Baseline Checklist

Minimal risk ≠ zero rules

Even if you only use AI to draft emails or summarise meetings, there are rules to follow and safeguards worth putting in place. Start with these six steps.

General examples apply to any organisation. If your organisation is in scope for Finance & DORA or NIS2, look for the labelled sector notes alongside the relevant steps.

0 / 6 in place
  1. 1

    Keep an internal AI inventory

    Good practice

    Keep a simple list of the AI tools people use, including their vendor, purpose, owner and assessed risk level.

    In practice

    One spreadsheet row per tool — e.g. “Copilot in Outlook · Microsoft · email drafting · all staff · Minimal”.

    General example

    The IT team adds Copilot to its software inventory and records that staff use it to draft emails, not to make decisions about people.

    EU AI Act reference

    • Art. 6 & Annex III — These help you classify how a system is used. The AI Act does not require a general internal register of all minimal-risk tools.

    Finance & DORA

    If Copilot is provided under a Microsoft 365 agreement, record the relevant ICT service arrangement in your DORA register of information. That register covers contractual arrangements, not individual AI tools.

    Art. 28(3)

    NIS2

    Include AI tools in your existing asset management where relevant. For example, record the owner and access rights for an AI meeting assistant. NIS2 does not prescribe a separate AI register.

    Art. 21(2)(i)
  2. 2

    Train staff in AI literacy

    Legal duty

    Everyone with a licence understands the basics: how generative AI works, that it can hallucinate, and its limits.

    In practice

    A short onboarding session or one-page guide before Copilot is switched on for a user.

    General example

    Staff compare a Copilot summary with the original meeting notes and spot that it omitted an agreed deadline.

    EU AI Act reference

    • Art. 4 — Providers and deployers must take measures to ensure a sufficient level of AI literacy for staff and others operating AI on their behalf, taking their experience, the context and the people affected into account.
    • Art. 3(56) — Defines AI literacy: the skills and understanding to use AI in an informed way and be aware of its risks.

    Finance & DORA

    Include relevant AI risks in ICT security awareness and resilience training. For instance, have staff practise spotting sensitive data in a prompt and brief the management body on risks from AI-enabled ICT services.

    Arts. 5(4) & 13(6)

    NIS2

    Cover relevant AI-related threats in cybersecurity training, such as phishing messages drafted with AI. Management bodies must receive cybersecurity training under the national rules implementing NIS2.

    Arts. 20(2) & 21(2)(g)
  3. 3

    A human owns every output

    Good practice

    AI drafts, people decide. Nothing goes to a customer or into a decision without someone reading it first.

    In practice

    House rule: “Read and approve before you hit Send.”

    General example

    A customer support adviser checks a Copilot-drafted reply against the published returns policy before sending it.

    EU AI Act reference

    • Art. 14 — Human oversight is required for high-risk AI. For everyday drafting, reviewing outputs is a safeguard, not what determines the risk level.
    • Recital 165 — Encourages applying high-risk safeguards voluntarily to other AI systems through codes of conduct.
  4. 4

    Set clear data boundaries

    Supports GDPR compliance

    Staff know what must never be pasted into a prompt or summarised by AI.

    In practice

    Set rules by tool and data type. For example, never paste passwords or API keys; only process personal or confidential data in approved tools, for approved purposes.

    General example

    Staff can summarise a public product guide but must not paste an employee’s medical certificate into Copilot.

    EU AI Act reference

    • Art. 2(7) — The AI Act does not replace GDPR — data protection law applies in full to data processed by AI.
    • Art. 10 — Data governance rules for high-risk AI set the benchmark for handling data responsibly.
  5. 5

    Use approved AI tools and work accounts

    Good practice

    Approve the tool, account type and permitted uses before allowing staff to use it for work.

    In practice

    Check the vendor’s terms for data processing, retention and model training. Don’t put work content into an unapproved personal account.

    General example

    An organisation approves Microsoft 365 Copilot through staff work accounts for email drafting, but not personal chatbot accounts for the same task.

    EU AI Act reference

    • Art. 25 — Sets out responsibilities along the AI value chain — deployers should know what their provider guarantees.
    • Art. 53 — General-purpose AI providers (behind tools like Copilot) must give downstream users documentation on capabilities and limits.

    Finance & DORA

    Use your ICT third-party risk process before contracting for an AI-enabled service. Check what it supports, assess its risks and carry out due diligence on the provider.

    Art. 28(4)

    NIS2

    Use your supply-chain security process to assess direct AI service providers. For example, check how a vendor protects access to your data and handles vulnerabilities and incidents.

    Art. 21(2)(d) & (3)
  6. 6

    Write a one-page AI policy

    Good practice

    Short internal rules: approved tools, who approves new ones, and how to report mistakes.

    In practice

    One page on the intranet, owned by a named person, reviewed yearly.

    General example

    A small employer names its IT lead as approver and asks staff to report accidental disclosures promptly through the existing incident channel.

    EU AI Act reference

    • Art. 50 — Explain when transparency duties apply, for example when people interact directly with an AI chatbot. Not every AI-assisted email needs a disclosure.
    • Art. 95 — Encourages voluntary codes of conduct for non-high-risk AI — your policy is your internal version.

    Ready-to-use template

    [Company Name] – Workplace AI Policy

    Version: 1.0 | Effective: [Date] | Owner: [CISO / Head of IT / Compliance]

    Applies to: All employees, consultants and temporary staff

    1. Purpose & scope

    This policy defines the acceptable use of everyday AI assistants (Microsoft 365 Copilot in Outlook, Teams, Word and Excel). We do not develop our own AI models, nor do we use automated decision-making that affects customers or employees.

    2. Approved tools only (no "shadow AI")

    • Permitted: company-provided enterprise tools with commercial data protection (your corporate Microsoft 365 Copilot account).

    • Prohibited: personal, free or unvetted AI tools (e.g. free ChatGPT, personal Claude/Gemini accounts, browser plugins) for any work-related task. Free consumer tools may use your data to train public models.

    3. What can and cannot go into AI

    • Allowed: drafting internal emails and documents, summarising internal meetings, analysing internal spreadsheets, proofreading.

    • Never: passwords, keys or IT infrastructure details; special category personal data (GDPR Art. 9); non-public financial results, M&A information or sensitive customer contract terms.

    4. Human accountability

    • You own the output: you are fully accountable for anything you share, even if AI drafted it.

    • Always verify facts, names, figures and sources – AI can "hallucinate".

    • No fully automated customer actions: a human reviews and sends.

    5. Transparency (EU AI Act Art. 50)

    • Disclose AI use when a client, partner or authority would reasonably expect the work to be your own.

    • Never present unedited AI output as professional advice.

    6. Incidents, new tools & review

    • Report sensitive data entered into an unapproved AI tool to [security@company.com] within 24 hours.

    • New AI tools or extensions must be approved by [Named role / IT].

    • This policy is reviewed annually and supports AI literacy under EU AI Act Art. 4.

Legal duty marks the AI Act’s literacy requirement. Supports GDPR compliance marks a practical safeguard when personal data is involved. Good practice is not a specific AI Act duty for minimal-risk use.

DORA and NIS2 have their own scope and do not change an AI system’s risk classification. Check the national rules implementing NIS2. For financial entities covered by DORA, DORA takes precedence for equivalent cybersecurity risk-management and incident-reporting requirements under NIS2 Article 4.

Not sure your tool is minimal risk?

Answer ten quick questions to check its likely risk level.

Take the self-assessment