Our approach
LightSpeed uses AI to improve research, drafting, prototyping, development support, content operations and discovery workflows. We do not treat AI output as final work. Human review stays central to how we plan, build, publish and maintain websites.
Where we use AI
We use AI as an assistant, always with the conditions below:
- Drafting website copy, with human review and claim checking
- Summarising source content, with the source kept
- Suggesting FAQs, schema and meta descriptions, which are reviewed and technically validated
- Code assistance, with developer review, security review and a GPL compatibility check
- Design prototyping, checked against our Figma design system and accessibility standards
Human accountability
AI may assist with drafts, summaries, translation, ideation, code suggestions and content analysis. Final content, code, design and recommendations are reviewed by people, and people remain accountable for accuracy, quality, ethics, accessibility, security and client suitability.
Some tasks stay with people:
- Final approval of public website copy
- Legal and privacy review of policy wording
- Approval of statistics, outcomes and client claims
- Security-sensitive technical decisions
- Client-specific commercial recommendations
Transparency
Where AI meaningfully assists public content, code or workflows, we aim to be open about it. We avoid presenting AI-generated output as unchecked human expertise, and we are building internal records of the AI tools and prompts we use so that important AI-assisted work can be traced and reviewed.
Privacy and data protection
We do not knowingly put sensitive client data, private customer data, credentials, payment details or confidential business material into public AI tools without approval and appropriate safeguards. Any AI-enabled workflow is reviewed against privacy, security and data-protection requirements before launch.
Open source and licensing
As a WordPress-focused team, we pay close attention to GPL compatibility, licensing, copyright and open-source obligations. AI-assisted code, media and content are reviewed before they are used in client or public work.
Quality over volume
We do not use AI to mass-produce low-quality content. AI supports structured thinking, research, content operations and technical workflows, while human editors and developers keep responsibility for the final result. We do not publish hard return-on-investment claims without evidence.
Responsible AI services
When we plan AI-powered services such as chatbots, search visibility support, analytics or content workflows, we define the approved sources, excluded topics, fallback behaviour, escalation routes, privacy rules and testing requirements before launch.
Review and improvement
AI governance is an ongoing process. We are putting in place a regular review cycle for our tools, prompts, outputs and policies, and a simple incident log, so that problems lead to better workflows and not only quick fixes.
Questions
If you have a question about how we use AI in your project, get in touch.
This page describes our own approach to responsible AI. It is not legal advice. Legal, privacy and regulatory requirements should be confirmed with a qualified adviser.