Policy on the use of artificial intelligence

At LEAD we guide leaders and teams in navigating complexity. Part of that work -writing, structuring, summarising, formatting -is today supported by artificial intelligence tools.

We neither hide this use nor present it as a selling point. We treat it like any other professional practice: with clear rules, carefully chosen tools, and one requirement that nothing can replace -human judgement remains visible and is fully owned.

This policy sets out the rules we apply.

This policy applies to all LEAD employees and to any freelancer or subcontractor working on LEAD's behalf. They undertake to observe it whenever they use AI tools in the context of an assignment.

Today we use two AI tools in our work:

• Claude, developed by Anthropic, on the Team plan, for writing, structuring and analysing content.
• Plaud, on the Pro plan, for transcribing and summarising meetings -only where all participants have been informed and have given their consent.

If we adopt other tools in the future, they must meet the same criteria (section 5).

We chose these tools against strict data-protection criteria.

We conclude a data processing agreement (DPA) with each supplier within the meaning of Article 28 GDPR. We work exclusively with paid, professional accounts -Claude Team at Anthropic, Plaud Pro -which provide these contractual guarantees. In that relationship our suppliers act as processors, and Lead remains the data controller.

Part of the processing carried out by these tools takes place on servers outside the European Economic Area, in particular in the United States. That transfer is covered by the European Commission's Standard Contractual Clauses, incorporated in our suppliers' data processing agreements, and supplemented by appropriate technical and organisational measures.

Under the contractual terms of the professional accounts we use, the content we submit to these tools is not used to train the AI models. Where a tool excludes training only subject to an explicit setting, we make sure our accounts are configured correctly.

LEAD does not pass your data to third parties for commercial, marketing or analytics purposes. Our suppliers act solely as processors, on our instructions.

We keep the data we submit to these tools no longer than necessary for the assignment. Meeting recordings and transcripts are deleted after 120 days. On request, we delete data relating to you.

Before we integrate a new AI tool, we check four requirements:

1. A GDPR-compliant data processing agreement -the supplier can produce the corresponding contractual guarantees.
2. Transparency about the storage location, and, for transfers outside the EEA, valid transfer mechanisms (SCCs).
3. No use of our data for model training.
4. No transfer to third parties, in any form.

A tool that fails to meet any one of these criteria, we do not use.

Just as important as the choice of tool is what we put into it. Our consultants do not enter special categories of personal data, nor confidential client data that is not covered by a data processing agreement, unless this has been expressly agreed in writing with the client. When in doubt, we work without AI support.

This is the most important rule.

AI helps us format, structure and accelerate. It does not decide in our place.

In practice:

• Every piece of content produced with AI support is re-read, evaluated and weighed by a Lead consultant before it is delivered to a client or published.
• The recommendations, diagnoses and strategic options in our assignments rest on the judgement of our consultants -not on a model output.
• We name what we reject, what we adjust, and why. Our added value lies in those judgements, not in the act of generating.

We regard this human control as non-negotiable.

Our conviction -reinforced by recent research on trust and AI -is that useful transparency does not consist in naming a tool, but in making the reasoning visible.

In our deliverables we aim to make explicit as much as we can:

• the trade-offs we have made,
• the options we rejected, and why,
• the limits of what we affirm.

That work -and not the mention of a tool -is what builds trust.

This policy is owned by the Managing Partner of LEAD. If we identify an error or a data breach involving an AI tool, we act in line with our internal incident procedure and, where required, inform the data subjects and the competent authority within the statutory deadlines.

We monitor developments in the European regulatory framework, including the AI Act (EU AI Act), and align our practice accordingly. Our emphasis on visible human judgement is consistent with the transparency intent of that framework..

Do you have a question about how we use AI in the context of an assignment? Would you prefer your file to be handled without AI support? Contact us at info@lead.be. We will process your request and confirm the measures we take.

AI tools evolve quickly, and so do suppliers' contractual terms. We review this policy at least once a year, and on any significant change to our toolset or our suppliers. The version in force is always the one published at www.lead.be/ai-policy.