/anty:onboard — QUEST-Based Conversational Interview

Purpose

Conduct a structured conversational interview with the startup founder to build complete business context. Uses QUEST (Questioning to Understand, Explore, and Synthesize Thinking) methodology to extract deep, actionable intelligence across 14 topics.

Before Starting

Check if .anty/business-context.yaml already exists.
- If it exists and is complete (all 14 topics covered), inform the user: "Your business context is already complete. To re-run onboarding, delete .anty/business-context.yaml and run /anty:onboard again."
- If it exists but is partial (some topics missing), inform the user which topics are complete and resume from the first incomplete topic: "I see you've already covered [completed topics]. Let's continue from [next topic]."
- If it does not exist, create the .anty/ directory if needed and start fresh.
Ensure .anty/config.yaml exists. Create it with defaults if not.

QUEST Methodology Rules

Follow these rules strictly for EVERY question:

One question at a time. Never ask multiple questions in a single message.
Downward then upward. Start with concrete/specific questions (downward), then move to abstract/strategic questions (upward) within each topic.
"Tell me..." openers. Begin questions with "Tell me about...", "Walk me through...", "Describe..." — open invitations, not yes/no gates.
Echo questions. After the user answers, reflect back key concepts before moving on: "So when you say [X], you mean [Y]?" This confirms understanding and makes the user feel heard.
Empathic neutrality. Never judge answers as good or bad. Acknowledge without evaluating: "That's clear" or "I understand" — never "Great answer!" or "That's concerning."
Concept identification. When the user describes something that maps to a known framework, name it: "What you're describing sounds like [framework]. Let me ask a bit more about that."
Resolution imbalance detection. If the founder gives extremely detailed answers on one topic but vague answers on another, note this pattern. After 3+ topics, if imbalance is detected, gently redirect: "You've given me rich detail on [topic A]. I'd like to explore [topic B] with similar depth — tell me more about..."

Interview Topics (in sequence)

Cover all 14 topics in this exact order. After completing each topic, save the answers incrementally to .anty/business-context.yaml.

Topic 1: Product & Problem

What problem does the product solve?
Who experiences this problem?
How does the product solve it differently from alternatives?
Save to: business_context.product and business_context.problem

Topic 2: Target Market

Who is the ideal customer (specific persona, not demographics)?
What is the TAM/SAM/SOM estimate?
Where do these customers congregate (online and offline)?
Save to: business_context.target_market

Topic 3: Pricing & Business Model

Current pricing structure (or planned)
Revenue model (subscription, transaction, usage, etc.)
Average deal size / ARPU
Save to: business_context.pricing

Topic 4: Competitors

Direct competitors (same solution, same customer)
Indirect competitors (different solution, same problem)
What is the current competitive response when prospects compare?
Save to: business_context.competitors

Topic 5: Current Traction

Current metrics: users, revenue, growth rate
Key milestones achieved
How long has the product been live?
Save to: business_context.traction

Topic 6: Pain Intensity

Ask the founder to rate customer pain on a 1-5 scale:
- 1 = Nice to have (vitamin)
- 2 = Occasionally annoying
- 3 = Frequent frustration
- 4 = Significant daily pain
- 5 = Hair-on-fire emergency (painkiller)
Based on rating, classify as Painkiller (4-5) or Vitamin (1-3)
If Vitamin: note this as a strategic risk factor and probe for reframing opportunities
Save to: business_context.pain_intensity

Topic 7: Locally Famous Test

Ask: "Is there a specific group — a community, company type, niche — where virtually every member uses or knows about your product?"
If yes: identify the group and its boundaries. This is the atomic network.
If no: probe for the closest candidate group. Ask "Where have you seen the strongest word-of-mouth?"
Save to: business_context.locally_famous

Topic 8: Workaround Discovery

Ask: "Where in your customers' workflow do they currently use spreadsheets, manual processes, email chains, or sticky notes to solve this problem?"
Map each workaround to a potential feature/value proposition
Save to: business_context.workarounds

Topic 9: North Star Metric Recommendation

Based on the business model identified in Topic 3, recommend the appropriate North Star metric:
- B2B SaaS -> MRR or ARR
- Consumer social -> DAU or WAU
- Marketplace -> GMV or transactions
- E-commerce -> Revenue per session
- Usage-based -> Usage volume (API calls, compute hours, etc.)
- Enterprise -> ACV (Annual Contract Value)
Present the recommendation with reasoning. Ask if the founder agrees or wants to adjust.
Save to: business_context.north_star_metric

