Target Audience¶

Phase: 3 — Synthesis Project: likeness Date: 2026-05-09 Confidence: Low-Medium overall. Macro / public-sentiment signals are Tier 2 academic. Creator-specific signals are inferred, not validated. This is the load-bearing discovery gap.

Honest framing¶

The founder Q&A confirmed zero structured creator interviews to date. Everything below is logically derived from public data, industry coverage, and platform analytics. It's a useful starting hypothesis. It is not validation. Phase 8 should weight creator-discovery experiments very heavily.

Primary Persona — Mid-tier verified adult creator with existing audience¶

Profile: - Age 22-40, US-based at MVP - Active on OnlyFans / Fansly / direct-platforms; possibly cross-posts to multiple - Existing fan audience: 1K-50K active subscribers - Monthly gross revenue: $5K-$80K (top 1-10% of OnlyFans creators band) - Tech-comfortable but not necessarily an AI specialist - Existing experience with AI on creator platforms is mixed — some use AI chatbots for fan responses; some have been deepfaked without consent and have no recourse

Representative composite quote (synthesized, NOT actual):

"I already saw stuff with my face on Civitai that I didn't make. The platforms that 'protect' me are the ones that ban me. If you can show me a way to actually license my likeness to my own fans on my own terms, and the money's real, I'd talk."

[Critical caveat]: This quote is illustrative. It needs to be replaced with actual quotes from real creators in the validation phase.

Secondary Persona — Engaged adult creator subscriber¶

Profile: - Existing OnlyFans / Fansly subscriber across 2-5 creators - Pays $20-$200/month subscriptions + $50-$500/month in PPV / customs / tips for top creators - Some willingness to pay for "extras" — has paid for custom videos, 1:1 messages - Parasocial engagement is a real driver — fans are invested in specific creators, not generic content - May or may not be an existing AI girlfriend / generator user

Representative composite quote (synthesized, NOT actual):

"I'd pay for AI of [specific creator] over generic AI characters because I actually care about her. The Candy.ai stuff feels hollow."

[Critical caveat]: Same as above.

Anti-Persona — Who NOT to target¶

Fans seeking deepfakes of celebrities or non-consenting third parties. Hard-blocked by design. The platform's value proposition only works against the consenting-creators audience.
Creators looking to sell their likeness for unrestricted use. Likeness's revocation premise is its differentiator. Creators who want a "sell once, license away" model are not Likeness's customer.
Creators with audiences that don't translate to AI engagement. A creator whose fans are paying for raw real content may not convert to AI even if the creator opts in. Test in concierge.
Mass-market AI girlfriend users seeking generic synthetic personas. Different category. Not addressable.

Customer Pain Hierarchy (synthesized — needs validation)¶

Ranked plausible importance, NOT validated:

#	Pain	Evidence	Confidence
1	Unauthorized AI use of creator likeness with no recourse, no revenue	MIT Tech Review on Civitai; OnlyFans 12K+ deepfake bans 2024-2025; 90% of deepfake bounties target women	High (data-backed)
2	Revenue ceiling on real content; AI is potential additive line	OnlyFans 2024 creator median $131/mo with 4.6M creators; AI as 15% of Fanvue revenue	High
3	Platform risk (Mastercard/Visa rule changes; account bans)	Fansly photoreal AI ban June 2025; OnlyFans deepfake bans; Civitai processor cutoff	High
4	Trust erosion with fans if AI is used wrong	Quoted creator concern in Yahoo / industry coverage	Medium
5	Compliance overhead (2257, ID verification, takedowns)	Industry-standard knowledge	High
6	Time-cost of producing real content; burnout signals	Implicit in creator economy reports	Medium

The interesting hypothesis to test: is #1 (defense) more compelling than #2 (offense) for actual creator decision-making? The founder brief frames this primarily as a #2 opportunity. Variation B in 00-intake/brainstorm.md proposes leading with #1. This is one of the highest-value questions for creator discovery interviews.

Jobs-to-be-Done Framework¶

Functional jobs¶

"Help me earn money from AI of my likeness without learning to train models myself."
"Help me defend against unauthorized AI use of my likeness."
"Give me workflow tools (approval queue, license configuration) that respect my time."
"Make sure my fans can actually pay (no processor surprises that cut off my revenue overnight)."

"Let me show my fans I'm in control of my own image."
"Let me participate in the AI economy without being seen as either a victim or a sellout."
"Don't make me look naive or out-of-touch to my industry peers."

