DitchCarbon vs CO2 AI: two routes to supplier-specific Category 1 data

DitchCarbon and CO2 AI both collect primary emissions data from suppliers, and both work with product carbon footprints. So the useful comparison is not which one does those things. It is how much of the work you can finish without asking a supplier anything at all, and what the platform gives you once the baseline exists.

DitchCarbon provides verified emissions data for over 2 million organisations, so procurement, sustainability and finance teams can measure and act on supply chain and portfolio emissions from one source.

How much can you do without sending a survey?

With DitchCarbon, often all of it. A company-specific figure already exists for a great many of the organisations on a typical Category 1 list, so you can rank the spend by embodied emissions, calculate the baseline and start acting before a single request goes out. A request is a tool for closing the gaps, not the way the data arrives.

That is worth more than the convenience. A supplier's sustainability team is usually one or two people, and every buyer running its own programme asks them the same questions in a slightly different order. Time spent on the fifteenth version of the same form is time not spent cutting emissions. Any model that has to ask every supplier before it knows anything adds to that queue, whoever is running it.

CO2 AI collects through a supplier portal. The supplier is given access, logs in, and compiles a submission: primary emissions data, product footprints or life cycle assessments. The access costs the supplier nothing, which is worth stating precisely because the cost was never the licence fee. It is the hours. Nothing reaches the buyer until somebody at that supplier sits down and does the work, so coverage is bounded by how many suppliers will log in and how much time each of them has.

The DitchCarbon equivalent is not a better portal. It is a profile the organisation owns. It arrives prepopulated from reports they have already published, they can correct it, and any buyer who asks reads the same answer, so the work is done once rather than once per customer. Charging an organisation to share its own emissions data is a model DitchCarbon will not run: profiles, data updates, reduction recommendations and a calculator for organisations measuring for the first time are free to the organisation, permanently.

Where a request is still needed, it goes out from the portfolio view, prepopulated, addressed to the named sustainability contact at that organisation rather than a generic procurement inbox. The supplier is reviewing and correcting rather than compiling, which gets a higher response rate than a cold survey for the obvious reason: the work has already been done and they are confirming it.

Where does the Category 1 figure come from before anyone is asked?

Here the two are structurally different, and it is the reason the collection models differ.

CO2 AI holds emission factors. Its published library is over 110,000 emission factors covering materials and industrial processes. Nothing in its public material describes a database of company-level emissions. Customer activity, ERP and spend data comes in through connectors and gets matched to factors, and company-specific figures arrive when the customer supplies them or the supplier responds.

DitchCarbon holds company records as well as factors. Verified emissions data for over 2 million organisations, including corporate-level GHG inventories, with supplier-specific emissions at spend, activity and product level and a generic factor library underneath drawn from ecoinvent, CEDA, EPA, DEFRA and EXIOBASE. Each organisation carries fifteen years of reported history, eighteen attributes deep, with restatements merged into a single series rather than left as competing versions, and the assurance status recorded per scope against the source document it came from.

Matching a spend line to the right legal entity is the first step rather than the product. What it buys you is a ranked list on day one, a Category 1 figure you can calculate immediately, and a shortlist of the organisations worth a conversation. One toggle switches the headline between organisation-specific factors and industry averages, and the difference between those two numbers is the honest measure of how much of your baseline rests on what organisations actually published.

Which changes what engagement is for. If a platform holds nothing about a supplier until that supplier answers, every supplier has to be asked. If a company-specific figure already exists for most of the list, the request goes only where it will improve something, and the maturity ladder shows where that is: every organisation sits on one of four steps, from not disclosing through to publishing product carbon footprints, sorted against embodied emissions, so a large supplier on the bottom step is immediately visible.

One comparison to avoid, because it gets made constantly and it is meaningless: a factor count set against a company count. Those measure different things.

How do product carbon footprints get handled?

Both platforms work with product footprints, in opposite directions.

CO2 AI builds them, producing PCFs for a manufacturer's own products at SKU level, aligned to PACT, PEF, TfS and ISO 14067, and exchanging them over the PACT network.

DitchCarbon holds and grades them. The product carbon footprint library carries published footprints for an organisation's products, each row with its total kgCO2e per unit, the standard it was prepared to such as ISO 14044 or ISO 14067, the third-party verifier where there is one, whether the boundary is cradle to gate or cradle to grave, and a quality score. That boundary field is not housekeeping: a cradle-to-gate figure and a cradle-to-grave figure for the same item are different numbers answering different questions, and mixing them quietly corrupts a category total. Where a footprint exists for something you actually buy, it feeds the calculation directly through spend allocation, and the emission factor method marker shows exactly which rows used a product footprint instead of a spend figure.

