A useful ChatGPT branding session starts with a small job you can review. For an independent specialist, that might be a service-page opening with a clear audience, a defined service and one next step. You can then check whether the draft preserves those decisions before publishing it.
This guide shows that process for a fictional US solo founder who offers onboarding audits to founder-led B2B SaaS companies. The brief and both candidates below are editorial illustrations prepared on September 30, 2026. They are not live ChatGPT outputs, customer stories or evidence of business results.
→ Jump to: Set Up Context | Approve the Brief | Use the Exact Prompt | Review and Repair | Keep a Run Record
Set Up Context
ChatGPT can help you ask discovery questions, explore positioning alternatives and draft language. The founder still needs to approve the business decisions and substantiate any claims. A prompt alone cannot verify a founder's career history, customer results or permission to use a testimonial.
As checked on September 30, 2026, ChatGPT projects group chats, instructions and uploaded or connected sources. Put a short approved brief in the project instructions or directly in the prompt, and confirm the intended files are available. A ChatGPT project does not imply access to a local folder on your computer.
App instructions and an API system message are separate controls. This workflow uses the ChatGPT app; it does not assume that the app exposes a field called a system prompt or that an API integration has been configured.
For a task that needs current external facts, ChatGPT web search can provide cited sources, subject to workspace settings. Open the supporting pages and inspect the relevant claim. The service-page fixture here needs only the approved brief, so a future test should disable external research where possible and record any tools used.
If the audience or service is still unclear, start with the brand strategy foundation guide. AI can suggest questions and candidate answers, but unsupported answers should remain hypotheses until you have evidence.
The BrandKernel vs ChatGPT comparison distinguishes defining brand context from supplying that context to a general-purpose assistant.
Approve the Brief
Here is the completed miniature brief for the fictional founder. It describes an approved offer, without inventing client outcomes.
- Audience: founders at founder-led B2B SaaS companies who want to examine their new-user onboarding.
- Service: an onboarding audit of the current signup flow, welcome emails and in-product guidance.
- Deliverable: a written list of observed friction points with recommended changes in priority order.
- Evidence available: the approved service description only. No customer results, testimonials, revenue figures or activation measurements are supplied.
- Voice: direct, concrete and restrained. Name the work. Avoid hype, jargon and labels such as revolutionary or world-class.
- Excluded promises: guaranteed activation or revenue gains, implementation services, ongoing coaching and invented qualifications.
- Next step: ask the reader to request an onboarding audit and share the current flow.
This is a working editorial brief. It is not a complete product schema or a promise that any app will follow every constraint automatically. For help turning voice adjectives into observable choices, use the brand voice examples.
Use the Exact Prompt
Paste this complete input into a fresh chat when you want to try the fixture:
Draft a service-page opening for this approved brief.
Audience: founders at founder-led B2B SaaS companies who want to examine their new-user onboarding.
Service: an onboarding audit of the current signup flow, welcome emails and in-product guidance.
Deliverable: a written list of observed friction points with recommended changes in priority order.
Evidence: only this approved service description. No customer results, testimonials, revenue figures or activation measurements are supplied.
Voice: direct, concrete and restrained. Name the work. Avoid hype, jargon and labels such as revolutionary or world-class.
Excluded promises: guaranteed activation or revenue gains, implementation services, ongoing coaching and invented qualifications.
Next step: ask the reader to request an onboarding audit and share the current flow.
Return one headline, a two-sentence opening and one CTA sentence. Use first person singular for the consultant and US English. Use only supplied facts. If a requested claim has no support, omit it rather than inventing evidence.
The output format is short enough to review against each field. Keep the brief intact when comparing revisions; silently changing the input makes it harder to tell what improved.
Review and Repair
Generic candidate, authored illustration, not a ChatGPT response:
Unlock your business's full potential
I help businesses transform their customer journeys with revolutionary growth solutions. My proven approach guarantees higher activation and revenue.
Book your growth transformation today.
Apply five review criteria before accepting it:
- Audience fit: fail. Businesses is broader than the approved founder-led B2B SaaS audience.
- Service scope: fail. Growth solutions does not describe the onboarding audit or its deliverable.
- Factual claims: fail. Proven approach and guaranteed activation and revenue have no supplied evidence. The guarantee is expressly excluded.
- Voice constraints: fail. Unlock, revolutionary and transformation replace concrete work with hype.
- Reader next step: fail. The CTA does not ask for the current onboarding flow or name the audit.
Decision: reject the candidate. A sentence that sounds persuasive cannot compensate for an invented guarantee. Request a revision that restores each failed brief field, then check the complete replacement.
Repaired candidate, authored illustration, not a ChatGPT response:
An onboarding audit for founder-led B2B SaaS
I review your signup flow, welcome emails and in-product guidance to identify friction in new-user onboarding. You receive a written list of observed friction points and recommended changes in priority order.
Request an onboarding audit and share your current flow.
Applied verdict: accept this passage for the fixture's editorial brief. It names the audience and audit, preserves the deliverable, introduces no unsupported outcome, uses concrete language and gives the specified next step. Acceptance here means the passage meets these five checks. It does not establish conversion performance or confirm that the service is right for a real customer.
Keep a Run Record
A future live evaluation should preserve the run date, app, displayed model, relevant settings, available files, search or other tool access, exact input and unedited raw output. Add your verdict and the specific lines that passed or failed. This article records no live model or subscription test.
If a real run produces a guarantee, retain that raw draft as evidence of the failure before requesting a repair. Review the repaired output again. If the actual service changes, revise the approved brief first and evaluate the next draft against that new version.
The cross-app evaluation guide explains how to keep those conditions comparable. BrandKernel's research proposal describes Specification Fidelity as a proposed evaluation approach; empirical validation is deferred. The five checks in this article are an informal editorial rubric, not a validated BrandKernel benchmark.
Frequently Asked Questions
What should I give ChatGPT before asking for branding copy?
Provide an approved audience, service, deliverable, available evidence, voice constraints, excluded claims and reader next step. Keep unverified customer results out of the brief, or clearly mark them as claims that cannot be published yet.
Where should I put the brief in the ChatGPT app?
Place it directly in the prompt or in project instructions and confirm which supporting files are available. App instructions are distinct from an API system message, and a project does not automatically expose files from your computer.
Can ChatGPT help me explore positioning before I have a finished strategy?
It can suggest questions, alternatives and language for review. Treat those alternatives as hypotheses, then use your service knowledge, customer evidence and deliberate business choices to approve the positioning.
How do I decide whether a draft is ready to publish?
Check audience fit, service scope, factual claims, voice constraints and the next step. Reject or repair unsupported guarantees and altered services even if the writing otherwise sounds polished.
Were the example candidates generated by ChatGPT?
No. Both candidates are authored editorial illustrations dated September 30, 2026. No live ChatGPT run, model comparison, customer outcome or validated benchmark is reported here.
Build a Brief You Can Review
Use the brand strategy template to document the decisions behind your next draft. Start with one passage, check every claim and save the accepted version with its brief.