The sustainability report was never the point

For most of the past decade, the hardest part of corporate sustainability was not deciding what to do about climate change or labour rights. It was assembling the paperwork. A large company's sustainability team spent its year sending data requests to procurement, finance, HR and operations, chasing suppliers for numbers, reconciling spreadsheets that disagreed, calculating emissions, running a materiality exercise, mapping the results to a disclosure standard, and pushing a draft through internal review and assurance before publishing a document few people outside the reporting world would read closely. The function became the place where fragmented information from the rest of the business was collected and turned into a report.

 

Two developments are now changing that workload at the same time, and the more interesting question is what the function will do with the room they create.

 

Automation and simplification

 

Take the technology first. The progression runs from spreadsheets to sustainability software to connected data systems and, lately, to AI. The shift that matters is not that a chatbot can write a paragraph of disclosure text. It is that these tools are moving upstream into the work itself. Reporting platforms now ship AI agents that draft and check disclosures against the ESRS and ISSB standards, produce gap assessments and return a compliance scorecard. Vendors and audit firms increasingly describe tools that trace where a number came from, classify supplier documents, read invoices and flag anomalies in a metric before a human notices. PwC's 2025 survey of 496 executives across 40 countries found that among companies already reporting, use of AI for sustainability reporting had risen to 28%, from 11% a year earlier, most commonly for drafting disclosures, identifying risks and pulling data together from different systems.

 

The second direction is regulation, and it is easy to describe too kindly. On 3 July 2026 the European Commission adopted a revised set of ESRS that cut mandatory datapoints by more than 60% and total datapoints by more than 70%, changes it expects to lower reporting costs by more than 30% per company, against its own 25% burden-reduction target. The revised standards still had to clear European Parliament and Council scrutiny before taking effect, and apply from financial year 2027, with early adoption permitted. The official language is about focus: shorter standards, fewer items of marginal use, a simpler materiality test. The context is blunter. This is part of the Omnibus package pushed through after sustained complaints about European competitiveness, and it does more than trim duplication. The new scope threshold, €450 million in turnover and more than 1,000 employees, removes thousands of previously in-scope mid-market companies from mandatory reporting altogether, and the earlier plan to raise assurance toward the standard applied to financial statements was dropped, leaving limited assurance as the only requirement. Less reporting, yes. But also, in places, less scrutiny.

 

Why the data was collected

 

Put the two together and the same conclusion arrives from both. If the machinery around reporting is getting cheaper and the rulebook is getting shorter, the report stops being the point. It never should have been. Companies did not start measuring their emissions, water use and supplier exposure because they enjoyed producing PDFs. They did it, in theory, to know things: where the carbon actually sits, which suppliers carry human-rights or environmental risk, which sites are exposed to water stress or flooding, whether a transition plan is on track. All of that is useful before it reaches a disclosure, and most of it stays useful even if the disclosure requirement goes away. More than two-thirds of companies already reporting under CSRD or ISSB tell PwC they get value from the data beyond compliance.

 

On one hand, a decade of reporting built a data foundation companies can now use for other things. On the other, regulators have just deleted most of the datapoints. The two are talking about different kinds of data. The datapoints that survived simplification are, roughly, the ones with an operational reader in mind. The long tail that got cut produced effort without a decision attached to it. If that is right, the reform does not shrink the useful foundation so much as strip the packaging off it.

 

Where the data could be used

 

The opportunity is to use this information while the decision is still open, rather than describing it a year later.

 

• Procurement: a supplier's emissions and labour record can shape which supplier wins the contract and what the contract demands, rather than being tallied after the purchase order has gone out.
• Operations: unusual energy or water use at a single site can be caught while a plant manager can still act, not summarised once a year.
• Supply chain: climate and nature exposure can point to the vulnerable suppliers, commodities and routes before a drought or a flood interrupts them.
• Finance: the same climate, energy and transition data can inform whether to re-fit a plant or wave through an acquisition, instead of sitting apart from the investment decision.

