Find out what people actually do, not just what procurement has formally approved. Include casual experimentation, personal accounts used for work and AI features embedded in everyday software.
Purpose
Create a useful first view of AI use across the organisation, with named contacts and enough context to decide what needs follow-up.
Recommended timing
Weeks 2–3 · Indicative timing. Adjust to the organisation’s size, complexity and available capacity.
What to do
- Send a short, plain-English AI usage check to all employees. Make clear that one-off and occasional use counts.
- Include examples: drafting or rewriting, summarising, translation, meeting notes, research, image creation, coding, data analysis, chatbots and AI features built into office or business software.
- Ask departments to nominate a contact who can help clarify responses. Do not ask employees to classify systems under legislation.
- Set a clear response deadline and send one reminder. Offer a simple route for employees who are unsure whether a tool uses AI.
- Collect tool names, purposes, users, frequency, whether work information is entered, whether outputs influence decisions and who can explain the use.
- Add identified uses to the central inventory in your organisation’s own approved environment. Record uncertainty rather than guessing.
A simple employee AI usage check
Keep this to five minutes. Send it to everyone, including people who do not think of themselves as AI users. A one-off experiment, an AI button inside familiar software or a personal account used for work all count. The aim is to understand what is happening, not to catch people out.
Suggested questions
- In the last three months, have you used an AI tool or an AI-powered feature for work? Yes / No / Not sure. Occasional or experimental use counts.
- What tool or feature was it? Name it if you know. Examples: ChatGPT, Microsoft Copilot, Gemini, an AI meeting note-taker, an image tool or an AI feature inside another business application.
- What did you use it to do? A short description is enough—for example, drafting, summarising, translating, research, coding, analysis, customer support or making a decision.
- How often do you use it? Once or twice / Monthly / Weekly / Daily / Built into a process or application.
- What kind of information did you enter or make available? Public information only / Internal business information / Personal information / Confidential or sensitive information / Not sure. Do not include the information itself in your response.
- Is the AI output used to make or materially influence a decision about a person, customer, employee or access to a service? Yes / No / Not sure. If yes, briefly describe the decision.
- Who uses or relies on this tool? Just me / My team / Several teams / Customers or members of the public / Not sure.
- Who can help us understand this use? Your name and team, or the name of the person who owns the process. If you are unsure, say so.
Optional: Is there anything else we should know, or any AI use you are unsure how to describe?
How to introduce it: “We are building a picture of how AI is actually used across our organisation so that we can support people, manage risks sensibly and make informed decisions. This includes occasional use and AI features built into everyday tools. Please answer openly; this is not a performance assessment. Do not paste personal, confidential or sensitive information into this questionnaire.”
For the programme lead: distribute the questionnaire using an internal channel you control, collect responses in your organisation’s approved environment, restrict access appropriately and maintain the resulting inventory in your own systems. Do not send completed questionnaires, inventory records or sensitive operational information to Lucentreg. Follow up on unclear answers. Do not ask employees to decide whether something is high-risk or legally compliant. Record uncertainty and escalate it for appropriate review.
Meetings to arrange
Use these default meeting formats. Combine meetings or shorten them where the same people are involved and the decisions can be made together.
Discovery launch
Duration: 45 minutes
Attendees: Programme lead, department contacts, internal communications/HR and IT.
Agenda: Explain what counts as AI use, why informal use matters, how responses will be handled and where employees can ask questions.
Expected output: Questionnaire issued, contacts briefed and response deadline agreed.
Department follow-up
Duration: 30 minutes per department, as needed
Attendees: Department contact, programme lead and relevant system or process owner.
Agenda: Clarify unclear responses, identify embedded AI features and distinguish personal experimentation from business use.
Expected output: Useful inventory entries with owners, purpose and missing information recorded.
Decisions to make
- Which departments need targeted follow-up?
- Which uses involve sensitive, confidential or personal information?
- Which uses influence decisions about people, access to services, employment, credit or other significant outcomes?
- Which suppliers or software products need further investigation?
Definition of done
- Initial AI inventory with named owners where known.
- List of departments that have not responded.
- List of unclear or potentially significant uses for triage.
- Supplier and embedded-AI follow-up list.
Keep it proportionate
Use existing roles, committees and processes wherever they work. Do not create a new approval layer or collect more information than is needed to make a sound decision.
Resources to use
Use a blank AI usage questionnaire and discovery guidance supplied with the customer resources. Distribute and collect it through your own approved internal channels. Maintain the completed responses and resulting AI inventory in your organisation’s environment; Lucentreg does not need to receive them. Use the governance workbook as your internal record and share only the minimum necessary information with authorised colleagues.
Important: Lucentreg provides educational and operational guidance, not legal advice or a determination of compliance. Screening prompts and suggested actions are starting points. Your organisation remains responsible for its decisions and should obtain qualified legal or technical advice where appropriate.