AI work policy

Preamble

I work with AI every day and I think it should be talked about openly. AI takes over the part of the work that needs time but not judgement. A shareholder report, for example, used to take me 3 hours. Now it takes 30 minutes. A fully structured project plan used to take up to 8 hours, now 1-2 hours. I put the time this frees up where my experience adds the most: complex questions and decisions that call for experience, sound arguments and strategic thinking.

This policy sets out how I work with AI, where the limits are and how I handle the information entrusted to me. I follow it on every Consilium Digitale project. It rests on five principles:

1. I

   Human judgement

   AI helps me work, but I draw the conclusions and I am accountable for them.

2. II

   Confidentiality

   I protect the information entrusted to me and keep each partner's separate.

3. III

   Transparency

   I can always say where and how I used AI.

4. IV

   Partner independence

   What I build for a partner's project stays with the partner.

5. V

   Competence

   Every month I set time aside to learn and I test new tools on real tasks.

Definitions

Partner.

A company or organisation I work with as an advisor, consultant or board member.

AI tool.

Software that analyses data, prepares documents and runs repetitive tasks automatically. I currently use these features:

• Projects - separate workspaces, each with its own files, instructions and memory.
• Skills - instruction sets that teach the AI a specific task, such as preparing a report in a set format.
• Connectors - access to my email, calendar and documents and, with a partner's permission, to the partner's systems.
• Scheduled tasks - work the AI runs automatically at a set time, such as a weekly report.
• Analysis and documents - calculations, data analysis, research in public sources, reports, decks and spreadsheets.
• Programming - the AI writes code to my brief: scripts for processing data, webpages, apps, work tools and automations.

Partner information.

Data and documents I receive from a partner or create on its behalf. I sort it into three levels:

• Public - published information.
• Internal business - sales, finance, plans, board materials.
• Strictly confidential - deals, negotiations and anything the partner marks as highly sensitive.

Working provisions

1. Human judgement

1. 1.1

   I am accountable for every conclusion and recommendation. An AI mistake I did not catch is my mistake.

2. 1.2

   I do not make a recommendation I have not checked myself and cannot argue for. I recalculate figures or check them against the source.

3. 1.3

   I use AI to analyse data, prepare documents and decks, research public sources, automate repetitive work, build work tools and test my own arguments.

4. 1.4

   I do not use AI to assess people: candidates, executives or team members. I make those calls myself.

5. 1.5

   I do not use AI in negotiations or sensitive conversations. I do not record or transcribe meetings without the consent of everyone present.

2. Confidentiality

1. 2.1

   Each partner has a separate AI Project with its own memory. One partner's material is not visible in work for another.

2. 2.2

   I use a paid AI subscription with the use of my conversations for model training switched off. I do not use free or unvetted tools for partner information.

3. 2.3

   I use internal business information only in the tool named in 2.2 and only where our contract allows it.

4. 2.4

   I do not put strictly confidential information into AI tools unless the partner and I agree to it separately in writing.

5. 2.5

   I do not put personal data into AI tools unless the task needs it. Where possible I anonymise it.

6. 2.6

   I connect AI to a partner's systems only with the partner's permission.

7. 2.7

   If a partner has its own AI or IT security rules, I follow them, even when they are stricter than this policy.

8. 2.8

   I start using a new tool for partner information only after updating this policy.

3. Transparency

1. 3.1

   If a partner asks which part of the work used AI, I give a specific answer.

2. 3.2

   If I suspect a partner's information has gone somewhere it should not, I report it within 24 hours and explain what I am doing about it.

3. 3.3

   I publish this policy and record every change in its change log.

4. Partner independence

1. 4.1

   A skill or other AI tool I build for a specific partner project belongs to the partner as well. On request or when our work ends, I hand it over in a format the partner can use in its own AI systems.

2. 4.2

   My general methodology, such as strategy and sales planning frameworks and writing rules, stays with me.

3. 4.3

   Before handing anything over, I check that it contains no other partner's information.

5. Competence

1. 5.1

   I spend at least 16 hours a month on structured learning: courses on official certified learning platforms, events and hands-on projects.

2. 5.2

   I try every new thing on a real work task within the same week.

Final provisions

1. 6.1

   I review this policy every 6 months and after any significant change to the tools I work with.

2. 6.2

   If you believe I have not followed this policy, write to info@consiliumdigitale.lt. I will reply within 5 working days.

3. 6.3

   Change log: 1.0, 22 October 2026