� COVER STORY �
Sustainability teams are tasked with driving decarbonization, managing community engagement projects( such as recycling, renewable energy, green technologies, human rights, and labor standards), and keeping up with a flood of incoming requests. These often arrive in different formats and languages, sometimes with poor translations or slightly varied phrasing even when the core question is similar.
Questions might include:“ Is your company regularly reporting its GHG( greenhouse gas) emissions?”,“ Do you have a carbon-neutral steel offering available today or planned in the future?” or“ How do you ensure your supply chain is free from child labor?”
Multiply that by hundreds of variations, and it is clear why sustainability teams are stretched. Responding to the most important stakeholders often means putting other strategic initiatives on hold, simply due to the time and resources required.
In the era of uncertainty and constant changes in sustainability demands, how can this challenge be solved?
One option could be to create a single, detailed standard that covers all stakeholder requirements. But we all know
Criteria |
Without AI |
With AI |
Time spent |
High |
Reduced by 50-70 % |
Manual effort |
High |
Low |
Risk of error Moderate-High Low
Collaboration
Insights
Scalability
Slower, more back and forth Reactive, based on human recall
Difficult
The table above compares key differences between working on sustainability data disclosure with and without AI.
how that usually goes: try to unify 14 standards, and you end up with a 15th. While there are some moves toward simplification, the regulatory and stakeholder landscape is still fragmented. So rather than aiming for full unification, what is actually needed is something that can adapt, something flexible enough to transform existing company data into the different formats required, while also flagging gaps and inconsistencies.
This is where tailored AI tools can help. AI can reduce the manual work of copy-pasting from one document to another. More than that, it can analyze incoming requests, spot patterns, and carry out gap assessments based on internal data. But this does not work out of the box. It requires a solid internal setup, clear disclosure rules, and the expertise of the sustainability team to ensure that what is shared is accurate, relevant, and safe, especially when it comes to sensitive information.
Streamlined within the AI tool Proactive: flags trends, gaps, and risks Easily scalable across requests and frameworks
So how does AI help and what are the limitations?
First, AI brings clear efficiency gains. Large volumes of data can be processed in minutes, which significantly reduces the time teams spend on manual tasks. It also handles documents in different formats and languages, making it easier to analyze and compare incoming requests no matter how they are structured or translated. Another key advantage is its ability to turn scattered information into usable insights, helping teams make sense of what is being asked and what is missing, and use these insights in strategic planning. At the same time, there are important limitations. AI still lacks deep contextual understanding. It cannot replace the judgment of someone who knows the business, the material issues, and the company regulatory context. Working with sensitive data also brings risks; companies need to have clear rules and secure systems in place to control what gets shared and with whom. Finally, AI models are only as
Green Steel World | Issue 18 | June 2025 7