In its recent white paper, The State of Global Sustainability Disclosures, Sprih Inc. analysed more than 200,000 reports from over 80,000 companies worldwide, creating one of the largest repositories of corporate sustainability data ever assembled. The findings show that sustainability reporting is no longer a fringe exercise.Yet comparability and consistency remain mainly out of reach for many businesses.
According to Sprih, this is where artificial intelligence must move from being a reporting tool to becoming the backbone of ESG intelligence.
Increasing visibility
The white paper, powered by SustainSense, Sprih’s climate AI engine, reveals a paradox. Disclosure rates for Scope 1 and Scope 2 emissions are relatively mature across many regions and sectors and near-term targets are widely adopted. Energy consumption is commonly reported in aggregate.
Yet when we move beyond headline figures, fragmentation becomes obvious.
Scope 3 emissions, which are often the largest share of a company’s footprint, remain inconsistently disclosed. Water reuse and rainwater harvesting data are scarce and waste categorisation varies widely. Smaller firms, particularly those under US$100mn in revenue, lag significantly in both completeness and consistency.
The paper explains that without standardisation, sustainability disclosures risk becoming a patchwork of narratives rather than a coherent dataset. This makes investors struggle to benchmark risk, while regulators face uneven compliance landscapes. Moreover, procurement leaders lack visibility across supply chains and executives are left navigating strategy with incomplete maps.
But AI can help change this equation.
Teaching machines the language of sustainability
One of the most powerful insights from the white paper is methodological. SustainSense does not merely collect documents; it extracts, classifies, validates and normalises data across languages, formats and reporting frameworks. In other words, it teaches machines to understand sustainability.
This matters because ESG data is not structured by default. It sits inside PDFs, integrated annual reports, regulatory filings and standalone sustainability documents. Terminology can differ across jurisdictions and definitions evolve. Units can vary and even the placement of data within reports is inconsistent.
Agentic AI architectures, as described in the paper, create a structured layer on top of this chaos. They identify emissions figures, distinguish between location-based and market-based Scope 2 data, harmonise water metrics and align targets to recognised definitions such as near-term, long-term and net zero.
The result is not just a larger dataset, but a comparable one.
When thousands of disclosures are translated into a common analytical framework, patterns emerge. Europe’s leadership in comprehensive target-setting becomes quantifiable. Asia’s relative lag in Scope 3 transparency becomes measurable. The maturity gradient between large enterprises and SMEs becomes visible at scale.
According to Sprih, this is not anecdotal ESG, but rather "it is systemic ESG intelligence."
A strategic asset
For many companies, sustainability reporting continues to feel like a compliance obligation. But the white paper offers some hope.
Executives can use AI-driven benchmarking to understand where their disclosure quality signals strength – or exposes weakness. Investors can assess governance resilience by examining not just target announcements, but the consistency of underlying metrics. Regulators can identify sectors where harmonisation efforts must intensify.
Crucially, AI can also surface blind spots. The analysis shows that while total energy consumption is widely reported, the breakdown between renewable and non-renewable energy is less consistent. Water withdrawal is commonly disclosed, but treatment and reuse metrics are rare. Waste generation is more visible than circularity performance.
These gaps, it seems, are not simply technical. They represent risk. In a climate-constrained world, incomplete value-chain data or poor resource visibility translates into financial exposure. AI could help transform ESG into static into dynamic risk management.
Better AI systems
Perhaps the most compelling idea in the white paper is the call for a global climate intelligence layer. If corporate disclosures are the raw material, AI is the infrastructure that makes them usable.
Imagine a landscape where investors can benchmark Scope 3 intensity across sectors in seconds; where procurement teams can map supplier emissions maturity; where policymakers can evaluate regional adoption of net-zero commitments with precision rather than estimates. Sprih says that this is not speculative, as it is already emerging.
However, the technology community must recognise that scale alone is insufficient. AI systems must be transparent, auditable and continuously learning. They must adapt as reporting frameworks evolve and new regulatory requirements emerge. They must balance automation with validation to ensure trust.
Equally, companies must view AI not as a shortcut to green credentials, but as a tool for accountability. The question for the market is no longer whether AI will shape ESG. It is whether organisations are ready to operate in a world where sustainability performance is no longer hidden in footnotes, but illuminated by intelligence at scale.
How can AI enhance ESG reporting?
