Responsible Use of AI

We take a cautious and responsible approach to AI. The technology is powerful, but it has real limitations — and real risks if deployed carelessly.

We consider where data is processed, how it is stored, and whether AI is appropriate for the task. If something needs to be exact, auditable, or legally precise, AI may not be the right tool — and we will say so.

Our principles

The guidelines we follow when implementing AI solutions

Human oversight

AI assists, it doesn't decide. Every workflow includes review points where humans can verify, correct, or override.

Data boundaries

We're clear about what data AI systems can access, where it's processed, and how long it's retained.

Appropriate use

We won't use AI where deterministic logic would be more reliable, or where the stakes demand perfect accuracy.

Transparency

Users and staff should know when they're interacting with AI-generated content or AI-assisted decisions.

Reversibility

If an AI system makes a mistake, it should be easy to identify, trace, and correct.

Proportionality

The complexity of AI solutions should match the problem. We don't over-engineer or add AI where simpler tools would work.

What we won't do

Deploy AI for high-stakes decisions without human review
Send sensitive data to third-party AI services without explicit consent and clear data handling agreements
Use AI to generate content that will be published without human editing
Claim AI can do things it demonstrably cannot
Staying current. AI is moving fast. We keep up with developments in capability, security, and regulation — and we'll tell you when something changes that affects your systems.

Using AI responsibly: common questions

Is it safe to put customer data through AI?+
It can be, with the right service and settings. Before we start we agree which data is used, where it is processed, and how long anything is kept.
Will AI ever send something to my customers on its own?+
Not in anything we build unless you decide it should. Our inbox triage, for example, only ever writes drafts; a person sends them.
How do I know when AI has been used?+
We recommend labelling AI-assisted content and decisions in the systems you use, so staff can see what came from where, and we can build that in.
What do I need to think about under UK data protection law?+
The same as for any supplier that handles personal data: what is shared, why, and for how long. We help you write that down. For formal legal advice, speak to a solicitor.
When do you recommend not using AI?+
When the answer must be exact every time, such as prices, calculations or legal wording. Ordinary software does those better.

More AI Services

OverviewEmail triageBusinessAutomationContentEthics

Questions About AI Ethics?

Get in touch to discuss responsible AI implementation.

Contact Us