How we use AI at Liquid Light

We've spent more than 30 years helping ambitious organisations tell their stories and solve complex challenges. Throughout that time, one thing has stayed constant: the thinking behind our work comes from humans. Your brand, your website, your next campaign: these are shaped by a team who care deeply about getting them right.

AI has quickly become part of the modern digital toolkit, and large language models are changing how agencies work. In our experience, these tools can be genuinely useful when applied thoughtfully. They also raise fair questions about quality, sustainability, originality, and privacy, and we try to be honest about the tensions that come with them.

This policy explains why we use AI, where it fits into our work, the tools we rely on, and where we choose to be cautious.

Why do we use it?

The answer is time, and for a useful second opinion. Take the discovery phase of a project: a set of workshop notes and several rounds of stakeholder feedback hold real insight, but drawing that into a clear, usable summary can take time. AI helps us reach a first structured version quickly, freeing our team to interpret what the findings mean for you rather than wrangling the raw material. We also use it to sense-check our work, testing an approach or spotting gaps in our thinking before we take it further. That reclaimed time and added rigour go straight back into your project, giving us more room to think, test, and craft the work that matters most.

We're realistic about what AI can and can't do. It's a capable assistant for structure and speed, though it has no judgement of its own: it recognises patterns rather than understanding meaning. We tend to prompt it to ask us questions rather than hand us answers, and to lay out considerations rather than recommend conclusions. Everything it helps produce is reviewed and reworked by our team before it reaches our clients.

Where does AI fit into our work?

We use AI to take care of the groundwork: the time-consuming, behind-the-scenes tasks that keep a project moving. Handling these more efficiently frees our team to focus on the creative and strategic work that makes the real difference to your project.

In practice, that might look like:

• Pulling together meeting transcripts and workshop outputs into clear summaries and actions.
• Early-stage research, such as sector or competitor analysis.
• Data analysis, such as surfacing patterns and themes across survey responses and interview transcripts.
• Drafting the framework for reports, which our team then writes and refines in our own voice.
• Creative exploration, such as sparking early ideas and testing visual directions, which our designers then take forward and shape by hand.
• Rapid prototyping: turning initial concepts and sample data into wireframes and layouts we can react to and refine quickly.
• Code reviews, along with support for troubleshooting and debugging.
• Setting up internal task lists and project admin, always checked by a person before anything is shared.

Our commitment to you: a finished deliverable will always be led, shaped, and signed off by a human. We've built a talented team over three decades, and their skill is exactly why organisations choose us.

Which tools do we use?

Our team draws on a small, carefully approved set of tools, each chosen for specific jobs:

• Google Gemini, our primary language model, is used within our secure Google Workspace so it sits naturally within our workflow while meeting our privacy and security standards.
• Claude, for reviewing logic and summarising complex project material, with model-training settings switched off.
• GitLab Duo, connected to our self-hosted GitLab, to support our developers.
• IntelliPHP, which runs on an offline model, to assist with routine code.

We keep this list short and deliberately chosen. Tools that don't meet our standards for privacy, security, or quality don't make the cut, and our approved list is reviewed as the technology evolves. We only use AI where a project's information stays private and isn't used to train anyone's models.

Protecting your data sits behind these choices. Where we can, we anonymise before we begin, substituting identifying details with generic terms so a project can't be traced back to you. Confidential project details, credentials, and anything genuinely sensitive stay out of AI tools altogether. Whether AI is involved or not, our team treats your information with the same care, because your trust underpins everything we do.

Where we are cautious

We don't use AI-generated imagery in finished client work without an explicit conversation with you first. The legal and ethical position on training data is still being tested, and you deserve the final say on whether AI-generated visuals appear under your name.

We don't use AI to write articles, case studies, or opinion pieces on our behalf. We may use it to think through structure or challenge an argument, but the ideas and the voice are ours. There's already enough generic AI slop online, we don't want to add to it.

The environmental question

Many of our clients work in environmental and sustainability spaces, and we're conscious of the energy cost of generative AI. We use it deliberately rather than by default. We've written before about the carbon footprint of AI and how to mitigate it, and we apply the same thinking to our own work.

What this all means in practice

• We use AI as part of our process, with people leading and finishing the work.
• Someone on our team reviews everything before it reaches you or goes out into the world.
• We're open within the team about where AI has contributed, so it can be checked before it goes further.
• We talk to you before any AI-generated visuals appear in final deliverables.
• We review this position regularly, because the tools, the law, and our thinking all keep changing.

The principle behind it all

We believe AI works best when it's used intelligently: in the right place, for the right task, always under human direction. It streamlines our work and sharpens our thinking, while the creativity, strategy, and craft remain firmly with our team.

Above all, your final work belongs to people. Your brand, your website, your designs, and your campaigns are made by us, for you: that's where the value lives, and it's what you can expect from us on every project.

We'll always be open about how AI has supported your work. If you have questions about how we've used AI on a project, just ask.