AI belongs in the work, not just the interface.

Where the time actually goes

Most sustainability teams don’t lose their weeks to analysis. They lose them to opening invoices, re-keying meter reads, chasing suppliers, mapping spend lines to categories and checking that last year’s numbers still reconcile. A chat interface doesn’t remove any of that. It just answers questions about data that someone still had to prepare by hand.

Putting AI where the effort is

Purpose-built AI is most valuable when it does that preparation: extracting values from documents, classifying activity, flagging anomalies, matching records to sources and running the same checks every cycle. When that work is automated, the answers an assistant gives are grounded in data that is complete, structured and traceable.

The test for sustainability AI isn’t how fluent it sounds. It’s whether every answer can be traced back to source.

Keeping people in control

Automation only works if teams trust it. That means showing the source behind every figure, routing uncertain cases for human review and keeping a record of what was changed and why. The aim isn’t to remove judgement from sustainability work, but to make sure judgement is spent where it matters.