A naming exercise can produce many attractive words while missing basic constraints: difficult pronunciation, an unwanted meaning in a target language, or a resemblance to terms the team has already excluded. A shortlist is more useful when every candidate carries a review record explaining why it remains under consideration.
This guide creates an original name screening sheet with ChatGPT. It focuses on linguistic and editorial rules, with fictional examples and documentation checked on October 6, 2026. It does not establish trademark clearance, domain availability, or final suitability. Those are separate checks that a generated shortlist cannot confirm from the naming brief alone.
Define the naming brief and audience
State what is being named: a product, internal program, publication, or feature. Identify the audience, intended associations, target languages, and where the name will appear. A name read in a document may face different pronunciation demands from one spoken frequently in meetings.
Separate desired qualities from mandatory constraints. “Friendly” is a preference requiring judgment; “must not use a listed internal project name” is an exclusion rule. Keep both categories explicit so a appealing candidate does not bypass a mandatory condition.
Record the brief version and decision owner. If the owner later changes a target language or brand requirement, the shortlist should be reviewed against that new version. Do not retain an earlier pass label as if the criteria remained unchanged.
Write exclusions that can be applied consistently
Specify the actual prohibited words, forms, or associations using the team's approved list. Define whether a rule applies to exact matches, substrings, similar sounds, or semantic associations. These are different tests and should not be collapsed into a vague instruction to “avoid confusion.”
For a fictional English-language internal program, a mechanical rule could limit candidates to a stated number of letters and exclude spaces. That rule can be checked directly. Pronunciation ease and tone require reviewer judgment and should not be reported as mechanical passes.
Use the current OpenAI prompting guide as a capability reference for expressing context and boundaries. The screening rules below are an original method; listing them in a prompt does not guarantee that every generated candidate satisfies them.
Build the screening sheet
Give each candidate an ID and preserve its original spelling. Record the rule version used for screening and the result for each mandatory rule. Keep uncertain judgments visible rather than treating them as approvals.
| Screening field | Required content |
|---|---|
| Candidate ID | Stable shortlist reference |
| Proposed spelling | Exact form generated or supplied |
| Intended association | Rationale tied to the naming brief |
| Mechanical checks | Length, spacing, and exact exclusion results |
| Linguistic review | Pronunciation and meaning notes by target language |
| Similarity concern | Possible confusion requiring review |
| Disposition | Retain, reject, or hold with a reason |
Add reviewer and date fields for linguistic judgments. A model-generated explanation of a word's meaning should remain provisional until checked by an appropriate language reviewer or approved reference.
Generate candidates separately from validation
Ask ChatGPT for a manageable candidate set rather than hundreds of names at once. Require it to explain the intended association briefly and identify the rules it attempted to follow. Then run a separate screening pass against the supplied list.
Generate a small naming candidate set for the supplied brief.
Preserve the mandatory exclusion rules and target-language scope.
Give each candidate an ID, exact spelling, and short intended association.
Do not claim trademark, domain, or product-name availability.
In a separate screening table apply each explicit mechanical rule.
Mark pronunciation, cultural meaning, and sound-alike concerns as provisional review items.
Return rejected candidates with reasons and a shortlist awaiting human language review.Check mechanical rules independently rather than relying on the model's self-report. A candidate can exceed a length limit despite being marked compliant. Screening should identify the actual result and retain the rejected candidate's ID for review history.
Review target languages and pronunciation
Ask appropriate reviewers to examine the exact spelling and likely spoken forms in each target language. Consider stress, syllable boundaries, transliteration, and resemblance to unwanted terms. A meaning note for one language does not establish suitability in every market.
Avoid claims that a name is universally pronounceable or culturally safe. The reviewed scope should name the languages and contexts checked. If a target language lacks a reviewer, mark that screen pending rather than turning the absence of a reported concern into a pass.
If the name will be written in more than one script, record the approved form for each script and review them separately. A transliterated form can introduce a different sound or association. Preserve that relationship in the sheet instead of assuming the same result applies automatically.
Work through a fictional screening case
Imagine Candidate C-01 satisfies the fictional brief's spelling and length rules but resembles an excluded internal program when spoken. The mechanical screen can pass while the pronunciation screen remains on hold. Those results are compatible because they evaluate different conditions.
Candidate C-02 may fit the desired tone but exceed the length rule. Record the rejection rather than quietly shortening it and retaining the original result. A shortened version is a new candidate spelling requiring another screen.
Candidate C-03 may have no reported issue in the first language and no completed review in the second. Its status should remain pending for the second language. Do not label it approved simply because the first reviewer liked it.
Rank only candidates that meet the reviewed rules
After mandatory screening, compare retained candidates against the desired qualities. Use the owner's approved rubric, with definitions for terms such as memorable, professional, or distinctive. Explain preferences through the brief rather than present subjective scores as objective measurements.
Avoid making the shortlist consist of minor spelling variations of the same candidate unless that comparison is deliberate. A useful shortlist offers meaningfully different directions while staying inside the same constraints. Record the differences so the owner can choose a direction rather than guess at the rationale.
If reviewers disagree, preserve their reasons and the target context. The decision memo workflow can organize competing preferences without implying that a model's ranking settles the decision.
Keep linguistic review separate from clearance
The final screening sheet should identify which checks were completed and which remain outside the linguistic exercise. Do not use “available” as shorthand for “passed the word rules.” A candidate's suitability also depends on checks the naming team performs through its normal process.
Record the selected candidate and actual decision owner separately from the proposed shortlist. If the chosen spelling changes, rerun affected screens and preserve the new version. Keep rejected names with their reasons so later brainstorming does not repeatedly reintroduce the same excluded idea.
Maintain the brief, exclusion list, screening sheet, reviewer notes, and final naming decision together. The result is a traceable shortlist with explicit limits, rather than a collection of appealing names accompanied by unsupported assurances.
Frequently asked questions
Can ChatGPT guarantee that a name works in every language?
No. Define the target languages and obtain appropriate review. Unreviewed languages should remain outside the stated approval scope.
Are mechanical checks enough to approve a candidate?
They cover explicit spelling rules, not pronunciation, associations, or other suitability checks. Keep those review categories separate.
What if a name is changed after screening?
Treat the revised spelling as a new candidate version and rerun the affected checks. Do not reuse the old pass labels automatically.
Can the shortlist claim domain or trademark availability?
Only a separate appropriate check could support those claims. This linguistic worksheet does not establish clearance or availability.
Why retain rejected candidates?
Their recorded reasons prevent repeated proposals and help reviewers understand how the shortlist developed under the approved rules.
