Here are our 9 golden rules for a successful, AI-powered tone of voice programme.

1. Prompt an LLM to audit hundreds of comms

from across your business, and your competitors’ comms too (alongside a human).

Why? You get a full sense of what’s working and what’s not.

2. Create specific, detailed guidelines.

‘We’re bold’ won’t cut it. Do you start with opinions? Do you ask controversial questions? Do you drop in one-word sentences?

Why? You’ll need specific rules if you ever want AI (or humans, for that matter) to write in your tone. Speaking of which…

3. Create a tone of voice prompt

alongside the guidelines.

Why? It’ll mean everyone can create on-tone comms in a flash.

4. Create even more prompts,

like a newsletter writer, a social post writer, a whitepaper writer… or whatever’s best for you. Include your tone prompt in them all.

Why? Speed, efficiency, consistency.

5. Run prompt training

alongside your tone of voice training.

Why? It’ll mean even the first output is in a good state.

We can run prompt training for you.

6. Have a writing amnesty,

asking heads-of across your organisation to send you the comms most in need of a rewrite.

Why? It’ll involve them in the process, and help you with the next step…

7. Use your LLM to rewrite all those comms

(and get good human writers to do the most important ones).

Why? It’ll show everyone there’s a big change happening.

8. Put those rewritten comms into your prompt,

or use them to fine tune an AI.

Why? Your AI’s writing will get even better.

9. Keep human tone of voice experts involved

throughout.

Why? AI can expand their influence, but it can’t replace them. 

Interested in hearing more? We ran a ‘Tone of voice in the time of AI’ webinar, covering:

  • How to build an AI-friendly tone of voice
  • How other brands are already using GenAI in their tone of voice programmes
  • What you need to do to get ahead of the game
  • How tone of voice programmes might look in five years’ time
You can watch that here.

Nick Padmore Screen

Written by Nick Padmore, Head of Language at Definition.