If you run sustainability at an SME, you already know the problem. Here’s the shift that changes it.

Every sustainability lead at an SME has had this conversation. A customer asks for the carbon footprint of a product. A regulator asks for a category-by-category Scope 3 inventory. An investor asks where the SBTi target lives. And you, the team of one or maybe two, have to find a way to answer with a spreadsheet, a part-time contractor, and a supplier base that doesn’t know what an emission factor is.

Scope 3 has been the wall. Not because anyone disagrees that it matters, it’s typically 70 to 90% of an enterprise footprint, but because the operating model has been broken. Until very recently, you had three options, and all of them failed in different ways.

Why the old playbook stopped working

Option 1: Spreadsheets

You start with a tab per category, optimistic about hierarchy. Six months in, you have 30+ tabs, broken VLOOKUPs, three versions named final_v3_REAL.xlsx, and no idea which emission factor vintage anyone used. Spreadsheets are great for analysis, terrible for state. Scope 3 is a state problem, with thousands of suppliers, products, and shipments changing every week. The math isn’t hard. Keeping the math current is impossible.

Option 2: Outsource to a consultancy

You pay between EUR 40k and EUR 200k for a screening. You get a PDF. The data inside that PDF is a snapshot of last year, built from spend-based estimates and industry averages, with no underlying model you can update. Customers ask for product-level numbers; the PDF can’t answer. Twelve months later you pay again.

The trap: the deliverable is a report, not a system. You don’t get more capable over time; you get a habit of buying reports.

Option 3: Buy enterprise carbon software

This works if you have a six-figure budget, a year for implementation, and a team to maintain it. Most SMEs have none of those. The category was built for FTSE 100 buyers, and the pricing, complexity, and onboarding requirements show it. You end up with a tool you can’t afford to use properly, configured by a consultant who left, sitting unused while you go back to the spreadsheet.

The pattern: every option is built around the assumption that gathering Scope 3 data is human work.

Spreadsheets assume a human stitches data together. Consultancies assume humans do it once a year. Enterprise software assumes a team of humans operates it. All three break for SMEs because the work is too expensive, too slow, or too brittle to scale to a one-person sustainability function.

What changes with agentic AI

The shift isn’t “AI helps you with Scope 3.” It’s deeper than that. Agentic AI changes who does the work. The agents don’t assist the analyst; they replace the data-gathering layer entirely, and the analyst moves to oversight, judgment, and engagement.

Concretely, that means:

  • Agents reach into source systems directly. ERP, procurement, freight portals, supplier specification PDFs, EPDs. They parse the documents, normalise the units, and map every line item to the right GHG Protocol category.
  • Agents engage your suppliers. Personalised requests in the supplier’s language, tied to the SKUs they actually ship to you. One-click upload. Automated follow-ups. Escalation paths if a supplier goes silent.
  • Agents assemble the inventory. Primary data where it exists. Intelligent, GHG-Protocol-compliant estimation where it doesn’t. Every estimate flagged with a confidence band and a remediation path to primary data.
  • Agents maintain the audit trail. Every number traces back to a source, a method, a timestamp, and a person who can be asked. Auditors stop asking for evidence – it’s already there.

The human role doesn’t disappear. It changes shape. Instead of being the bottleneck for data collection, you become the strategist for what to do with the data: where to push suppliers, which products to redesign, which targets to commit to, which customers to lead with. That’s the job sustainability leads actually want.

“The work that used to take a quarter takes a Tuesday. What you spend the rest of the quarter on is the part that actually moves the footprint.”
— Common refrain from SustainZone customers

“How do I know it’s real?” A Checklist

If you’re evaluating an agentic Scope 3 platform, here are the five questions that separate marketing from method. We use them ourselves.

1. Show me a single emission number, end to end.

Pick one line item in the inventory. Ask the platform to trace it: which invoice, which BOM, which supplier response, which emission factor, which version of the GHG Protocol guidance. If the answer is anything other than a click-through audit trail, walk away.

2. What happens when a supplier doesn’t respond?

Real platforms have a hierarchy: primary, activity-specific secondary, spend-based with confidence scoring. Each fallback should be flagged in the inventory. “We’ll get back to you” is not an answer.

3. Can the same data layer produce CBAM, EPR, SBTi, and CSRD outputs?

Scope 3 is the underlying data layer. If the platform can’t reuse it for CBAM filings, EPR jurisdictional reports, SBTi target validation, and CSRD disclosures, you’re going to end up buying four platforms. The whole point is to solve the data problem once.

4. Who owns the data?

You should. If the supplier engagement happens through their portal, the relationships should be yours; the data should export cleanly; the model should be re-runnable on your systems. “Data lock-in” is the second-most-expensive mistake in this category, after “no data at all.”

5. What’s the proof timeline?

Twelve-month implementations are not implementations — they’re long sales cycles. A real agentic platform should produce a credible first-pass inventory on one product line or one category of spend within 30 days. If it can’t, the agents aren’t doing the work.

Where to start

Figure: Overview of GHG Protocol scopes and emissions across the value chain (Scope 1, 2 and 3).
Source: GHG Protocol (WRI & WBCSD), Corporate Value Chain (Scope 3) Accounting and Reporting Standard (2011).

Don’t start with all 15 GHG Protocol categories. Start where the pressure is loudest:

  1. If you’re exporting to the EU and have any covered goods (steel, aluminium, cement, fertiliser, hydrogen, electricity), CBAM is your first wedge.
  2. If your top customers are asking for product carbon footprints, start with the SKU catalogue covering 80% of revenue.
  3. If you’ve signed an SBTi commitment letter, work backwards from the validation deadline.

The work compounds. The same agents that build the CBAM dataset feed the customer PCFs. The same supplier relationships that supply primary data for one product supply it for the next. The same audit trail satisfies the auditor for limited assurance, the regulator for filings, and the customer for tender responses.

The takeaway: Scope 3 isn’t unmeasurable; it’s unstaffable with the old playbook. Agentic AI changes the staffing model, and the wall comes down. You don’t need a bigger team. You need agents that do the data work, an audit trail by default, and a platform that turns the same data layer into every disclosure you owe.

About SustainZone

SustainZone is the agentic AI platform for the value chain. We help SMEs measure, report, and reduce Scope 3 emissions and reuse the same data layer for CBAM, EPR, SBTi, and CSRD obligations.

Every number is traceable to its source. Every update is automated, reducing manual work and improving reporting confidence.

Start with a 30-day proof of value at SustainZone.