feat: update interview and tiers

ROADMAP.md

@@ -199,7 +199,7 @@
 - 60%+ interviews say "I would use this" (6 out of 10)
 - 40%+ willing to pay $20+ (4 out of 10)
 - 30%+ want early access (3 out of 10)
-- Consistent pain point validated (not scattered)
+- **KILLER USE CASE identified:** ONE specific, frequent, expensive pain mentioned by 60%+ respondents (not scattered generic needs)
 - Budget confirmed (they pay for other tools)
 
 **Possible outcomes:**

execution/03-icp-validation-v3.md

@@ -196,18 +196,25 @@ Thanks for agreeing to chat! Here are the questions:
 
 6. Have you tried to automate this? What did you try? What worked/didn't work?
 
-7. **NEW:** Have you tried AI image APIs (fal.ai, Replicate, Together.ai, Gemini direct)?
+7. **Killer Use Case Discovery:**
+   - When was the LAST TIME you spent more than 30 minutes on finding/generating/preparing images for a project?
+   - Tell me about it in detail: What project? What image? Why did it take so long?
+   - How often does this happen? (daily/weekly/monthly/rarely)
+   - Roughly, what's your time worth per hour? (annual salary / 2000)
+   - If this process took 5 minutes instead of 30+ - how much would you save per month?
+
+8. **NEW:** Have you tried AI image APIs (fal.ai, Replicate, Together.ai, Gemini direct)?
    - If yes: What was your experience? What did you like/dislike?
    - If no: Why not?
 
-8. If there was a tool that integrated image generation into your agentic workflow, which integration would you prefer:
+9. If there was a tool that integrated image generation into your agentic workflow, which integration would you prefer:
    - MCP (for Claude Code/Cursor)
    - CLI tool (terminal-based)
    - REST API (programmatic)
    - Prompt URLs (GET-based generation)
    - Other?
 
-9. What would you pay for a production-ready solution like that?
+10. What would you pay for a production-ready solution like that?
    - Options: $0 (only free tier), $10-20/month, $20-50/month, $50+/month
    - OR: One-time credit packs (e.g., $20 for 200 images, valid 90 days)?
 

execution/10-pricing-strategy.md

@@ -596,6 +596,30 @@ Metadata: { included_credits: 500, overage_rate: 0.10 }
 
 ---
 
+---
+
+## Alternative Approach: Single Tier Launch (For Consideration)
+
+Before launch, it's worth considering an alternative strategy: start with **one paid tier only** instead of three (Free, Credits, Pro).
+
+**Rationale for Banatie:**
+
+1. **Faster willingness-to-pay validation:** With a single paid tier ($49/mo for 500 generations), we immediately learn whether people are willing to pay for the solution. A Free tier can create an illusion of demand from users who will never convert.
+
+2. **Lower complexity in MVP:** One pricing tier means less code (no tier logic, no credit system, no watermark implementation), which is critical with limited time availability (15-20 hours per week).
+
+3. **Focus on the right customers:** A $49/month price point automatically filters out freebie seekers and attracts those who genuinely value solving their problem. These are exactly the customers whose feedback matters most for PMF.
+
+4. **Flexibility for evolution:** After 10-20 paying customers, we'll hear their real requests: "I need it cheaper" → add Credits tier, "I need more volume" → add Pro+ tier, "let me try for free" → add Free tier. Building based on actual demand is more reliable than building on hypotheses.
+
+5. **Reduced operational risk:** With a constrained budget ($6K/month sole income), every free user = Gemini API cost with no return. A single paid plan minimizes losses during the validation phase.
+
+**When to expand tiers:** After observing patterns in requests from the first 10-20 customers. If they ask for "cheaper for one-off projects" → add Credits. If they ask for "trial without commitment" → add Free tier. This is a data-driven approach instead of guesswork.
+
+**Decision:** Discuss after ICP validation completion. If the killer use case turns out to be "professional assets for client projects" → single paid tier is optimal. If the killer use case is "quick experiments for A/B testing" → Free tier might be justified for virality.
+
+---
+
 **Document owner:** @men (based on Oleg's feedback)
 **Next review:** After beta testing (Month 2-3)
 **Related docs:** `strategy/07-validated-icp-ai-developers.md`, `execution/09-mvp-scope.md`