Why your regulatory team loses days to customer compliance questionnaires

Picture the Tuesday morning. A product manager opens the inbox and finds three new questionnaires from three customers. All three want to know broadly the same things about the same product. None of them want it in the same format. One is a spreadsheet with locked cells, one is a portal that times out, and one is a PDF that has to be printed, signed and scanned.

The facts have not changed since last week. The product is the same product. Yet the work of restating those facts, in each customer's template, falls to whoever understands the chemistry well enough to answer without getting it wrong. In most UK chemical businesses that is one or two people, and their time is the scarcest you have. This pattern sits underneath the whole documentation burden in the chemical and process industries, and questionnaires are where it shows up first.

It helps to remember that none of this is your customer being difficult. Each request exists because the customer has a downstream obligation of their own, and your answer is an input to it. Understanding the driver behind each request is the first step to answering it once and reusing the answer.

The SVHC case is the clearest example of the driver. When a substance is added to the REACH Candidate List, businesses that supply articles containing it above the 0.1% by weight threshold have a duty to pass information down the supply chain under REACH Article 33. Your customer is asking so they can meet that duty. Confirm the current Candidate List and the exact wording of the duty against ECHA and HSE, because the list changes and the UK and EU regimes are maintained separately.

The work looks trivial from the outside. It is not complex. It is not creative. Most of it is looking up a fact you already hold and transcribing it into someone else's box. That is exactly the trap. Because the work is low-judgement, it feels like it should be cheap. Because it needs real product knowledge to get right, it can only be done by expensive people. High volume, low judgement, high skill required: that combination is what makes it hurt.

Put a rough number on it for your own business. The point is not the precise figure, it is the shape. Swap in your own numbers:

Take the middle of those ranges. Twelve hours a week, across roughly forty-six working weeks, is about 550 hours a year. Multiply by your own fully loaded hourly cost for the people actually doing it. For most SMEs the answer lands somewhere that makes people go quiet, and that is before you count the opportunity cost of those same people not doing the work only they can do. Use your own loaded rate rather than a headline salary, and treat the output as a range, not a single number.

Everyone has tried to solve this. The usual attempts help a little and then quietly fail for the same reasons.

• A shared folder of past answers. It works until a fact changes. Then nobody knows which of the forty saved answers are now wrong, and the folder becomes a liability instead of an asset.
• Templates. They save time only when the customer accepts your format. Most large customers insist on their own, so the template becomes one more thing to maintain rather than the answer.
• A generic AI chatbot. It does not know your products. It will write a confident, fluent answer that is not grounded in your data, which is worse than no answer because it looks right.

The pattern that holds up has four parts, and none of them is magic:

• A single verified source of truth for your product regulatory facts, owned and maintained deliberately rather than scattered across an ERP, spreadsheets and people’s memories.
• Drafts generated from that source, with a citation on every answer pointing back to where the fact came from, so a reviewer can check it in seconds.
• A named approver. A person signs off before anything leaves the building. The system drafts; the competent person approves.
• An auditable record of what was sent, to whom, and when, so that if a customer or an auditor asks later, you can show it.

Notice what is not on that list: replacing your experts, or trusting a machine to decide. The gain comes from taking the transcription work off your specialists while keeping the judgement firmly with them.

You do not need to buy anything to make this better. Three steps help even with no budget:

1. 1Write down the master facts once. For your top products, capture the answers to the questions you get asked most, in one controlled document, with a named owner and a review date. This is the source of truth you will build everything else on.
2. 2Log every questionnaire for a month. Which customer, which product, how long it took, which facts they wanted. A month of this data tells you exactly where the time goes and which requests are worth automating first.
3. 3Separate drafting from approving. Let a more junior person assemble the draft from the master facts, and reserve your specialist for the review and sign-off. Even without any AI, splitting these two jobs recovers a surprising amount of expensive time.

When you are ready to take the transcription off your specialists entirely, that same source of truth is what a retrieval system drafts from, with citations and a review gate. If your product facts also live in safety data sheets that drift out of date, fix that source first, because everything downstream inherits its errors.

Regulatory references in this article point to primary authorities. The Candidate List and the precise wording of duties change over time, so confirm the current position against the source before relying on it. This article is practical guidance, not legal advice.

• ECHA — Candidate List of Substances of Very High Concern
• HSE — UK REACH