Value-chain emissions are usually the largest part of a company’s footprint and the hardest to measure, because the activity happens in other people’s businesses. Good Scope 3 work is honest about which numbers are estimates.
Scope 3 emissions are the indirect greenhouse gas emissions in a company’s value chain, from purchased goods and transport to the use of sold products, grouped by the GHG Protocol into 15 categories. Scope 1 covers direct emissions and Scope 2 purchased energy. AI helps by classifying spend into those categories, reading activity and supplier data from documents, and recording the method behind each estimate.
Scope 3 emissions are difficult to measure because the data sits with suppliers and customers. Most companies start spend-based (money spent times an industry factor), move to activity-based data (kilograms, kilometres, kilowatt-hours) where they can, and use supplier-specific figures for the few suppliers that matter most. The system keeps those three tiers visibly separate, so a reader never mistakes an average for a measurement.
and activity data
GHG Protocol category
supplier · activity · spend
source and version recorded
and approves the inventory
queued for a person to classify
Under the GHG Protocol, Scope 1 is direct emissions from sources a company owns or controls, such as its boilers and vehicles. Scope 2 is indirect emissions from the electricity, heat or steam it buys. Scope 3 is every other indirect emission in its value chain, upstream and downstream, split into 15 categories from purchased goods to end-of-life treatment of sold products.
Because the activity happens in other organisations. Suppliers may not measure their own emissions, may use different boundaries or years, or may not respond at all. Spend-based estimates fill the gap but move with prices rather than physical activity. Reliable Scope 3 work therefore combines several data tiers and is explicit about which numbers are estimates.
By keeping the method stable and recorded: the same category mapping, the same factor sources with versions, and a clear note whenever a supplier-specific figure replaces an estimate. Otherwise a change in method looks like a change in emissions. The system stores each year’s inputs and factors, so a recalculation of a prior year is reproducible.
It can automate the classification, document reading and factor matching that take most of the time. It should not choose the boundary, the factor dataset or how to treat a supplier’s figure; those are methodological decisions your analyst makes and documents. Uncertain classifications go to a person rather than being guessed.
What we built, what stage it is at, and a link to the live thing wherever there is one.
yourOS is an operating system with an AI at its heart. You talk to it like a person, out loud or typed. It makes your calls, plans your days and works on your projects with a crew of agents, on your own private server. Our own product, built from the first commit; the waitlist is open.
Website, backend and app assets for Sonora, a free AI sound-therapy app on iOS and Android published by Cognitive AI.
Brand, website, prototype, apps on five platforms, the AI backend and the algorithm-intelligence pipeline for Podswap, a social growth engine that runs without a human in the loop.
In closed beta: Roxy, a private AI assistant with permanent memory running on infrastructure the user owns. Brand, site, product, backend and conversational design.
Live at coherence.com: brand identity generated with AI agents we directed, the motion system and website for the group behind Sonora, AboutAI, WakeUp and MudShadows. Agent work remains in build.
Live at domainbooth.com for DomainBooth, a premium domain brokerage: a new animated mark, every page redesigned and rebuilt with designed mobile layouts, and a custom CMS for its sales record, featured names, press and enquiries.
Bring a year of purchase ledger. We will show the category split, the factor behind each line and where supplier data would change the total.
30-minute technical call · no deck