Guides
Brand voice, from the evidence rather than the adjectives.
Four short pieces on the questions people arrive with. Each one is written from the same measurements the free test runs, and each ends with something you can do this afternoon without buying anything.
A brand voice guide template, with the hard part marked
The twelve sections a Wordprint uses, free as a file, with an honest note on which five cannot be filled in without counting.
How to write brand voice guidelines that change what gets written
Why most go in a drawer within the month, and how to write one from evidence that a new hire, an agency or ChatGPT can follow.
ChatGPT custom instructions for your brand voice
Why "friendly and professional" changes nothing, what an instruction has to look like to bite, and where it goes in ChatGPT, Claude, Copilot and Gemini.
Tone of voice examples
The difference between an adjective and a rule, shown on pronoun policy, rhythm, pricing in prose and saying no.
The Check, inside your tools
Your Wordprint as something ChatGPT, Claude, Copilot and Gemini consult rather than remember: the page, the API and the MCP server.
How to measure brand voice
Twenty-two things you can count, the four that separate a real voice from a fake, and how to do it with a spreadsheet.
Benchmark: how UK accountancy firms sound
Thirty-one firms on one instrument: a median of 72, a firm at 100, the failure twenty-two of them share, and the word twenty of them use.
Benchmark: how US CPA firms sound
Thirty-five American firms on the same instrument as the British: level at 73, a fuller phrasebook, and a punctuation mark that proves the method is not an AI detector.
Benchmark: how UK agencies sound
Twenty-two agencies: the highest median of any sector, no shared phrasebook at all, and the fault half of them share with the accountants.
Benchmark: how UK recruitment agencies sound
Forty agencies: mid-table at 73, the most evidence of any sector, one agency at 96, and three in four unable to stop a sentence.
Benchmark: how cloud, data and AI consultancies sound
Forty firms on one instrument: a median of 56, nobody in the top band, all forty failing the same measure, and the two words thirty-four of them share.
Or skip the reading
Paste a page you have published into the free test and get the four scored measures, the phrases you share with your sector and the sentences that are genuinely yours, in about a minute.