AI first, not AI always
We use AI every day, in every team, because it makes our specialists faster and the work better. A person stays responsible for everything it touches, and we run it all under a certified AI management system. Here's exactly how.
How we use it
Where AI earns its place
AI is the starting point for our work wherever it can credibly help. It speeds up the craft. It doesn't replace the judgement or the accountability.
-
Strategy and growth
Research synthesis, competitor scans, audience analysis and first drafts.
-
Design
Mood boards, visual concepts, prototype concepts, layout and colour variations, image editing and design QA.
-
Development
Scaffolding, refactoring, documenting and reviewing code, drafting tests and debugging.
-
Delivery
Plans, risk reviews and status updates.
Where we draw the line
Decisions about people, such as hiring or performance. A person makes those, start to finish.
Legal, regulatory, financial or medical content, without a qualified professional checking it.
Personal data, passwords and anything else our policy keeps out of AI tools.
Any work where your contract or brief rules AI out.
Experts first
Specialists, with AI behind them
AI gets the first draft there faster. Our people make it yours.
AI gives everyone the same average answer. Our specialists are why your work isn't average.
A model trained on the whole internet knows what most companies say. It doesn't know your customers, your sector or what makes your brand different from the one next to it.
So our strategists, designers, developers and growth specialists use AI to reach a strong first draft faster, then spend the time they've saved on the judgement that makes the work yours.
Every engagement is treated as unique. Nothing produced for one client is reused, adapted or referenced for another.
A person checks everything
Every piece of AI-assisted work is reviewed by a specialist before it reaches you, to make sure it's correct, safe, ethical and legal. Whoever signs it off is responsible for it, whether AI helped or not.
- Facts, figures and references are checked against their sources. We never cite anything AI produced without verifying it.
- AI-assisted code goes through our normal review, testing and approval. Nothing merges without a developer reviewing it.
- Outputs are checked for accuracy, bias and whether they suit your brand and your audience.
- AI-generated imagery follows strict rules. We never use AI-generated people as customers, staff, testimonials or case studies. We never alter a real person's likeness without their written consent. We keep provenance labels intact, and we list any AI-generated assets for you at handover.
Your data
-
Business-grade tools only
We only use paid business licences from established AI providers, accessed through company sign-in with multi-factor authentication. Free and personal AI accounts are banned for any work.
-
Never used for training
We don't train, fine-tune or contribute to training any AI model with your data, and every tool we use is on terms that exclude it.
-
Every provider assessed
Each AI provider is assessed as a data processor, with a data processing agreement in place, before anyone uses it. We review them at least once a year, and again whenever their terms or features change.
-
Checked before it goes in
Before using any information with AI, we check what kind it is. Public and internal material can be used. Confidential material, like contracts and plans, needs approval first. Personal data, passwords, payment details and sensitive personal data never go into an AI tool, and nor does anything your NDA keeps from third parties.
-
Your work stays yours
Using AI doesn't change who owns what we make for you. That's set by your contract, as it is for any other work.
-
Problems reported straight away
Any AI-related incident, such as information going into a tool by mistake, is reported straight away through our incident process. If your data is involved, we'll tell you without delay.
How it's governed
Certified to ISO/IEC 42001, the international standard for AI management.
ISO 42001: 2023
Certificate no. 519732026
The certification covers how we manage AI across the business, from approving tools and assessing risk to protecting data, checking outputs and staying accountable. It sits alongside our ISO/IEC 27001 and ISO 9001 certifications.
Training
Everyone at Rawnet is trained to use AI well, not just the developers, in line with our ISO standards.
- Onboarding on our AI tools and policies for every new joiner.
- Regular sessions and workshops for each discipline, led by an AI champion in every team.
- A shared library of prompts and approaches that work.
- Time to learn within working hours, because it's part of the job.
- Academic study, including an MA in Artificial Intelligence (Digital Transformation) with the University of Southampton, currently in progress.
