ESG Reporting When a Customer Asks for Data

A questionnaire arrives from a customer you have supplied for nine years. Forty questions about emissions, waste, diversity and governance, with a deadline and a note that responses feed their annual reporting. Nobody at your company has ever produced any of this, and the request is not optional in any way that matters.

Why it reached you

Large buyers increasingly report on their value chain, which means the emissions and practices of their suppliers form part of their own figures. They cannot produce that from their own records, so the question travels down the chain until it reaches businesses with no reporting obligation of their own.

Which explains the shape of the questionnaire. It is written for the buyer's reporting framework rather than for your business, which is why a quarter of it will be irrelevant to what you do and why answering it feels like translating.

Most of it is already in your files

What they ask for	Where it already exists
Electricity and gas use	Twelve months of utility bills
Fuel consumption	Fleet card and fuel statements
Waste and recycling volumes	Hauler invoices
Headcount, turnover, training	Payroll and HR records
Materials and volumes purchased	Supplier invoices and purchase orders
Policies and governance	Documents you have, somewhere

The work is collation rather than measurement, which is good news and also why it takes so long the first time. Twelve months of bills across three sites in four formats is a tedious extraction job, and it is exactly the kind of structured reading that automates well, as in extracting data from PDFs.

Where the records are on paper, the scanning step comes first and decides the accuracy of everything after it, which is covered in turning paper into something useful. Purchase records are the hardest to assemble in most small businesses, and they are considerably easier where purchase orders exist.

Say when you do not know

The temptation on a question you cannot answer is to estimate something reasonable and move on. Resist it, because the number you supply becomes part of a figure your customer publishes and may have assured.

Not currently measured is a complete answer, and buyers encounter it constantly. Where you intend to start measuring, say when. That reads as a supplier who understands the question rather than one who is hiding, and it is far safer than a figure you cannot substantiate if anyone follows up.

The same applies to policies. Claiming a formal policy you do not have creates an exposure that outlasts the questionnaire, and an honest we do not have a written policy, here is our current practice is both accurate and frequently sufficient.

Answer once, reuse many times

The second questionnaire from a different customer asks the same things in a different order with different wording. Businesses that rebuild the response each time spend the full effort repeatedly, which is how a tolerable task becomes an annual ordeal.

Keep a maintained answer set with the figure, the period it covers and the source document for each. Mapping a new questionnaire onto existing answers is a matching problem that automates well, and it turns a week into an afternoon.

Give it an owner and a refresh date, because a stale ESG answer set is worse than none: last year's consumption figure submitted as this year's is a misstatement rather than an approximation. That maintenance discipline is the one in document control.

Check every figure against its source

Extraction from bills and invoices is reliable on clean documents and less so on scans, multi-site statements and anything with a correction on it. A consumption figure that drifted during extraction is a wrong number in a customer's published report with your name attached.

Score per field rather than per document, keep the source reference beside each answer so it can be re-checked quickly, and have a person confirm the totals before submission. The general method is in quality assurance when AI writes the first draft, and the underlying record quality question is in data cleansing.

What you get out of it

Keeping the customer is the obvious one, and the questionnaire is increasingly a condition of the relationship for anyone supplying large buyers, public bodies or regulated chains.

The less obvious one is that most businesses doing this exercise properly find something. Twelve months of consumption data assembled in one place tends to reveal a site using far more than the others, or a waste line that grew without anyone noticing. Those are cost findings that arrived disguised as compliance work, and they belong on the same view as everything else you track, which is what a simple dashboard is for.

Frequently Asked Questions

Why is a customer asking a small supplier for ESG data?

Because they have reporting obligations that depend on their supply chain. A large buyer reporting on emissions across its value chain needs figures from the businesses it buys from, so the request travels down the chain to companies with no obligation of their own. Understanding that the question is driven by their requirement rather than by judgment of you makes it easier to answer the parts you can and decline the rest.

What can a small business answer from records it already has?

More than most expect. Electricity and gas consumption come off utility bills. Fuel use comes off fleet card statements. Waste volumes come off hauler invoices. Headcount, turnover and training hours come off payroll. Materials purchased come off supplier invoices. Those cover a large share of a typical questionnaire and require collation rather than measurement.

What should you refuse to answer?

Anything you would have to invent. If a question asks for a figure you do not measure, saying so is a better answer than an estimate presented as data, because the number you supply becomes part of someone else’s published report. Write not currently measured, and where it matters add when you intend to start. Buyers encounter that answer constantly and it is far less damaging than a figure that cannot be substantiated.

Can AI help with ESG questionnaires?

With the collation and the repetition, which is most of the work. Pulling consumption figures out of twelve months of utility bills, mapping a new questionnaire to answers you have already given, and keeping a maintained library of standard responses are all tasks it handles well. Deciding what to claim and verifying every figure against its source stays with a person, because these answers end up in a customer’s published reporting.

Is this worth the effort for a small supplier?

It depends on whether the customer asking is one you want to keep. For businesses supplying large buyers, public bodies or anyone in a regulated chain, the questionnaire is becoming a condition of the relationship rather than a courtesy. Doing it once properly and reusing the answers is far cheaper than reconstructing it each time, which is how most suppliers currently handle it.