Our AI Policy
How we use AI
We use AI tools to aid our development work. There's nothing mysterious about it and we want to be clear and upfront about how and why we use it.
Why we're talking about this
AI is everywhere now, and most people use it without being upfront or explaining what it actually means to their work.
We could do the same, but instead, we're being upfront: we use it selectively, in specific parts of our workflow, and we think you should know where and why.
Not because we're defensive about it, but because trust matters.
When you understand what goes into building your product, you can make better decisions about it and you know exactly what you're getting.
Where AI helps us
Where it makes sense, using AI speeds up our process and keeps project costs down.
Boilerplating
Setting up project structures, scaffolding components, generating repetitive configuration. It's work that needs doing, but it's not where the thinking happens.
Prototyping
When we're exploring an idea or testing an approach, AI can help us spin something up fast. We use it to validate a direction before we commit to building it properly.
Code review
A second set of eyes catching edge cases, suggesting cleaner patterns and spotting bugs before they reach production. It's good at catching what humans naturally skim over.
Bug fixing
Debugging is pattern-matching, and AI's useful for that. It can spot common issues and suggest fixes. We verify, validate and implement the actual solution.
Refactoring & code cleanup
Identifying opportunities to simplify or reorganise existing code, suggesting cleaner patterns. AI can spot where code is redundant or could be more efficient. We make the final call on whether the refactor is worth doing.
Documentation
Writing technical docs is necessary and tedious. AI helps us draft and structure them. We make sure they're accurate and actually useful.
Where AI doesn't feature
We only use it where it actually adds value. You're not paying for AI in every line of code.
Client communications
Understanding your needs, asking the right questions, building the relationship that makes the project work. That's human work, and there's no substitute for it.
Strategy & direction
How your product should work, what problems it solves, which features matter. That comes from conversations with you and deep understanding of your business & your customers.
Architecture decisions
How systems are structured, how services communicate, what patterns we use. These are deliberate choices based on your actual needs, the scale you're aiming for, and long-term maintainability.
Design & user experience
Aesthetic direction, interaction design, information architecture, accessibility. These come from our designers and developers working with you to understand your users.
Security & compliance
Decisions about how we handle sensitive data, what security measures matter for your specific project, compliance requirements. These require strategic judgment, not automation.
Testing & quality assurance
While AI can help with bug-spotting, the strategy behind testing, i.e. what matters to test and how thoroughly, comes from understanding your product and users.
Why this matters
Strategy, architecture, and design are where the real value sits. AI is useful tooling, nothing more.
We use it to remove friction and work faster where it makes sense. We don't use it where decisions need to be made, and we don't use it just because it's there.
All code generated with AI goes through the same review, testing, and quality standards as anything else we ship. The tool doesn't change our standards.
When we can deliver your project faster without cutting corners, you benefit from that. Lower turnaround time means lower costs. But only where it's genuinely useful, not padding the process with unnecessary AI.
Everything we build gets human review, human judgment, and human accountability. That's how we work.