# How AI finally joins sustainability to the business > The real work of sustainability is connecting what sustainability means to how the business actually operates. - URL: https://annwn.ai/insights/joining-sustainability-to-business - Author: Rory McCabe, COO (Annwn) - Published: 2026-09-09 - Updated: 2026-10-09 - Category: Insight - Reading time: 4 min ## **Research says most executives believe the business case for sustainability is clear, but only two in five can prove it. We have won the broad argument — boards don't dismiss sustainability the way they used to (I can remember when they did) — yet not enough money is moving, and more businesses than ever are cutting sustainability investment. We have the plans and the targets., but we find it much harder to make the case for action.** ## The gap It's easy to understand. Executive belief is cheap; capital allocation, operational change and procurement action all cost money. In an uncertain, somewhat less green world, the question of return needs an answer. The age of pure green ambition has passed. We in the sustainability profession know what needs doing; we've known for years. But the job is changing, from reducing impact to making the case for a return on investment. That's a very different ask for a profession that grew up focused on externalities. Now we have to translate those externalities into internal decisions, and that is just hard. Meanwhile, many of us spend our weeks reconciling data, chasing versions, fitting it to frameworks, answering RFPs, trying to improve the scores. When the decision meeting comes we have our reports and our compliant plans, but too rarely the business case for the proposal we know is right. That hard part is the one that  is often left undone.  For example, linking a procurement ledger to an emissions plan, or showing what sustainability action might do to a customer strategy, or modelling how decarbonisation investment impacts the P&L. These are hard for good people with good data — and, as budgets reduce,  there aren't going to be more of us. ## Why it's structural This isn't a failure of effort. Finance and procurement have ERP. Sales has the CRM.  HR has Workday. Those systems work because the data belongs to one function and one domain. Sustainability doesn't work that way, it crosses finance, procurement, operations, HR and suppliers, so the data we need doesn't sit in one box. *We are the only function asked to drive decisions across the whole business without a system that can see across the whole business.* We do have sustainability systems, and they give us the externalities — emissions, water, social risk. But they're built around the disclosures we must make: CSRD, TCFD, GHG Protocol. Necessary, but built to show the world what happened, not to help a business decide what to do next. So when we build a story from that data it comes out in emissions and decarbonisation, while every other function talks in cost, profit and revenue. We're not choosing the wrong language; we don't have the data to speak the right one. The translation needs someone fluent in sustainability, finance, procurement and the operations of the business. That's rare, so mostly the work doesn't get done and we stay in our silo — which is not the job any of us came to do. ## What is AI doing now? Everyone is being asked to use it, and most of us don't know where to start. Build, buy, chat, vibe? What about security? It has become: work out how to use this, and don't get it wrong. Where AI is being used in sustainability today, it's broadly one of two ways. The first bolts AI onto an existing system of record — the same reporting platform, the same incomplete data, with a chatbot on top. It can only reason over whatever survived that process: the spreadsheet nobody finished, the supplier who never replied. It knows what your number is; it doesn't know why. The second automates tasks — read a PDF, extract a figure, draft a paragraph. Genuinely useful, real hours saved. But pulling a number out of a document is not the same as reasoning across a value chain to find where the opportunity sits. *Both make the reporting faster. Neither helps us make the case.* ## What should it look like? Say the Head of Sales asks: are we exposed on our customers' sustainability requirements? Contracts have requirements, customers have policies we need to line up with, and sustainability may be in the next RFP. Not a compliance question — a revenue question. Answering it means joining things that live in completely different places: the contracts, CRM data on revenue and deal history, supplier scorecards, our customers' published commitments. Nobody has those in one place, so we skate across the top of the question. Join them and we get a view of every customer — where we meet the requirements, where we don't, and what the gap might cost. So we can say: “Roughly £25m of deals carry sustainability risk. We can reduce that to £5m if we act in the next six months.” A number the sales team gets straight away. The next question is tougher: what would it take to address that £20m? Now we're making the case for intervention — the fleet, the factory, the retrofit, the supply chain. What it does to capex and opex, to win probability, to the balance sheet. And it has to survive contact with the CFO: where did that assumption come from, who said the lease renews then, what's that win probability based on? Those answers aren't in any single system. They have to be built across sustainability, finance and operations. ## What we are building Neither answer comes from AI writing better prose. It comes from joining two things that in most businesses have never met. Our ESG system might know we use 800,000 litres of diesel. It won't know the fleet comes off lease in eighteen months, that two routes can't be electrified, that a factory upgrade is already in the plan, or that our largest customer has a 2030 requirement we're busy failing. Our business systems know all of that. Nowhere are the two joined up — and where they are, it's one-off, manual, the sources get lost, the assumptions contradict each other, and the multi-tab spreadsheet creaks the moment it's challenged. What's changed is that AI can finally read the unstructured material — the contract clause, the customer's commitment, the note in the capital plan — and work out what relates to what. That's the thing no ERP could ever do. That's what we've built at Annwn. Not another database, not a chatbot on top of a report, not an API. It's the app for the real job: somewhere secure where sustainability and business data sit together, connected, the dots joined, the full picture drawn — and every number in it traceable back to its source. We're sustainability people who happen to be software engineers. We're early, building with a small number of design partners, and we'd like a few more: teams with a genuinely tricky problem at the intersection of business and sustainability. It'll be a handful of businesses for now. We're ambitious about where it goes.