There is also a check on the way in, and it does more than test the format. The PCF Evaluator takes a supplier product carbon footprint or an EPD and returns a graded report. It opens with an overall quality band and a plain verdict on whether the footprint can enter a calculation, then grades data quality, whether the figure sits within the expected range for what is being described, and PACT alignment scored as required fields present out of the full set. Underneath sits an assurance breakdown: whether the calculation methodology is disclosed, whether the lifecycle boundary is broken out beyond the headline number, and whether an independent verifier has signed the footprint off. Each dimension comes back marked good or needs review, so a footprint that looks fine on its cover page and thin underneath does not get through on appearance.

A worked example is public. The evaluation of Ferro Chrome from Tata Steel comes back high quality and ready to use at 5,310 kgCO2e per 1,000 kg, with 23 of 38 PACT fields present and 4 of 8 assurance checks met. The report is shareable, exports as a PDF, and comes out as structured JSON so the verdict travels with the footprint into your own systems instead of living in somebody's inbox. It is free and needs no registration.

For a team measuring Category 1, that direction matters. You are not usually publishing footprints. You are receiving them from suppliers of varying rigour and deciding which ones can enter a baseline.

What happens after the baseline?

This is where a comparison usually stops, and it should not, because a baseline nobody acts on is an expensive spreadsheet. Everything below runs off the same company layer, on the organisations you already buy from.

Forecasting and targets

The forecast runs an organisation's actuals forward on a two-year and a five-year trend, out to 2050, and draws alongside them the trajectory that organisation's own stated target requires. The gap between trend and trajectory is the answer, and it is switchable by scope, absolute or intensity. Where the underlying Scope 3 reporting has missing years, large swings or patterns suggesting a reporting change rather than a real movement, a notice appears above the forecast telling you to treat it with caution, linked to the disclosure it was built from. Being told when not to trust a chart is worth more than a chart that never warns you.

Worth saying plainly, because it is the kind of thing vendors blur: forecasting is projection rather than measurement, and the UL Solutions verification that covers the calculator does not extend to it. What is projected is the organisation's own reported history carried forward.

Targets sit in a register rather than a badge. Each one carries its scope, on-track status, classification such as absolute or net zero, base year, target year, reduction rate and its source, with SBTi-validated targets marked separately from self-declared ambitions and the original wording quoted in full. Where an organisation has published several overlapping and occasionally inconsistent goals, all of them are shown rather than the most flattering one. Across a portfolio, SBTi progress can be weighted by share of embodied emissions, share of spend or share of organisations, and the three give different answers: weighted by count a portfolio can look poor while the organisations carrying most of the footprint are all committed.

Deciding what to do about it

Every organisation and emissions category in the portfolio is ranked by the share of your total footprint that acting on it would move. Each line carries a realistic reduction percentage, the tonnes behind it, and a named peer already reducing at that rate. Peer intelligence sits behind that: a comparable organisation, the scope or Scope 3 category it cut, the reduction achieved on a two or five year trend, and the specific action, each carrying the published report it came from. The peer set splits into direct competitors and similar companies, and it is editable, because your view of who you are compared against is usually better than ours.

None of that is generated advice. It is drawn from what organisations have published about reductions they actually made.

Scoring and coverage

Every organisation carries a score from 0 to 100 with its industry benchmark beside it, built from emissions intensity, disclosure quality, climate commitments and the direction of travel over time. The breakdown is the useful part: six groups covering industry, region, reporting, initiatives, reductions and reporting quality, each row naming the criterion, its weight, the points earned out of those available, the source consulted and the value found. Filtering to unearned criteria turns a scorecard into an engagement list.

Across the portfolio, disclosure coverage runs ten measures as a percentage of organisations: whether a disclosure exists at all, whether they disclosed to CDP, whether near-term and long-term targets are set, whether Scope 1, 2 and 3 data exists, and whether the relevant upstream and downstream categories are disclosed. Coverage gaps are shown rather than hidden, and the thin bars are the list of who to engage first.

DitchCarbon and CO2 AI at a glance

Table compiled August 2026 from each vendor's public documentation. Capabilities change, so check the current position with any vendor before a decision.

What happens at supply chain scale?

DitchCarbon runs supply chains of over 100,000 suppliers. That is worth stating because it settles the question a large manufacturer or pharmaceutical group asks first, which is whether the approach survives contact with a real supplier base rather than a pilot.

The list a procurement team actually holds is a spend export, not a clean supplier register. It has duplicates, trading names and entities that no longer exist. DitchCarbon takes it as it is. There is no cleanse to run first and no integration project to schedule, which matters because the implementation is where these programmes usually lose a quarter and a budget line. Upload the file and you have a ranked list the same day, with the import history recording how many rows were processed so a partial match shows up as a number rather than as a gap discovered later in a baseline.

Scale is also where the two collection models separate hardest. A portal-based programme is bounded by how many suppliers will log in, which holds up across a few hundred large, engaged counterparties and thins out fast down the tail. A prepopulated request against a standing database does not depend on the tail responding at all, because a company-specific figure is already there for a great many of them, and the request is reserved for the ones where it changes the answer.

Which of the two publishes independent verification of its calculation?

