SOURCE-BASED ANALYSIS · NO HANDS-ON RATING
Choose samples that represent the voice you want to repeat
Jasper documents that Brand Voice can be created from text, uploaded files or URLs and can be kept private or made visible to the workspace. Start with approved material that reflects the intended tone, not merely the content that is easiest to find. Exclude outdated campaigns, unreviewed drafts and pages whose tone is intentionally exceptional.
Create a small acceptance fixture
Before generation, write three short briefs that exercise different parts of the voice: a factual service explanation, a restrained customer email and a concise social post. Keep the permitted facts, audience and required call to action constant. Jasper's preview compares output with and without the selected voice; use that contrast to identify what the profile changes rather than treating either draft as automatically correct.
| Fixture | Expected signal | Failure to flag |
|---|---|---|
| Service page | Plain, specific claims | Invented benefit or inflated certainty |
| Customer email | Warm and direct | Casual language that weakens the message |
| Social post | Recognisable rhythm | Slogan repeated without context |
This is a hypothetical review fixture, not a measured Jasper test.
Separate voice from facts and style rules
A voice describes how content sounds. Product facts, approved claims and disclaimers belong in maintained knowledge sources, while terminology and punctuation rules belong in a style guide. Do not use a voice sample as the only authority for a price, capability or legal qualification. Record which layer should resolve each correction.
Review visibility before setting a default
- Name the owner and intended audience for the voice.
- Confirm every sample is approved for that audience.
- Generate the same fixtures with and without the voice.
- Check factual fidelity, tone, terminology and exclusions separately.
- Keep the voice private while revising it.
- Only an authorised administrator should consider a workspace default after review.
Keep a change record
Store the sample list, profile description, fixture results, reviewers and decision date. When the brand changes, rerun the same fixtures before replacing the default. A stable comparison set makes changes visible without inventing a numerical quality score.
Sources and verification
Official documentation checked on 2026-10-29. Worked examples are hypothetical planning exercises, not measured product benchmarks.