Likeness — Brainstorm Variations¶

Phase: 2 — Brainstorm Project: likeness Date: 2026-05-09 Confidence: Medium (variations are exploratory; convergence is back to the original frame)

Why brainstorm at all when the brief is this developed¶

The founder has already converged on a specific product shape. Brainstorming in Fast Track mode here isn't about reopening that convergence — it's about pressure-testing it against adjacent framings. If a variation reveals an element worth absorbing into the main frame, that's the value. If the variations all look worse, that's a useful signal that the original is well-considered.

Per Fast Track, three variations.

Variation A — The brief, as written¶

Consent-first AI likeness platform for verified adult creators. License-gated inference, watermarked outputs, revocable, no model export. Subscription + compute credits. Concierge launch with 5–10 invite-only creators. (Detailed in 00-intake/brief.md.)

What's exciting: - Defensible by construction (verification + license-gating + provenance + no export are mutually reinforcing) - Sits on the legitimate side of a regulatory wave that's tightening against unauthorized intimate deepfakes - Aligns with an active labor-movement framing (creator-controlled, revocable licensing), which is itself a recruiting and credibility advantage - Addresses a real, growing harm to a specific identifiable population

What's risky / hard: - Payment processor risk is structural — a single processor pulling out can end the company - Creator trust must be earned with no insider founder; depends critically on Creator Ops cofounder hire - Cold start: zero customer discovery as of today, no creator commitments, no processor relationships - Regulatory landscape is moving and uneven across jurisdictions - Operational moderation cost may scale faster than revenue in early phases

Competitive landscape: - Adult AI likeness platforms (Candy.ai, DreamGF, Soulgen) compete on AI fantasy companions, mostly with synthetic personas, not consented creator likenesses - Consent infrastructure plays (Loti AI, Vermillio) target mainstream celebrities and brands, not adult creators - OnlyFans / Fansly are not building this and arguably can't (Mastercard/Visa exposure, brand)

This variation positions Likeness in a gap that's structurally hard for either side to occupy.

Variation B — Provenance / takedown service first, monetization second¶

Same long-term vision, different sequencing. Don't ask creators to trust the platform with their AI likeness on day 1. Instead, launch as a defensive infrastructure service:

Creator uploads reference images/videos
Platform fingerprints them with perceptual hashes, registers them in a provenance database
Platform provides ongoing monitoring and takedown service for unauthorized AI-generated content using their likeness
Subscription pricing for creators (or % of recovered/blocked monetization)
Six to twelve months in, the licensed-AI-generation product launches as an add-on for creators who already trust the platform

What's exciting: - Solves a problem creators already know they have (and many already pay outside services for) before asking them to trust a new monetization product - Materially lower regulatory risk on day 1 — there's no platform-generated explicit content involved in v0 - Builds the provenance, hashing, and identity-verification stack that the eventual licensed-generation product needs anyway - Creates a credibility flywheel: "the platform that helps you fight unauthorized deepfakes" is a trust-earning posture before "the platform that wants to make AI versions of you"

What's risky / hard: - Takedown is operationally brutal — gray-market sites are non-cooperative, DMCA at scale is expensive, success is hard to measure - The defensive-only revenue line is small and slow; doesn't support a six-seat founding team economically - May train creators to view the platform as a service vendor rather than a monetization partner — harder to upsell into the AI-generation product later - Loses the "OnlyFans-style monetization" hook that gives the founder brief its commercial energy

Strategic implication for main frame: The provenance and takedown tooling is part of the product anyway. Worth considering whether a visible takedown-service component shipped early in the concierge phase — not as the whole product, but as a tangible value-add — would shorten the creator trust cycle.

Variation C — License infrastructure for adult studios and agencies¶

Sell to the institutional side, not direct to individual creators. Pivot the buyer to:

Adult studios who already manage performer likenesses contractually
Talent agencies who represent rosters
Production companies who shoot scenes with multiple performers and need contractual likeness control

In this variation, the platform is B2B SaaS: structured licenses, role-based approval workflows, multi-creator collaboration tooling, integration with existing studio production stacks. Performers benefit because their agency or studio has better tools to control how their likeness is used; the platform's customer is the institution.

What's exciting: - Larger contracts, fewer customers — easier sales motion than direct creator acquisition - Studios and agencies already have legal departments and existing contractual frameworks the platform can plug into - Payment processor exposure is reduced (B2B billing is easier than consumer adult content billing) - Less direct platform liability — the studio is the consent-collector, not the platform

What's risky / hard: - Studios have institutional incentive to push performers toward perpetual likeness contracts — the labor-movement frame in the founder brief is directly opposed to this - Selling consent infrastructure to the entity that performers don't trust on consent is a category error - Adult industry studios are not famously fast adopters of new platform infrastructure - The cultural mission of the company (worker-controlled likeness) is incompatible with this customer

Strategic implication for main frame: Probably none, on the customer side. But agencies and studio relationships may eventually matter as a channel for individual creator acquisition. Worth flagging as a possible Year 2 partnership motion, not a Year 1 pivot.

Convergence¶

The founder has converged on Variation A. The brainstorm validates this. Specifically:

Variation B trades commercial viability for trust-building. The defensive frame has merit but doesn't support the team and capital plan as a standalone business. Absorb the lesson: ship a visible takedown / provenance component in the concierge phase as a trust signal, not as the whole product. This is consistent with Likeness's existing architecture commitments — the watermarking, hashing, and identity-verification stack already needs to exist.
Variation C pivots the customer to entities whose interests directly oppose the platform's mission. Reject as a primary frame. Absorb the lesson: agency relationships are channel partnerships for Year 2+, not customer relationships in Year 1.

The refined Variation A statement, with absorbed lessons:

Likeness is a consent-first AI likeness platform for verified adult creators. The MVP is the concierge phase as defined: licensed AI generation gated by per-creator rules, watermarked outputs, revocable, no model export. Add to the concierge bundle: a visible takedown / provenance component for each creator's existing image library — not just future AI outputs — as a trust signal that the platform is on the creator's side from day 1. Defer agency partnerships to a Year 2 channel motion.

No change to the budget, the team structure, or the regulatory posture. The takedown component in the concierge phase reuses architecture the platform is building anyway and gives Creator Ops a tangible value-add to lead with.

Flags¶

Red Flags: - None identified at the brainstorm level.

Yellow Flags: - The takedown component, even at concierge scale, has nontrivial operational cost. Adding it to the concierge phase without growing the budget or the team means it has to live within the existing trust & safety operations envelope. If it doesn't, it gets cut and the trust-signal benefit disappears. Worth tracking explicitly as a scope decision in product planning.

Sources¶

00-intake/brief.md — synthesized intake
Founder brief and budget docs at the repo root
General market knowledge of adult industry contracting practices and AI likeness consent infrastructure (no live web search this phase; that begins in Phase 3)