One does. DitchCarbon's Portal calculator is verified to ISO 14064-3 at limited assurance by UL Solutions, renewed annually, and DitchCarbon was the first company to earn UL Solutions' Sustainability Information Calculator Verification, in June 2025. Separately, DitchCarbon data has been used in emissions reports subsequently assured by ten different third-party assurance providers, including Big Four firms.

CO2 AI publishes no evidence of independent verification of its calculation engine. What it publishes is security material, described as SOC 2 aligned practices and ISO 27001, and statements that its computations are traceable and verifiable by external auditors. Auditable is not audited: the claim is that someone else's auditor could check the output, not that anyone has checked the engine.

It is worth knowing what PACT conformance does and does not cover, since it appears in this market as though it settled the question. PACT describes a conformant solution as one that has completed the technical interoperability test, so it can share and receive product carbon footprint data by API, and PACT's own documentation for the conformance tool states that it does not evaluate any aspect of the calculation engine. PACT also declines the word certified, using conformant instead. So conformance tells you a footprint can move correctly between systems. It does not tell you the footprint was calculated correctly.

When the auditor arrives, the output matters as much as the declaration. DitchCarbon exports one Excel workbook ordered the way an auditor asks: totals by methodology, a summary per organisation, the line-by-line calculations and the emission factors applied, with every figure carrying its source and change history. The calculation method breakdown states what share of the total came from organisation, industry, product and activity data, so data quality is a proportion you can point at rather than a claim you have to make.

The full working on verification across this market, including every vendor checked and the evidence for each, is in our benchmark, Who's really verified? A reality check on carbon software assurance. The underlying documents are downloadable from the DitchCarbon trust centre.

Ask either vendor for the declaration itself, not a page that says audit-ready. Audit-ready without a named verifier and a named standard is marketing.

Where the two products stop overlapping

Two differences are structural rather than competitive, and both are worth stating flatly.

CO2 AI is a system of record for the full greenhouse gas inventory, across all 15 categories, with CSRD, CDP, GRI and SBTi output. DitchCarbon is not, and not by omission: it builds the Scope 3 baseline from your vendor list and feeds whatever you already report from. If your reporting stack works, replacing it is the expensive way to fix a data problem.

CO2 AI builds product carbon footprints for products you manufacture, to publish to your own customers. DitchCarbon works the other way round, holding and grading the footprints of what you buy. Which direction you need depends on whether your customers are asking you for footprints or you are asking your suppliers for theirs, and plenty of companies are doing both.

Can you run DitchCarbon alongside CO2 AI?

Not really, and it would be dishonest to suggest otherwise. The two overlap too much to sit on top of each other: both hold supplier data, both engage suppliers for more of it, both work with product carbon footprints. Running them together means paying twice for the same layer and then reconciling two answers for the same supplier. This one is a choice.

Where that layering does work is underneath the broader sustainability and ESG data platforms, which are a different shape of product: Watershed, Sweep, Microsoft Sustainability Manager, SAP. Those are systems of record for the corporate inventory and the disclosure, and they take a supplier data layer underneath rather than building their own. Run that way, DitchCarbon feeds them, Category 1 and Category 2 start from company-specific figures instead of sector averages, and nothing about the reporting stack has to change.

Run on its own, DitchCarbon is the platform. Measure the organisations you buy from or invest in, engage them for better information, forecast against their own targets, then track performance at counterparty, group, account or portfolio level.

What if you are the supplier being asked?

Many companies reading a page like this are on both sides of the request. You are measuring your own suppliers, and a customer is asking you for the same thing, possibly through a portal.

If that is you, your DitchCarbon profile is already there and already carries your published reports, listed by year with the assurance status found for each scope. Claiming it lets you correct what we hold, add reports or certifications by drag and drop, and answer the questions buyers ask most often once, so any logged-in viewer from another organisation reads the same answer rather than emailing you for it. A disclosure to-do list shows year by year what we hold on you, marked complete or partial, and names the categories buyers expect but cannot find. There is no fixed form, a file in whatever format you already hold is accepted, and it is free. Claim your profile.

How do you test the difference?

One test settles it. Take a real sample of your Category 1 suppliers, weighted the way your spend is weighted so the tail is properly represented, and ask both vendors the same question: how many of these can you give me a company-specific figure for today, before any supplier is contacted, and where did each figure come from?

Then ask the second question, which matters just as much. Of the suppliers you would need to contact at all, how many are there, what does the request look like when it arrives, and how much of it is already filled in?

Run the pilot with both. DitchCarbon will do its half free: send a sample of your supplier list to the coverage check and you get back what already exists for each organisation, the source behind every figure, and the gaps marked rather than filled in quietly. That is the honest comparison, because it happens on your list rather than on a vendor's reference set.

DitchCarbon will show you what it already holds on your list, with the source and change history behind every figure. Numbers you can defend within 2 weeks.

Last reviewed August 2026.