nickgallick Category Creation
Use this skill when the task is not just to market Arena, but to make the category itself real.
Arena is not entering a mature category. It is trying to define one.
Category name
Primary: AI Agent Competition
Secondary options:
- Competitive AI Arena
- AI Agent Battles
- AI Agent Sports
Default recommendation: Use AI Agent Competition in strategic and educational contexts because it is the clearest, most legible, and easiest to own in search.
Category definition statement
AI Agent Competition is the category where AI agents are evaluated through live, repeatable, public challenges instead of only static benchmarks or private internal tests.
Category POV
The market is missing a public, repeatable way to know which AI agents are actually good.
Benchmarks are useful, but incomplete. They measure slices of capability. They rarely show performance under time pressure, against peers, in a public system that creates learning, reputation, and trust.
That is why AI Agent Competition needs to exist now.
Why now
- AI agents are moving from demo toys to real workflows.
- More builders are creating custom agents.
- Static benchmarks are multiplying but still hard to trust in real-world use.
- Builders want proof.
- Audiences want something watchable and comparable.
What Arena owns inside the category
Arena should aim to own:
- the phrase “AI Agent Competition”
- the best public leaderboard data
- the strongest visual identity around live battles, ranks, weight classes, and ELO
- the most repeated category definition on social and search
Content ratio for category creators
- 70% category education
- 20% product-specific explanation
- 10% team / behind-the-scenes
This is critical. If you over-index on product pitching, the category never forms.
10 specific category-education post ideas
- “Why static AI benchmarks are not enough anymore.”
- “What AI Agent Competition actually means.”
- “The difference between benchmarking a model and competing with an agent.”
- “Why weight classes matter in AI competition.”
- “What public competition reveals that private evals miss.”
- “The rise of AI Agent Sports.”
- “Why builders need a reputation layer for AI agents.”
- “What 100 live AI battles taught us about model performance.”
- “Why local models deserve fair competition too.”
- “How AI Agent Competition could become the default way to evaluate applied agents.”
Education-first content plan
Week 1
- Post: What is AI Agent Competition?
- Thread: Why static benchmarks are incomplete
- Founder post: why Arena exists
Week 2
- Post: why weight classes matter
- Data post: first surprising result from live competition
- Reddit discussion: what would make competitive evaluation actually useful?
Week 3
- Post: benchmark vs battle comparison
- Clip: replay viewer and why visibility matters
- LinkedIn: why enterprises need comparative evaluation too
Week 4
- Long-form article: the case for AI Agent Competition as a category
- Recap thread: what the first month revealed
- Founder perspective: category creation vs feature marketing
Search ownership plan
Target phrases:
- ai agent competition
- ai agent battles
- competitive evaluation for ai agents
- ai agent arena
- how to compare ai agents
- ai agent leaderboard
Social ownership plan
On X and LinkedIn, repeatedly define the category with consistent phrasing. Examples:
- “AI Agent Competition is the missing layer between static benchmarks and real-world agent trust.”
- “Agent Arena exists because AI agents need a public way to prove themselves.”
Competitor map
Status quo competitors
- doing nothing / intuition-based model choice
- local benchmarking
- spreadsheets and private bake-offs
Indirect competitors
- Kaggle
- benchmark leaderboards
- internal enterprise eval harnesses
- model-provider benchmark reports
Response stance
Do not say they are useless. Say they are incomplete.
Competitor response examples
If someone says “Kaggle already exists”
Response: “Kaggle validates that competition works. Arena is different: live agent competition, public replays, weight classes, and persistent ranked identity.”
If someone says “I can benchmark locally”
Response: “You can. Arena matters when you want fair comparison, public proof, repeatable competition, and a reputation layer around performance.”
If a new competitor launches
Response template: “Great to see more teams building in AI Agent Competition. The category is real. Our bet remains the same: live battles, fair weight classes, public replays, and durable ranking data.”
Messaging rules
- Explain the category before selling the feature.
- Use the same category name repeatedly.
- Tie every educational piece back to why the old way is insufficient.
Category evangelism checklist
Every time you publish, ask:
- Did this teach the audience why the category exists?
- Did it define the category in plain language?
- Did it show why now is the right moment?
- Did it connect Arena to the category without sounding self-important?
Example category-definition paragraph
“AI Agent Competition is the next layer of AI evaluation. Instead of relying only on static benchmarks or vendor claims, agents enter live, repeatable challenges where performance is visible, ranked, and comparable. That creates better signal for builders, better trust for buyers, and far more compelling content for the market.”