Which AI tools we use, what happens to your information, and where a person stays in control.
AI is a big part of what we do, so we want to be clear about how we use it, what happens to your information, and where a person stays in charge. This policy applies to our own work and to the AI systems we build for clients.
1. Our principles
- People stay in control. AI drafts and suggests. A person checks anything important before it is sent, published or relied on.
- Only what’s needed. We share the minimum information an AI tool needs to do the job, and remove personal details where we can.
- No training on your data. We use business and API accounts with settings that stop providers using our inputs to train their models, wherever the provider offers this.
- Honest about limits. If AI isn’t the right answer for a problem, we’ll tell you.
2. The AI tools we use
| Provider | Tools | What we use them for |
|---|---|---|
| OpenAI | ChatGPT and the OpenAI API | Drafting, analysis and powering client automations |
| Anthropic | Claude and the Claude API | Drafting, coding help, analysis and client automations |
| Gemini and the Gemini API | Drafting, research and client automations | |
| xAI | Grok | Research and drafting |
These providers may process information in the United States. Our Privacy Policy explains the safeguards we use for international transfers. We review this list when we add or remove a tool.
3. How we use AI in our own work
- Drafting proposals, emails, website copy and guides, which we then edit and check.
- Helping write and review code, which we test before it goes live.
- Summarising notes from calls and organising project tasks.
- Monitoring the health of websites we look after, with alerts reviewed by a person.
We don’t put client logins, payment details or special category information (such as health data) into AI tools.
4. AI systems we build for clients
When we build an AI assistant or automation for your business, for example drafting replies to enquiries or preparing quotes:
- We agree with you in writing what it will do, what information it can see and which provider it uses.
- It includes a step where you or your team approve its output before it reaches a customer, unless you ask us in writing to remove that step.
- We test it before launch with realistic examples, including cases where it should hand over to a person.
- We help you tell your own customers that AI is involved where that’s appropriate, and update your privacy notice if needed.
- You stay the controller of your customers’ information. We act as your processor under our Terms and Conditions.
- Usage costs from the AI provider are passed on at cost, and you can see what you are paying for.
5. Accuracy and fairness
AI can be wrong, out of date or biased, and can sound confident when it isn’t right. That’s why we check outputs, keep a person in the loop for anything that matters, and design systems so that they don’t make decisions with legal or similarly significant effects on people without human review.
6. Things we won’t do
- Use AI to make up reviews, testimonials or results.
- Pretend AI-written messages come from a named person who hasn’t checked them.
- Build systems that profile people or make decisions about them in ways that would be unlawful or unfair.
- Use your information to train or improve AI products for anyone else.
7. Ownership of AI-assisted work
Work we create for you with help from AI is treated like any other work under our terms: once you have paid in full, the finished deliverables are yours. AI providers’ own terms still apply to their tools.
8. Questions or concerns
If you’d prefer we don’t use AI on your project, or want to know exactly how it’s used, email hello@wrkflw.co.uk. We’ll always respect a reasonable request to keep AI out of your work.