Category Positioning SOP
Intent
Turn a thin AI wrapper startup into a category leader. Most AI startups sound like wrappers over rented models. This SOP shows how to design a category narrative that makes a product the unavoidable answer in its category.
Inputs
- Product name and description
- Current positioning (how the company describes itself)
- Top 3 competitors and their positioning
- Target customer (who buys, who uses)
- Unique technical or workflow advantage
Process (SOP)
- Run the Wrapper Test and decide whether the product is a wrapper.
- Find the constraint outside the model: vertical, infrastructure, security, workflow, or data.
- Design the category narrative: enemy, new category, core idea, and proof.
- Write the one-liner: target customer, outcome, unique mechanism, proof metric.
- Build the proof stack with a receipt for every claim.
- Distribute the narrative through the channels where the audience already gathers.
If the product fails two or more Wrapper Test questions, position it as a feature or workflow layer until the constraint is real. When to stop: do not publish a category claim without at least one verifiable receipt.
Step 1: The Wrapper Test
Run the 10-minute test:
- Can a technical user get 80% of the output by pasting the core prompt into a general AI model directly? (If yes, it is a wrapper.)
- Can a competitor clone the UI in a weekend? (If yes, no moat.)
- Does a frontier-model vendor shipping the feature next quarter kill the business? (If yes, no defensibility.)
If the product fails 2+ tests, it is a wrapper. That is OK — most start here.
Step 2: Find the Constraint Outside the Model
Companies that survive pick a narrow surface where they have an unfair edge:
- Vertical: regulated industries (healthcare, legal, finance)
- Infrastructure: agent memory, reward models, cold starts, latency
- Security: securing AI agents
- Workflow: owning the workflow end to end
- Data: proprietary data loops, network effects
Step 3: Design the Category Narrative
Answer these questions:
- What enemy are you against? (e.g., "AI wrappers with broken onboarding")
- What new category are you creating? (e.g., "DevEx for AI startups")
- What's the one idea that defines you? (e.g., "Developer adoption is the byproduct, not the product")
- What proof do you have? (e.g., "1,000 paid customers, $0 ad spend")
Step 4: Write the One-Liner
"I help [target customer] [do outcome] by [unique mechanism]. Proof: [metric]."
Step 5: Build the Proof Stack
Every claim needs a receipt. Typical receipts:
- Adoption → GitHub stars, package downloads, active users
- Revenue → closed proposals, customer logos (with permission)
- Authority → awards, press, original research
- Content → assets produced, publishing cadence, measured reach
Step 6: Distribute the Narrative
- X (daily): Technical insights, build logs, positioning teardowns
- LinkedIn (weekly): Long-form articles, teardown analyses, and operating notes
- YouTube: Short explainers and demos
- Website: The hub — product pages, proof, and contact
Outputs
- Category narrative document
- One-liner positioning statement
- Proof stack (receipts for every claim)
- Distribution plan (X, LinkedIn, YouTube, website)
- Competitor positioning map
Pre-flight checklist:
- Wrapper Test completed
- Constraint outside the model named
- One-liner includes a proof metric
- Every claim has a receipt
- Distribution plan names an owner and cadence
Evidence
Category design is a well-established positioning discipline: name the enemy, define a new category, and attach a proof stack so the category claim is verifiable. The Wrapper Test operationalizes the common "thin wrapper" critique of AI startups.
Examples
Before: "We use AI to help teams work better." — generic, no category, no proof. After: "We make AI agents safe to deploy in regulated healthcare workflows. 3 enterprise design partners, SOC 2 certified." — names the category, states the constraint, attaches a receipt.
The shift: move from describing capabilities to naming a category and proving you own it.
Metrics
- Category named (Y/N). Target: one category name used consistently for 90 days.
- One-liner published (Y/N). Target: live on the homepage and in the founder bio.
- Proof stack complete (Y/N). Target: 3+ receipts.
- Distribution channels active (count). Target: 2+ channels with weekly cadence.
- Inbound conversations attributed to positioning (monthly). Target: 3+ per month by day 90.
Anti-Patterns
- Listing capabilities instead of naming a category — "AI + analytics + automation" is a feature list, not a position
- Leading with company age or team size instead of the customer outcome
- Claiming "interdisciplinary" or "full-stack" without specifying the constraint you own
- Skipping the Wrapper Test — if the product is a wrapper, admit it and fix the constraint before claiming a category