A research-based diagnostic of your AI potential: we take a workflow — the double materiality assessment, say — cut it into clean sub-tasks and assess each one. In the end you know which sub-tasks are worth it for AI, which aren't, and what it takes.
The question isn't whether AI helps in the sustainability function. It's: at which sub-task of a workflow — and how far.
"40% of sustainability work is automatable" sounds catchy and misleads. AI potential is decided at the sub-task level: a workflow contains sub-tasks a model handles reliably — and ones that stay with the human: judgement, stakeholders, accountability.
So we don't judge the workflow as a whole — we cut it into sub-tasks and assess each one.
We start with a workflow and break it into sub-tasks. Each sub-task passes through four checks — only when all four mark it suitable, and the operationalisation (Stage 0) is in place, is it genuinely AI-ready.
The dimensions are synthesised from the literature — grounded in 41 institutions across six groups: academia, AI labs, standard-setters, consultancies, law, international bodies.
Does a sub-task stay with the human — or does AI take it on, and how far? The more of the four checks a sub-task passes, the more independently AI can run it: from 0 of 4 (human) to 4 of 4 (AI as "Expert"). Plus the controls it takes.
A workflow, cut into sub-tasks and each one assessed:
A compact, fixed scope: in three weeks from three workflows of your choice to a clear AI roadmap. Fixed scope, fixed price — you know up front what you get.
The diagnostic stands on its own. If you want to go further, on request come build-out and ongoing operation — you decide after each step.
Cut workflows into sub-tasks, assess AI potential, build the roadmap.
You start hereImplement the prioritised workflows in your systems — through to handover.
Sparring after handover: model and regulatory updates, new use cases.
In a no-obligation intro call we talk through where you stand and where your biggest AI potential might sit. If it fits, I'll sketch a possible approach.