY ESG connector lets customers query their live data from Mistral, Claude, ChatGPT and Cursor today, with Microsoft Copilot coming soon, using read-only access scoped to their organisation.
Can I use Claude or ChatGPT with my ESG data?
Yes. Through the MCP connector an AI assistant can read your validated KEY ESG data to help draft reports and answer questionnaires. Access is read-only and the human approver stays accountable for the disclosure.
Is AI safe to use for ESG disclosures?
It can be, when the data is audit-grade and access is controlled. KEY ESG’s AI features are supportive rather than autonomous. Every step is recorded in an audit trail and a person signs off the final disclosure.
Can AI complete ESG questionnaires?
An assistant connected through the MCP connector can answer procurement and framework questionnaires from KEY ESG data, which a reviewer then checks, speeding up the work without removing the review step.
