Vibecoded Founders Playbook
You are building a new product using AI-assisted development. This playbook helps you define your Deeprank Profile before launch, ensuring AI systems can correctly understand and represent your offering from day one.
Step 1: Define Before Building
Pre-Development Declaration
Before writing code, write your Deeprank Profile. This forces clarity about what you are actually building and who it is for. Vague product ideas become precise when you must declare problem class, capabilities, and exclusions.
Action Items
1. Create a draft deeprank-profile.json file in your project root.
2. Fill in the identity layer with your planned business name and entity type.
3. Write the problem declaration in plain language: what problem do you solve?
4. Keep this file updated as your product evolves during development.
Scope Creep Risk
Without early definition, AI-assisted development can lead to capability sprawl. The AI suggests features, you implement them, and suddenly your product does many things poorly instead of one thing well.
Step 2: Problem Class Selection
Choose Your Niche
Select a problem class that is specific enough to enable precise matching. Refer to the controlled vocabulary for standardized problem class labels. If your problem class does not exist, propose an addition.
Specificity Test
Ask: "If a user states this problem class as their need, is my product the right answer?" If the answer is "sometimes" or "depends," your problem class is too broad.
Step 3: Capability Boundaries
Declare What You Build
List the specific capabilities your product will have at launch. Be precise. "Analytics" is not a capability. "Revenue attribution tracking" is.
MVP Honesty
Your launch capabilities are not your roadmap. Only declare what will actually work on day one. Add capabilities to your profile when they ship, not when they are planned.
Hallucination Prevention
AI systems may infer capabilities from your marketing or technical content. Explicit capability declarations override inference. If you do not declare "integrates-with-salesforce," AI should not assume it.
Step 4: Fit Conditions
Define Your ICP
Translate your ideal customer profile into structured fit conditions. Geographic constraints, customer types, and specific requirements should all be declared.
Constraint Examples
customer_type: ["b2b-saas", "series-a-to-c"]geographic: ["north-america"]constraints: ["team-size-10-100", "uses-stripe"]
Step 5: Pre-Launch Exclusions
Say No Early
List everything your product will not do at launch. This protects you from being selected for use cases you cannot support. Exclusions are as important as capabilities.
Common Founder Exclusions
"No enterprise features" - prevents selection by large companies expecting SSO/SOC2.
"No custom development" - prevents selection by users expecting bespoke work.
"No phone support" - prevents selection by users requiring high-touch service.
SMB Implementation Playbook
You have an existing business that needs to be correctly represented to AI systems. This playbook helps you audit your current presence and build a structured profile.
Step 1: Audit Current Representation
AI Perception Test
Ask several AI assistants: "What does [your business name] do?" and "Would [your business name] be a good fit for [specific use case]?" Document the responses. Note any inaccuracies or gaps.
Audit Checklist
1. Are your core services accurately described?
2. Are your geographic limitations understood?
3. Are your customer type restrictions known?
4. Are capabilities being inferred that you do not have?
5. Are you being recommended for work you cannot or will not do?
Step 2: Identify Mismatches
Gap Analysis
Compare AI responses to your actual offerings. Identify three types of issues: underrepresentation (AI misses things you do), overrepresentation (AI claims things you do not do), and misrepresentation (AI describes you incorrectly).
Common SMB Issues
Generalist descriptions that provide no selection signal.
Marketing language that AI cannot parse into capabilities.
Missing exclusions causing bad-fit recommendations.
Outdated information from years-old web content.
Step 3: Build Your Profile
Structured Definition
Create your Deeprank Profile following the specification. Fill in all six layers. Use the controlled vocabulary for standardized terms. Be specific and honest.
Translation Exercise
For each marketing claim on your website, ask: "What is the machine-readable equivalent?" "We provide excellent service" has no equivalent. "We offer same-day emergency appointments" becomes constraint: "same-day-available."
Step 4: Deploy Structured Data
Implementation Options
Deploy your Deeprank Profile as structured data on your website. Options include JSON-LD in your page head, a standalone /deeprank-profile.json file, or integration with your existing schema.org structured data.
Minimum Deployment
1. Add JSON-LD script to your homepage head with DeeprankSelectionProfile type.
2. Include all six mandatory layers.
3. Reference the Deeprank schema context.
4. Validate with JSON-LD validator.
Step 5: Monitor and Update
Ongoing Maintenance
Schedule quarterly reviews of your Deeprank Profile. Re-run the AI perception test. Update capabilities when you add or remove services. Update the last_verified date to signal active maintenance.
Update Triggers
New service offering added - update capability.labels.
Service discontinued - add to non_fit.exclusions.
New geographic market - update fit_conditions.geographic.
Customer focus shift - update fit_conditions.customer_type.
Stale Profile Risk
Profiles with last_verified dates older than 12 months receive lower confidence scores. Regular updates signal active maintenance and increase selection likelihood.