Free EU AI Act customer questionnaire answer sheet: ten questions, ready-to-edit answers, evidence to attach

Updated 27 Oct 2026 · Free. No sign-up. Disclosure: Unadhi also sells a paid EU AI Act Readiness Kit (US$29), linked near the end. The sheet works without it.

What you get: one Google Sheet tab with ten rows, one for each AI Act question EU customers commonly put in a supplier questionnaire. Each row has the question, an answer you can edit and paste, the evidence to attach and the official article it comes from. Fill the square brackets with facts that are true for you and delete any sentence that is not. Download the CSV, then in Google Sheets use File → Import → Upload → Replace spreadsheet.

The finished sheet (answer tab)

#	Question the customer asks	Answer template	Evidence to attach	Where it comes from
1	Do you provide or use AI systems in the service we buy?	We [use / provide / use and provide] AI systems in [service name]. The AI systems are listed in our AI inventory (last updated [date]). Systems used: [tool 1, tool 2].	Your AI inventory (one row per AI system)	Art. 3 definitions (official text)
2	Are you the provider or the deployer of each system?	For [system name] we are the [provider: we develop it, or have it developed and place it on the market under our own name / deployer: we use a system built by [vendor] under our authority].	Role column in the AI inventory	Art. 3(3) 'provider', Art. 3(4) 'deployer' (official text)
3	Does any use fall under the high-risk list (Annex III)?	We checked each use against Annex III on [date]. Result: [none apply / [use] may apply, reviewed by [name or adviser]]. For Annex III uses the high-risk rules apply from 2 December 2027.	Dated screening note per system, with your reasoning	Art. 6, Art. 113(c)(i) (official text)
4	Do you use any prohibited AI practice?	We have checked the prohibited practices in Article 5 and do not use [any / the following: ...] on [date].	Dated Article 5 check signed off by [role]	Art. 5; Art. 113(a) (applies from 2 Feb 2025; some newer points from 2 Dec 2026) (official text)
5	Do people know when they are talking to an AI system (chatbot, voice bot)?	Yes. [Chatbot name] shows this notice at the first message: "You are chatting with an AI assistant, not a person. Ask for a human at any time." [Delete the last sentence if you do not offer a human.] Where it appears: [first message / voice prompt / footer].	Screenshot of the notice with date; the exact wording	Art. 50(1) and Art. 50(5) (official text)
6	Is AI-generated audio, image, video or text marked as AI-generated?	[Provider answer] Our system marks its outputs in a machine-readable format: [method]. Systems placed on the market before 2 Aug 2026 have until 2 Dec 2026. [Deployer answer] We use [vendor tool]; the vendor confirmed on [date] that outputs are marked: [link or email reference].	Vendor confirmation or your technical note	Art. 50(2); Art. 111(4) (official text)
7	Do you publish deepfakes or AI-written text on matters of public interest?	[No. / Yes: we label [content type] as "AI-generated" before publishing, and [text only] a named person reviews and is editorially responsible.]	Label wording; name of the responsible editor	Art. 50(4) (official text)
8	Is your staff trained on AI?	We take measures to support AI literacy of staff and others who operate AI systems for us, based on their knowledge and the context of use: [what you did, date, who attended].	Training or briefing record with dates	Art. 4(1) (official text)
9	Which AI models or vendors sit behind your product, and what documentation do you have for them?	We use [model/vendor] under [contract/date]. We hold the vendor's documentation: [link, version, date received]. [If the vendor gave none, write that and say what you asked for.]	Vendor documentation, contract clause, date requested	Chapter V (general-purpose AI models) applies from 2 Aug 2025: Art. 113(b) (official text)
10	Who is responsible for AI in your company, and who do we contact?	[Role, not a personal phone number] is responsible for AI use and answers AI questions at [company email].	One named role and one shared email	Internal record

The row-5 notice wording is a suggestion. The Act does not prescribe exact words; it asks that people are informed, clearly and at the latest at the first interaction (Article 50(1) and (5)).

How to fill it in, in 20 minutes

List every AI tool your team uses or builds into your product (one line each: name, vendor, what it is used for, who can see its output).
For each tool decide your role (row 2) and write it next to the tool.
Answer rows 3 and 4 only after you have looked at each use against the official text of Article 5 and Annex III. If you cannot tell, write "under review" and the date, not "no".
Paste your chatbot notice and take a dated screenshot (row 5).
Ask each vendor in writing how outputs are marked and what documentation they provide (rows 6 and 9). Keep the reply.
Write what staff actually did on AI literacy (row 8). Do not describe training that has not happened.
Save a copy of the finished sheet with the date. Update it when you add a tool or when a date below changes.

The dates the sheet relies on

Date	What applies	Source
2 Feb 2025	Chapters I and II (including AI literacy in Article 4 and the prohibited practices in Article 5), except Article 5(1) points (ba) and (bb) and Article 5(1a), (1b), which apply from 2 Dec 2026	Art. 113(a)
2 Aug 2025	Chapter V (general-purpose AI models), Chapter III Section 4, Chapters VII and XII and Article 78 (not Article 101)	Art. 113(b)
2 Aug 2026	The Act applies in general, including Article 50 transparency duties	Art. 113
2 Dec 2026	Article 50(2) marking, for generative AI systems placed on the market before 2 Aug 2026	Art. 111(4)
2 Dec 2027	High-risk rules for AI systems classified under Article 6(2) and Annex III	Art. 113(c)(i)
2 Aug 2028	High-risk rules for AI systems classified under Article 6(1) and Annex I	Art. 113(c)(ii)

Want the AI inventory and policy behind the answers?

The answers above point to an AI inventory, a dated risk screening and an AI literacy record. If you do not have those yet, the EU AI Act Readiness Kit (US$29, once) is an Excel AI inventory with risk screening, an AI literacy and acceptable-use policy template in Word and a 5-page guide with a 90-day plan. 14-day money-back if it is not useful. It does not classify your systems for you and is not legal advice. You can also build the inventory yourself in any spreadsheet and use the sheet above as it is.

Frequently asked questions

Is this legal advice?

No. It is a drafting aid. The Commission's AI Act Service Desk and a qualified lawyer are the places to confirm how the Act applies to you.

Why is a notice enough for a chatbot?

Article 50(1) requires that people are informed they are interacting with an AI system, unless that is obvious to a reasonably well-informed person given the context. Article 50(5) requires the information to be clear, distinguishable and given at the latest at the first interaction.

Does my small business need to mark AI-generated content?

Article 50(2) puts the marking duty on providers of AI systems that generate synthetic audio, image, video or text. If you only use a vendor's tool, ask the vendor how its outputs are marked and keep the answer.

This sheet is general information and a drafting aid, not legal advice. Answers must be true for your business; nothing here is proof that you comply with the law. Unadhi is not affiliated with the EU or any EU authority. Spotted an error? Write to mininglife7@gmail.com and we correct it.

Sources

AI Act Service Desk, European Commission, consolidated text as at 27 July 2026, including changes by the Digital Omnibus: Article 50, Article 3, Article 4, Article 6, Article 111, Article 113 (all read 27 October 2026).
Regulation (EU) 2026/1744 (Digital Omnibus on AI): EUR-Lex.
Regulation (EU) 2024/1689 (AI Act): EUR-Lex.

← All guides

Unadhi Guides are free, independent explainers. We correct errors as soon as we find them: write to mininglife7@gmail.com.