Answer customer questionnaires from your own files.

Like ChatGPT — but it only answers from your own documents: safety data sheets, REACH files, specifications and procedures. Every answer shows where it came from. If it isn't in your files, it says so.

For HSE, regulatory and quality teams at UK chemical and formulation companies · 14-day free trial · No credit card · From $39/month

Compliance noteUnder UK REACH, if an article you supply contains a Candidate List substance above 0.1% by weight, you must tell your business customers — and answer a consumer's request within 45 days.Source: HSE, UK REACH substances of very high concern; HSENI (checked 2026-09-23)

QUESTIONS
DocumentsSDS · SOPs · specs · REACH
Items4, written not captured
AnswersFrom your own files, cited
THE QUESTIONNAIRE IS DUE
You ask
A customer questionnaire asks for the transport classification of DG-40. What does its SDS say?
You get
Section 14 of that safety data sheet quoted — UN number, class and packing group — with the page named.
From
Safety data sheets, one current revision per product
HOW DO WE CLEAN IT?
You ask
What does our work instruction say about cleaning the blending vessel between batches?
You get
The cleaning steps quoted from the work instruction, cited to the document and page.
From
Standard operating procedures and work instructions
WHAT DOES THE SPEC SAY?
You ask
What purity does the customer specification for PR-210 require?
You get
The requirement quoted from that specification — or a plain “not in your files” if you don't hold it.
From
Customer specifications and technical data sheets
NOBODY REMEMBERS WHY
You ask
The chemist who wrote this work instruction has retired. Why does it specify this cleaning sequence?
You get
The passage that states the reason, cited to the instruction or the validation record it came from. If only the step was written down and not the reasoning, it says so rather than inventing one.
From
Work instructions, validation records, change notes

Where the hours go

Customer questionnaires never stop
They arrive continuously, each in a different format, asking about composition, SVHC content, classification and transport. The answers are usually already in a safety data sheet or a declaration you hold — finding them is the slow part.
Every supplier's SDS reads differently
Safety data sheets share sixteen sections, but not a layout. Pulling the same fact from thirty suppliers' sheets means reading thirty documents.
A substance joins the Candidate List
Somebody has to work out which products and data sheets mention it. Today that means opening them one by one.
The knowledge walks out of the door
When the technical manager retires, knowing where everything lives often goes with them. The documents stay; the map doesn't.

What happens to your questions

Four kinds of question, and exactly what it does with each — including the one it can't handle yet.

ItemYou askSureCiteAIStamp
01What a section of an uploaded SDS, procedure or specification saysAnswers, and names the document and passage it came from.VERIFIED
02Regulatory lists or limits you haven't uploaded, such as the current Candidate List“I can't answer that from the uploaded documents.” It doesn't recite standards or rules from memory.NOT IN EVIDENCE
03Another company's filesIt can't reach them. Each company's documents sit in their own isolated workspace.SEALED
04Which revision of an SDS or assessment is currentNot tracked yet. Keep superseded revisions out of the workspace for now.OPEN

Upload these first

A trial is most useful when it holds the documents people already ask you about. Start here, then ask your first question.

  • Standard operating procedures and work instructions
  • Quality, environmental and safety manuals (ISO 9001, 14001, 45001)
  • Customer specifications and technical data sheets
  • Certificates of analysis
  • Safety data sheets — one current revision per product; archive superseded ones outside the workspace
  • COSHH assessments, REACH and SVHC declarations

What it won't do, and what it does instead

What it will not do, what happens instead, and where each one is a limit of document search in general rather than of this product.

It won't tell you which revision is current
If both revisions are in the workspace, you can be handed the superseded one — and the citation will look right, because that document is real. Revision tracking is on the roadmap. Until it ships: every answer names the document and page, so you are checking a named file rather than trusting a paragraph, and keeping one current revision in the workspace — superseded ones in a separate folder — makes the cited one the current one.
Also true ofNothing consumer-grade does this either. ChatGPT has no way of knowing which of two files you uploaded supersedes the other, and your shared drive will hand you a superseded safety data sheet without comment. What you get here is the document and page behind every answer, so a superseded file is visible instead of silent.
It won't promise you a complete list
Ask for every SDS that lists a Candidate List substance and what comes back is the most relevant passages, not a guaranteed complete set — treat it as a starting point, not a final count. It says so rather than handing you a short list that reads like the whole one. Ask about a single job, certificate or product and it answers from the passages that match, with the page.
Also true ofEvery search you already use works this way: ChatGPT answers from what it found, and so does the search box on your shared drive. Neither one warns you that the list may be short.
It hasn't been tested on chemical documents yet
Our published tests cover legal contracts, FDA drug labels and SEC filings. The FDA drug-label set is the closest match to a safety data sheet, but we haven't benchmarked chemical documents. Try it on your own files during the trial and check the cited pages.
Also true ofWe have not found a chatbot that publishes how it performs on documents like yours. Ours are public, including the hard suites where we score badly — which is why we can tell you this has not been tested yet.
It isn't a validated GMP system
For pharmaceutical work, use it to find information, not to hold the record: it is not validated to EU GMP Annex 11 or 21 CFR Part 11. Finding the answer and recording it are different jobs, and the record stays in the quality system your auditor already knows.
Also true ofNo general AI tool is validated to those standards either. The difference is scope: nothing here writes back into your quality system, so it cannot alter a record.

Try it on your own documents this week

14 days, no credit card. Upload the files people ask you about, and ask your first question.

Start free trial

SureCiteAI finds and cites information in the documents you upload. It does not classify substances, determine regulatory status or author safety data sheets, and it does not replace your competent person, your regulatory advisers or your quality system.