Do you use AI in client work?

Yes.

FortyTwo may use approved AI tools to support aspects of strategy, creative development, research, technical delivery, marketing operations, project management and internal workflow efficiency.

AI is used to assist the team, not replace professional expertise, judgement or accountability.

Examples of appropriate AI-assisted activities may include:

• research support
• ideation and exploration
• content structuring
• summarisation
• draft generation
• code assistance
• technical troubleshooting
• SEO support
• accessibility review
• project administration
• internal workflow support

All final outputs remain subject to human review, editing and approval.

Will AI generate final deliverables without human review?

No.

No AI-generated output will be:

• delivered directly to a client
• published to a live environment
• treated as factual
• approved as final creative work
• used in a campaign
• implemented in production systems

without appropriate human review by the FortyTwo team.

Responsibility for final outputs always remains with us.

What AI tools does FortyTwo use?

FortyTwo only permits the use of approved AI platforms for client-related work.

Approved platforms are reviewed for:

• security
• confidentiality
• commercial terms
• data handling practices
• intellectual property considerations
• privacy compliance
• training and retention policies

Where possible, FortyTwo uses paid, business or enterprise-level AI services that provide contractual protections regarding client data and model training.

We are happy to provide details of the tools we have approved and are using for any client project. 

Is client information used to train public AI models?

FortyTwo’s position is that confidential client information should not be used to train public AI models.

Where enterprise or business-tier AI platforms are used, FortyTwo seeks to use services whose terms state that customer prompts, uploads and outputs are not used to train publicly available models.

However, as we are beholden to the policies of the tools we use, if their terms change, we will review our position and discuss with clients before any changes are approved. 

Any exception to this approach would require explicit internal review and approval.

What types of information are not permitted to be entered into AI tools?

Unless specifically authorised and reviewed, the following types of information must not be entered into third-party AI systems:

• passwords or login credentials
• confidential financial information
• personal data
• HR records
• sensitive contractual information
• trade secrets
• unpublished intellectual property
• proprietary source code
• confidential client documents
• commercially sensitive information
• regulated or protected data
• security-sensitive technical information

Additional restrictions may apply depending on the project, client sector or contractual requirements.

How does FortyTwo classify information when using AI tools?

FortyTwo recognises that different categories of information carry different levels of risk.

As a general principle, information may be treated as:

• public
• internal
• confidential
• commercially sensitive
• regulated or protected

The sensitivity of the information determines whether AI tools may be used and which tools, if any, are appropriate.

Can clients request limitations on AI use?

Yes.

Clients may:

• request that specific AI tools are not used
• request additional approval steps
• specify project-specific restrictions
• request disclosure regarding material AI-assisted workflows
• request that AI is not used for specific deliverables or workstreams

Where practical and commercially reasonable, FortyTwo will work collaboratively with clients to respect these preferences.

If you would like to discuss use of AI on your projects, your designated account manager or key contact will be happy to help. 

Will clients be informed when AI has been used?

Not every minor operational use of AI will necessarily be disclosed. For example, internal formatting assistance, note summarisation or administrative support tools may not require explicit notification.

However, where AI materially contributes to:

• generated imagery
• written content
• strategic analysis
• code production
• voice or video generation
• client-facing deliverables

FortyTwo may disclose this where appropriate.

How do we treat AI material provided by clients?

Increasingly, clients are discovering how useful AI can be in helping to articulate a brief, generate early visual references, or collate thoughts and feedback. Used well, these materials can provide a valuable starting point for discussion and help accelerate early-stage communication.

However, we are careful to treat AI-generated visuals as moodboards rather than finished concepts. They can help us understand intent, but they are never a substitute for the collaborative, human-led creative process required to solve the problem properly. The same principle applies to copywriting.

Any AI-generated copy we receive is treated as a draft and subjected to the same rigorous review, refinement and strategic thinking as any other written material. After investing time in defining a clear tone of voice and message, it would be unwise to hand interpretation entirely to AI without consistent human oversight

 Likewise, while AI can be helpful in organising and summarising feedback, it must be accompanied by direct conversations to ensure expectations, priorities and interpretations are fully aligned. In every case, AI serves as a useful assistant to the process, not a replacement for it.

Who owns AI-assisted deliverables?

Current updates to copyright law have been paused, so there is no specific AI regulation in this regard. Therefore, ownership, licensing and transfer of deliverables remain governed by FortyTwo’s contractual agreements with clients.

The use of AI-assisted workflows does not alter agreed ownership arrangements unless explicitly stated otherwise in writing.

How are AI vendors assessed?

AI platforms may be reviewed against criteria including:

• access management
• training policies
• retention practices
• commercial licensing
• copyright position
• enterprise safeguards
• regulatory compliance

Frequency of vendor review:
Annually - Jan/Feb

Responsible reviewer/team:
Natasha Wilson - Senior Account Manager

Are prompts, uploads or outputs retained by AI providers?

Retention practices vary between platforms and providers.

FortyTwo seeks to use platforms with commercially appropriate retention, confidentiality and enterprise protections where possible.

Where retention or training policies create elevated risk, those platforms may be restricted or prohibited for client-related work.

We’ll be held to the standard of any AI laws, policies and requirements from AI tools themselves. As we do not own the AI tools, we cannot dictate retention terms, but we will consider these before approving a tool for use. 

What happens if there is an AI-related incident or concern?

Any suspected:

• misuse of AI systems
• confidentiality breach
• inappropriate disclosure
• security concern
• unauthorised AI usage
• inaccurate or misleading AI-generated content

must be escalated internally for review.

FortyTwo will assess incidents in line with its internal governance, contractual obligations and applicable legal requirements.

Incident escalation contact/process:
Gina Urquhart - Digital Account Director

Target response timeframe:
Initial response and receipt of issue - Within 2 hrs
Plan of Action and Execution of Plan - 2 - 5 days (depending on scale)

How often is this policy reviewed?

Because AI technologies, regulations and commercial practices continue to evolve, this policy will be reviewed periodically.

Review Frequency: Annually Jan/Feb

Policy owner: Mark Kemp - Managing Director

Where can clients find additional information?

Additional information may be available in:

• FortyTwo’s AI Charter
• FortyTwo’s Privacy Policy
• Relevant Project Agreements
• Confidentiality Agreements
• Client-Specific Contractual Documentation

For further questions regarding FortyTwo’s use of AI in client work, please contact Mark Kemp: mark@fortytwo.studio