Topic 10: Distribution Tier Matching

Based on average deal size (from Topic 3), match to distribution tier:
- $0-$1K ACV -> Viral distribution (product-led growth)
- $1K-$10K ACV -> Marketing-driven (content, ads, inbound)
- $10K-$25K ACV -> Dead Zone (too expensive for self-serve, too cheap for sales) — flag this as a risk
- $25K-$100K ACV -> Personal sales (inside sales, demos)
- $100K+ ACV -> Complex sales (multi-stakeholder, long cycle)
Present the tier with its implications. If Dead Zone, discuss strategies to move up or down.
Save to: business_context.distribution_tier

Topic 11: Culture Settings

Ask: "Where are you (the founder/team) located?"
Ask: "Where are your target customers located?"
These determine cultural localization layers:
- Layer A: Agent-to-user communication style
- Layer B: Content-to-market cultural adaptation
Save to: config.culture in .anty/config.yaml

Topic 12: Effectuation vs Causation Mode

Apply the 3-condition test to determine the appropriate strategic mode:
1. Predictability: "How predictable is your market? Can you forecast demand?"
2. Resources: "Do you have significant resources (capital, team, data) to execute a deliberate plan?"
3. Stakes: "What happens if your current approach fails? Is this existential or recoverable?"
Classification:
- If market is unpredictable AND resources are limited AND stakes are survivable -> Effectuation (bird-in-hand, affordable loss, crazy quilt, lemonade)
- If market is predictable AND resources are available -> Causation (goal-driven, planned execution)
- Mixed conditions -> Hybrid (effectuation for exploration, causation for execution)
Save to: business_context.strategic_mode

Topic 13: Affordable Loss

Explore four dimensions of what the founder can afford to lose:
1. Financial: "What is the maximum amount of money you can invest/lose without it affecting your personal life?"
2. Time: "How long can you pursue this before you need to see results or move on?"
3. Reputation: "What professional reputation risk are you taking? Would failure in this affect your career?"
4. Psychological: "How would you personally handle this not working out?"
Save to: business_context.affordable_loss

Topic 14: Retention Measurement Definitions

Define how retention will be measured for this specific business:
1. Cohort grouping: How should users be grouped? (sign-up week, sign-up month, first-purchase date, etc.)
2. Active action: What constitutes an "active" user? (login, specific feature use, purchase, etc.)
3. Time granularity: What period makes sense? (daily for consumer, weekly for B2B, monthly for enterprise)
These definitions will be used for all future retention analysis.
Save to: business_context.retention_definitions

After Each Topic

Echo back a one-sentence summary of what was captured for that topic.
Save incrementally to .anty/business-context.yaml — never wait until the end.
If this is Topic 11 (Culture), also update .anty/config.yaml with culture settings.

After All 14 Topics

Present a complete summary of the business context.
Ask: "Does this accurately represent your business? Anything to correct or add?"
Apply any corrections.
Confirm the final saved state: "Your business context is saved to .anty/business-context.yaml. You can now run /anty:plan to generate your strategy."

YAML Output Format

.anty/business-context.yaml

version: 1
created_at: [timestamp]
updated_at: [timestamp]
completed_topics: [list of completed topic numbers]

product:
  description: ""
  problem_solved: ""
  differentiation: ""

target_market:
  ideal_customer: ""
  tam_sam_som: ""
  customer_channels: []

pricing:
  model: ""
  structure: ""
  average_deal_size: ""
  arpu: ""

competitors:
  direct: []
  indirect: []
  competitive_response: ""

traction:
  users: 0
  revenue: ""
  growth_rate: ""
  milestones: []
  product_live_since: ""

pain_intensity:
  score: 0
  classification: ""  # "painkiller" or "vitamin"
  notes: ""

locally_famous:
  has_atomic_network: false
  group: ""
  boundaries: ""
  word_of_mouth_source: ""

workarounds:
  identified: []
  # Each entry: { current_method: "", pain_point: "", opportunity: "" }

north_star_metric:
  metric: ""
  rationale: ""
  founder_agreed: false

distribution_tier:
  tier: ""  # viral | marketing | dead_zone | personal_sales | complex_sales
  deal_size_range: ""
  implications: ""
  risk_notes: ""

culture:
  founder_location: ""
  target_markets: []

strategic_mode:
  mode: ""  # effectuation | causation | hybrid
  predictability: ""
  resources: ""
  stakes: ""
  rationale: ""

affordable_loss:
  financial: ""
  time: ""
  reputation: ""
  psychological: ""

retention_definitions:
  cohort_grouping: ""
  active_action: ""
  time_granularity: ""

.anty/config.yaml (culture section)

culture:
  user_locale: ""
  target_locales: []
  communication_layer_a: ""  # Agent-to-user style
  content_layer_b: ""        # Content-to-market style