🤖

BackOfficeStack: Business Panel

7 LLMs stress-tested the business concept. Enter hub password to see what they found.

That's not right. Try again.

🤖 BackOfficeStack Research

7-Model Adversarial Business Panel

Models: Claude Opus 4, o3, Gemini 2.5 Pro, Grok 4.3, GPT-4o, DeepSeek V4 Flash, OpenRouter Owl Alpha  ·  Run date: 2026-06-21

← Back to Hub
🚨

Unanimous Verdict: Do Not Invest $50k As Currently Structured

7 of 7 models
DeepSeek V4 Flash, capturing panel consensus

"morally admirable and practically doomed"

The bilingual mission is genuine. The legal and economic structure is not viable without significant changes.

All 7 models reached the same conclusion. The cooperative/worker equity vision is ahead of the market. The compliance and economics as currently structured aren't.

⚑

The 5 Things Every Model Flagged

All 7 models · all 5 issues
  1. 1
    Worker Misclassification Is a Legal Grenade A platform that sets prices, blocks low-pay bookings, controls scheduling, and owns the customer relationship IS the employer under Colorado's ABC test, regardless of what the contract says. "Owner-operator" and "co-op member" language provides no legal cover when the platform controls economic conditions. Exposure: back wages, unpaid unemployment insurance and workers' comp contributions, civil penalties up to $5,000 per worker per violation.
  2. 2
    Immigration Exposure Is an Independent Knockout Stating the mission as "maximize jobs for Mexican workers" while having a "no work-authorization collection" policy doesn't create plausible deniability, it creates constructive knowledge under IRCA. If any worker is unauthorized and ICE audits, "we didn't ask" fails when company documents establish targeting intent.
  3. 3
    The $26/hr Floor Doesn't Survive the Math To net $26/hr, a worker needs ~$35-40/hr gross after self-employment tax (~15.3%), travel time, insurance, and supply costs. A 3-hour job needs to pay the worker ~$105-120 gross. On a $200 cleaning: after Stripe (~$6) and supplies (~$20), $174 remains. Worker takes $105-120, leaving $54-69 for the platform. Google Ads customer acquisition in Boulder home services runs $60-100 per booked job. The math only works on $275-300+ jobs. Most volume is $150-200. At those prices, the platform loses money on every booking and the wage floor code would block most of those bookings anyway.
  4. 4
    Disintermediation Will Kill Retention Boulder already has Handy, TaskRabbit, Amazon Home Services, Molly Maid, Merry Maids, and 50+ local independents. Once a worker meets a homeowner through the platform, "pay me directly next time, I'll knock 15% off" is an existential churn event with no enforceable countermeasure.
  5. 5
    No Operations Experience Is a Structural Gap SEO expertise covers customer acquisition. Running a cleaning business is quality control, scheduling, complaint resolution, insurance claims, and knowing what to do when a worker no-shows or breaks a $4,000 vase. Software handles admin but doesn't replace domain knowledge.
✦

What Models Said Has Genuine Merit

Real competitive advantages
  • ✓ Annette's SEO skills are the strongest asset, called out by all 7 as a real competitive advantage in customer acquisition
  • ✓ The bilingual angle is differentiated and underserved in Boulder
  • ✓ The back-office stack concept (software as multiplier across verticals) is sound in principle
  • ✓ The cooperative/worker equity vision is ahead of the market
↻

The Pivot All Constructive Models Recommended

Consensus path forward

Hire 2 W-2 employees with proper insurance. Start as a traditional cleaning business with software as a competitive advantage, not a social-mission enforcement mechanism. Prove recurring customer acquisition first. Get the cooperative legal structure vetted by counsel as a parallel track, not a launch-day requirement.

Panel run 2026-06-21 · BackOfficeStack project · Private  ·  ← Hub