Swedish climate-tech company ClearTraced is using AI to automate the collection and verification of ESG data, with Pakistan-based CTO Abdullah Rashid leading the engineering behind the platform.

Companies today publish an enormous amount of information about their environmental, social and governance performance. Annual reports, sustainability disclosures, policies, governance documents and regulatory filings all contain pieces of the picture.

The problem is turning that information into data that can actually be used.

Much of the work is still manual. Researchers and analysts have to locate relevant disclosures, interpret them, extract the right figures, structure the information and check whether the result can be supported by the original source.

As ESG reporting grows more detailed and the demand for comparable sustainability data increases, that process becomes increasingly difficult to scale.

This is where artificial intelligence is beginning to change the equation.

Swedish climate-tech company ClearTraced is building technology designed to automate the collection, analysis and verification of ESG information from corporate disclosures. Its platform uses AI to identify relevant information, extract structured data and connect individual datapoints back to the underlying source material.

That last part is particularly important.

In many AI applications, producing a plausible answer is enough. In ESG, a number without supporting evidence can be difficult to trust. Users may need to know not only what the figure is, but where it came from and whether the original disclosure supports the interpretation.

ClearTraced’s approach therefore puts emphasis on traceability alongside automation. AI can handle much of the initial extraction and processing, while additional validation and human review are used when interpretation or judgment is required.

Among the people building that system is Pakistan-based engineer Abdullah Rashid, co-founder and CTO of ClearTraced, who leads the company’s engineering and AI development.

“The challenge is making the information reliable enough to use,” Rashid says. “Extraction is one part of the problem. You also need to understand whether the model interpreted the disclosure correctly, whether the data is consistent and where exactly the information came from.”

That distinction points to a broader challenge facing enterprise AI.

As companies move from experimenting with AI to using it in operational workflows, the standard for a useful system becomes much higher. The technology has to be capable of handling large volumes of information, but it also has to behave predictably enough for people to rely on its output.

ESG is a particularly demanding environment for that kind of system. Sustainability information can vary significantly in how it is presented, measured and described from one company to another. Relevant information may also be distributed across multiple documents rather than appearing in a single standardized dataset.

AI can help navigate that complexity, but automation alone does not solve the underlying trust problem.

ClearTraced has begun applying its technology in institutional settings, including work involving organizations such as Global Child Forum. The company has also worked with Swedish media and data organizations including Bonnier, Dagens Industri and Millistream.

Its collaboration with Global Child Forum, for example, involves AI-assisted data collection for a corporate children’s rights benchmark while retaining human methodological review. It illustrates a model that is becoming increasingly relevant across business applications of AI: using automation for scale while keeping people involved where context and judgment matter.

That human element may be particularly important during the current phase of AI adoption.

Businesses are still figuring out which tasks can be reliably automated, where oversight is necessary and how AI should fit into existing systems. In many cases, the immediate opportunity is not replacing an entire workflow, but making parts of it dramatically faster.

The same principle applies to ESG research.

Instead of asking AI to independently make a final judgment about a company’s sustainability performance, organizations can use it to process documents, identify relevant evidence and reduce the amount of repetitive work required from analysts.

The potential benefit is not simply speed.

If the resulting information is structured, continuously updated and connected to its source, it could become easier to compare companies, monitor changes and investigate individual datapoints when questions arise.

That could gradually change ESG data from something assembled periodically through manual research into something closer to an always-updated information layer.

The shift also illustrates where some of the most practical opportunities in enterprise AI may emerge.

The most valuable applications are not necessarily the ones that produce the most impressive demonstrations. They may be the ones that take work that is repetitive, time-consuming and difficult to scale and make it substantially more efficient without removing the human judgment that the process still requires.

For the ESG industry, that could mean spending less time searching through documents and more time interpreting what the data actually means.

The technology is still evolving, and questions around accuracy, validation and oversight will remain important. But the direction is becoming clearer.

ESG runs on data. Increasingly, the question is whether AI can help organizations collect that data faster without losing the evidence and accountability that make it useful in the first place.

That may be the more consequential shift happening beneath the surface of the AI boom.