Emotional jobs¶

"Make me feel like the platform is on my side, not extracting from me."
"Make me feel like the platform respects the work I already do, not like I'm a beta tester."
"Reduce the anxiety of unauthorized AI being out there about me."

Language Map¶

Words and phrases creators are likely to respond well to (or recoil from), based on industry coverage and platform-side communication patterns. NOT validated by creator interviews — replace with actual creator language post-discovery.

Likely positive¶

"Your likeness, your rules"
"Revocable license"
"You're in charge of every decision"
"We don't keep your model — it stays on the platform, period"
"Worker-controlled"
"You can shut it off anytime"

Likely negative / red-flag¶

"Deepfake" (carries victim/perpetrator framing; most creators react against the word)
"AI twin" / "AI clone" (cute branding that minimizes the worker-control issue)
"Passive income" (suggests platform extraction, dismisses the actual work)
"Synthetic" (often confused with non-consenting fabrication)
"Train your model" (sounds technical and creator-as-product)

Likely neutral¶

"AI generation"
"License"
"Consent"
"Approval queue"

Buying Behavior¶

Decision process: - Individual decision (creator chooses their own platform stack) - Agency or manager input for top-tier creators (10%+) - Trust is the gating factor — references from peers matter more than features

Decision criteria (hypothesized): 1. Does this platform respect me? 2. Does the money work? (volume × take rate × processor stability) 3. Do my fans actually use this? 4. What's the legal exposure to me as the creator? 5. Can I revoke / leave / take my data with me?

Cycle length: - Initial conversation to first concierge participation: weeks to months - Concierge participation to "this is my main AI revenue platform": 3-6 months minimum - Adoption is a creator-by-creator earned trust process; no shortcut at MVP scope

Common objections (anticipated): - "How do I know my model won't leak?" — answered by architecture commitments + audit logs - "What happens if I want to leave?" — answered by revocation premise + no-export - "What if my fans complain or don't pay?" — answered by per-generation pricing + creator-set rules - "What's your real take rate including processor fees?" — must be answered honestly with full margin breakdown

Where to Reach Them¶

Channel	Priority	Cost / time
Founder + Creator Ops cofounder 1:1 outreach	Primary	Time-intensive
Adult industry events (XBIZ, AVN, EXXXOTICA)	High	$5-20K/event
Existing Discord / Telegram creator communities	High	Time-intensive
Talent agency partnerships	Medium	Year 2
Press / earned media	Medium for credibility, low for direct creator pipeline	Time

The Creator Ops cofounder hire is the single biggest determinant of creator-acquisition velocity.

Demand Validation¶

Direct demand for creator-side AI on legitimate creator platforms: - 15% of Fanvue revenue is AI-generated [Sacra, Tier 2] — strongest single demand signal. - 12,000+ OnlyFans deepfake violation account bans 2024-2025 [Tier 2/3] — implicit demand for legitimate alternative. - Adult AI image generation ~$2.8B in 2025 [Tier 3, directional] — fans demonstrably pay for AI adult content. - Public sentiment supports consent-first framing (47.5% negative on adult deepfake abuse) [Tier 2 academic].

Demand validation score: Strong on the macro / category level. Weak on the Likeness-specific WTP at the proposed pricing tiers.

Data Gaps¶

The entire creator-voice section needs primary research. Phase 8 priority #1.
Specific WTP data for Likeness's tier structure ($15-$200/mo range) is unvalidated.
Pain hierarchy ordering #1 vs #2 is the most strategically important unknown.
Anti-persona boundaries are theoretical — could there be edge-case creators we don't expect, who would benefit and we'd want to onboard? Discovery would surface this.

Strategic Connections¶

The pain-hierarchy ordering question (#1 vs #2) connects directly to 00-intake/brainstorm.md Variation B — which sequenced takedown-first, monetization-second. Creator interviews will tell us whether Variation A or a hybrid is correct.
The "pricing tier WTP unvalidated" gap connects to 05-financial/revenue-model.md — the financial model has to acknowledge it's resting on inferred preferences.
The Vylit competitive signal in competitor-landscape.md raises pressure on the speed of creator interviews; discovery is competitively urgent, not just product-urgent.

Flags¶

Red Flags: - No primary creator data exists. Every creator-related claim is inferred. Validation Phase 8 must front-load creator interviews above all other experiments.

Yellow Flags: - The composite quotes in this document are illustrative only — replace with real creator quotes after the first 5-10 interviews. - Pain hierarchy ranking may be inverted in practice. Test this explicitly in interviews.

Sources¶

See 01-discovery/raw/customers-demand.md for full source list and per-claim citations.