 

The supply-chain case is not hypothetical. A 2026 Capgemini survey of 1,000 senior leaders found that 94% said climate-related disruption had already cost their company money over the past two years, almost a third of them more than a million dollars. PwC's own survey points to companies already using sustainability data for capital investment, supply-chain planning and physical climate risk.

 

The question changes: less "what do we have to disclose," more "what does this tell us to do."

 

The promise has been made before

 

There is reason to be cautious. The promise that a company's data would finally get used is not new. It was made for enterprise resource planning, then for governance and risk platforms, then for sustainability software itself, and in most companies the data mostly did not get used.

 

Nor has reporting actually become lighter yet. In PwC's 2025 survey, 66% of companies said the resources going into sustainability reporting had risen over the previous year, and 65% said senior leaders were spending more time on it, with only about one in twenty reporting any decrease. The tools are spreading faster than they are saving anyone work.

 

Confidence in the data appears to be getting worse, not better. In the Capgemini survey, 79% said their sustainability data was insufficient to inform strategy, up from 53% two years earlier. That number is moving the wrong way even as the tools multiply. Whatever AI is doing so far, it is not making the underlying data trustworthy enough to act on. If anything, leaders have grown more aware of how far short it falls.

 

Part of that is a data-quality problem and part of it is a limit on what automation can do. AI can classify a document and reconcile two datasets. It cannot decide whether to close a high-emitting plant, drop a supplier, absorb a cost to cut a risk, or rewrite a transition plan. Those are judgment calls with trade-offs, and someone has to own them. The reporting itself still rests on human choices about materiality, boundaries, estimates and assumptions, and an AI-drafted disclosure still has to survive assurance and, eventually, a regulator or a plaintiff's lawyer. Feed poor data into a capable model and you get wrong answers faster and with more confidence, which is worse than slow ones. The guardrails to prevent that are thin: in the same survey, only 40% of companies had a formal framework for how AI is used and who is accountable when it goes wrong.

 

What the team does instead

 

If the drudge work shrinks, the shape of a sustainability team should change with it. Less time chasing data, consolidating spreadsheets, cross-referencing frameworks and reformatting the same tables every year. More time interpreting what the numbers mean, sitting with procurement and operations, modelling scenarios, working through transition plans and engaging suppliers. The sustainability professional becomes a translator between the data and the people making decisions, and the walls between sustainability, finance, risk and procurement get lower.

 

There is already a sign of the gap this has to close. EFRAG's 2026 review of 905 assured sustainability statements for financial year 2025 found companies identified, on average, 6.4 material sustainability topics but had set measurable targets for only 3.3 of them; nearly two-thirds tied sustainability performance to executive pay, and the share disclosing a climate transition plan rose from 55% to 69%. Companies are getting good at naming what matters. Acting on it is another matter. They are identifying more material issues than they set measurable targets against.

 

The freed time won't allocate itself

 

That leaves sustainability teams with a harder question, one many have been waiting to answer without saying so. For years the case for more resources was that reporting ate the time that could have gone into actual programmes. AI and a shorter rulebook are about to test that case in public. If producing the report really does get much cheaper, where does the freed-up capacity go? Into supplier decarbonisation, climate adaptation, product redesign, energy and water efficiency, the transition plan? Or does a finance director look at a team that now spends far less time on disclosures and simply cut it, keeping sustainability where it has often sat, as a compliance cost?

 

I would not bet confidently on the generous outcome. The honest reading is that reporting labour was never the real bottleneck. The bottleneck is whether anyone acts on what the reporting reveals, and clearing a few analysts' calendars does not fix that by itself. It removes an excuse. What replaces the excuse is a management choice, and some companies will make it well while others quietly downsize the function and call it efficiency.

 

The reversal at the centre of this is simple to state. The old model collected sustainability information in order to produce a report. The emerging one runs the business on that information and lets the report fall out of the system, almost as a by-product. Reporting does not vanish in that world. Investors, customers and regulators will still want credible numbers, and the ISSB's standards are being adopted in more than 30 jurisdictions. But the report stops being the reason the data exists.

 

The real test of a sustainability function, once the reporting gets easy, will be how often the numbers in its annual report tell management something it did not already know and had not already acted on. In a company that has this right, the answer should be: almost never.