AI Policy

Generative AI creates new content from scratch. Give it a prompt, and it produces a blog draft, image, or meta description based on patterns it has learned. Examples: ChatGPT, Gemini.

AI-assisted software doesn’t create content independently. It supports your work by suggesting keywords, highlighting technical issues, or refining what you’ve already written. Examples: Grammarly, Google Search Console.

The important distinction is that generative AI always needs human editing for accuracy and brand alignment. AI-assisted tools improve precision and strategy but rely on expert input. 

1. Purpose of this policy

This document sets out how we use Artificial Intelligence (AI) safely, responsibly, and transparently in our digital marketing work. It applies to both our team and our clients, and reflects best-practice guidance from the Parliamentary Office of Science & Technology (POST), UK GDPR, and wider ethical standards.

AI supports our creativity, analysis, and efficiency, but it never replaces human expertise. Every strategic decision, creative output, and client deliverable is checked, reviewed, and approved by a person.

2. How we use AI tools

AI tools help us to improve the speed, quality, and effectiveness of our work. 

We use AI to help us:

• Generate ideas and first drafts
• Identify patterns in data and performance
• Support research and analysis
• Automate repetitive processes
• Enhance content for SEO purposes 
• Improve internal efficiency so we can spend more time on strategy

AI is always a supporting tool, not a standalone decision-maker. 

Even though we use AI, it doesn’t erase the need for creativity. But it often sparks ideas that we can develop in our own unique way.

3. Things we don’t do when using AI tools

To protect quality, fairness, and trust, there are clear limits on what we use AI for.

We don’t use AI to:

• Speak directly to our clients, i.e., using AI chat bots
• Produce finished client content without human editing, adapting and approval
• Access, store, or process confidential client information (specifically within generative AI software)
• Generate claims or factual statements without verification
• Create misleading, discriminatory, or harmful outputs

Human oversight is always required.

4. Transparency and Accountability

We are open about how and when AI is used. Clients can expect:

• Internally, team members adhere to a checklist for AI use
• Assurance that all outputs are checked by a human
• Full accountability for the accuracy and quality of the work we deliver

5. Data Protection & Privacy

No sensitive or confidential data is entered into AI systems.

This includes:

While we may record meeting audio using AI-assisted tech, we always ask permission first and never share transcripts

• Personal data
• Client-specific confidential information
• Any information governed by NDAs
• While we may record meeting audio using AI-assisted tech, we always ask permission first and never share transcripts

All data handling complies with UK GDPR as set out in our Privacy Policy

We only use AI tools with transparent data handling policies.

We disable the ability to use data to train generative AI models.

6. Fairness, Bias, Inclusivity

AI systems can reflect or amplify bias. 

To prevent this:

• We check AI outputs for stereotypes, discriminatory phrasing, or unfair representations
• We ensure content is inclusive and respectful
• We prioritise balanced, representative imagery and language
• We will not manipulate AI tools to produce work that may mislead audiences, smear competitors, create fake personas for commercial gain, or perform tasks that are against best practices.
• This applies equally to internal work and client work

If any output is generated that contravenes any of the above, we will report it via the proper channels.