| name | company-namer |
| description | This skill should be used when naming companies, products, features, apps, or brands. It applies when the user asks to brainstorm names, rename something, evaluate name candidates, or run a structured naming process. Triggers on "name my company", "help me name", "rename", "brand name", "product name", "what should I call", or any naming/branding request. |
Company Namer
Professional naming methodology based on Lexicon Branding and Lunour agency practices. Guides the user through a structured process that moves from strategy to generation to evaluation -- producing names with energy and distinctiveness, not safe descriptive labels.
Core Principles
These principles override default AI naming behavior, which tends toward the generic and descriptive:
-
Name the meaning, not the function. Name what the company represents at the emotional/aspirational level, not what it literally does. "CloudTech Solutions" describes a function. "Vercel" evokes velocity, versatility, and excellence.
-
Seek the tension zone. The best names polarize. If everyone in the room nods and says "yeah, that's fine," the name is in the invisible zone -- forgettable. Aim for names where half the audience loves it and half questions it. That energy is memorability.
-
Quantity drives quality. Do not stop at 50-100 candidates. Professional firms generate 2,000+. Target at least 300 candidates before filtering. The breakthrough name is rarely in the first batch.
-
AI generates, humans judge. AI excels at producing volume. It cannot feel which name has energy, sense tension in a room, or distinguish "surprisingly familiar" from "generically pleasant." Generate with AI, evaluate with human judgment.
-
Sound carries meaning. Hard consonants (K, T, P) feel strong and precise. Soft sounds (L, M, S) feel smooth. V is daring. Say every name out loud. If it feels wrong in the mouth, it is wrong.
Workflow
Follow these five phases in order. Do not skip phases. Each phase builds on the previous one.
Phase 1: Map the Landscape
Before generating a single name, understand the competitive territory. The most important decision in naming is deciding what the company is not.
To gather context, ask the user:
- What does the company/product do? (Keep this brief -- details come in Phase 2)
- Who are the top 10-20 competitors?
- Which competitor names do they admire, and why?
Then analyze:
- Classify each competitor name by pattern (descriptive, invented, metaphorical, compound, acronym, founder name, Latin/Greek root, fragment blend). See
references/naming-knowledge.md for the full pattern taxonomy.
- Identify which patterns are over-represented in the space. These are off-limits.
- Declare clear "we're NOT going there" boundaries with the user.
This constraint is liberating -- it eliminates the obvious and forces creative territory.
Phase 2: Find the Ultimate Benefit
This is where most naming efforts go wrong. They name the feature instead of the feeling.
To climb the benefit ladder, walk the user through four questions:
- Feature: What does the product/service actually do?
- Functional benefit: What problem does that solve?
- Emotional benefit: How does solving that problem make someone feel?
- Ultimate benefit: What bigger aspiration or identity does that connect to?
The ultimate benefit is the creative springboard. Present it as 1-2 words.
Example: A project management tool.
- Feature: Task tracking and team collaboration
- Functional: Projects get done on time
- Emotional: Feeling in control, less chaos
- Ultimate: Clarity
"Clarity" as a concept opens up optics (lens, prism, focus), water (crystal, pure, spring), weather (clear sky, horizon), and dozens of other adjacent fields. "Project management" opens up nothing.
Also ask the user for 3-5 adjectives describing how they want people to feel when they hear the name.
Phase 3: The Treasure Hunt
Generate name candidates at scale. Use the ultimate benefit concept as the seed and explore from every angle:
- Synonyms and near-synonyms of the ultimate benefit
- Latin and Greek roots (use etymonline.com mentally -- explore clar-, luc-, vid-, etc.)
- Adjacent fields -- at least 5 different domains that metaphorically connect to the concept (optics, water, weather, astronomy, music, architecture, etc.)
- Sound and fragments -- which syllables feel right? Combine word fragments: first half of one word + second half of another
- Metaphors -- what objects, experiences, or moments represent the concept?
- Invented words that follow English phonetic rules and feel like they could be real
Use "approximate thinking": sit in the zone between absurd and obvious. Half-formed ideas, weird combinations, names that provoke a laugh -- write them all down. Do not judge early. Refinement comes later.
Generate candidates in batches of 50-100. Present them organized by exploration angle (synonyms, roots, adjacent fields, etc.). Target 300+ total candidates across multiple rounds.
Critical AI bias to counteract: First-pass AI-generated names tend toward the bland and generic. After the first batch, explicitly push toward more inventive, surprising, and emotionally charged territory. Include at least 30% invented/fragment-blend names in every batch.
Phase 4: Filter and Test
Apply filters in this order:
- Say it out loud. Remove anything hard to pronounce or that sounds awkward.
- Competitor collision. Remove anything that sounds like an existing competitor.
- Descriptive trap. Remove anything that literally describes what the company does.
- Processing fluency. Remove anything over 3 syllables or hard to spell.
The surviving 40-60 names move to the tension test. For each finalist, ask: "Does this name make me feel something? Or do I just... not hate it?" Names that provoke no reaction are in the invisible zone. Cut them.
Narrow to the top 10.
Phase 5: Final Evaluation
Score each top-10 finalist on six dimensions (1-10 scale). Consult references/naming-knowledge.md for the full scoring rubric.
| Dimension | Key Question |
|---|
| Distinctiveness | Does it stand out from competitors? |
| Processing Fluency | Easy to say, spell, remember? |
| Emotional Resonance | Connects to the ultimate benefit? |
| Energy | Creates a reaction? Has tension? |
| Longevity | Works in 10 years? Doesn't box in? |
| Availability | Domains, social handles, trademark? |
The winner is rarely the highest average score. Look for the name scoring 9-10 on distinctiveness and energy, even if it scores 6 elsewhere. Remarkable beats balanced.
Present the final shortlist with scores, a brief rationale for each, and a clear recommendation with reasoning.
Quick Mode
When the user wants fast brainstorming without the full process (e.g., "give me some name ideas for X"):
- Ask 2-3 quick questions to understand the product and desired feeling
- Skip directly to Phase 3 treasure hunt, generating 50-100 candidates
- Highlight the top 10 with brief rationale
- Offer to run the full process for deeper exploration
Anti-Patterns to Avoid
- Do not default to descriptive names. "SmartTask," "DataFlow," "CloudSync" -- these are the invisible zone.
- Do not optimize for consensus. If presenting options, flag which ones are polarizing and explain that polarization signals energy.
- Do not let domain availability kill a good name. Domains can be acquired or modified (get[name].com, [name]hq.com). A mediocre name is forever.
- Do not present names without saying them aloud mentally. Sound matters as much as meaning.
- Do not stop generating too early. If the user seems satisfied with the first batch, gently push for one more round -- the best name is often in the second or third pass.
Resources
Detailed case studies (Swiffer, BlackBerry, Impossible Foods, Pentium, Vercel, Raremarq, Axio), the full scoring rubric, sound symbolism guide, benefit ladder framework, naming pattern taxonomy, renaming signals, and common pitfalls are in references/naming-knowledge.md. Consult this reference throughout the process, especially during Phase 3 generation and Phase 5 evaluation.