Making your voice machine-readable
Something fundamental is changing about how content gets created. AI writing tools are no longer a novelty – they're part of the workflow. People use them to draft emails, generate social posts, write first versions of web copy, summarise documents and brainstorm ideas.
And most AI tools have no idea what your organisation sounds like.
The copy-paste problem
The way most people feed brand voice into AI tools is laughably crude. They copy a few sentences from their brand guidelines into a system prompt and hope for the best. "Write in a warm, professional tone." "Be friendly but authoritative." "Sound like our brand."
It sort of works. But it's inconsistent, unreliable and entirely dependent on whoever wrote the prompt remembering to include the right instructions. There's no structure, no persistence, no shared standard. Every time someone starts a new conversation with an AI tool, they're starting from scratch.
Structured voice as infrastructure
Imagine a different approach. Instead of copy-pasting vague instructions, your AI tools could query your style guide directly. They could pull your voice definition, your tone guidelines, your terminology preferences and your editorial conventions from a structured, machine-readable source – automatically, every time, without anyone having to remember to include them.
This is where style guides are heading. Not just documents for humans, but structured data that machines can use too. A canonical, authoritative definition of who you are and how you sound, available to any tool that needs it.
Your visual brand already has a structured format. You have hex codes for colours, font files for typography, SVGs for logos. These structured formats mean that any design tool can render your brand correctly. Your verbal brand deserves the same treatment.
What this means in practice
A machine-readable style guide could:
- Feed AI writing tools with your actual voice, tone and style – so generated content sounds like you from the first draft
- Integrate with your CMS – surfacing relevant guidelines as people write, right where they need them
- Connect to content quality tools – checking whether published content matches your stated guidelines
- Enable API access – so any tool in your stack can query your style definition programmatically
The technology exists. What's been missing is the structured format – a standard way for organisations to define their voice that tools can reliably consume.