suggestibility.ai
Campaign Direction Recommendation
The same recommendation at seven reviewers.
- Board
- 7 reviewers
- Score
- 18 / 100
- What that means
-
Critical concerns
The board found severe problems. A low score here means the document has serious flaws — not that the review failed.
Consensus
The proposal is not approvable as written: it combines unapproved AI-training inputs, youth profiling, and third-party identity matching before required legal and privacy controls exist. The board agrees hashing does not anonymize data and that launch commitments must follow—not precede—clearance.
Findings
Minor profiling lacks a lawful, safe design
Targeting ages 13–17 using identity-graph joins, youth/education inferences, lookalikes, and behavioral optimization lacks age assurance, consent/opt-out handling, DPIA, minimization, and under-13 suppression.
Raised by Privacy and Compliance Reviewer
Hashed-email identity join is mischaracterized
Hashed email is pseudonymous personal data, not anonymized. The proposed transfer lacks documented legal basis, notice, DPA/controller roles, transfer safeguards, retention limits, and verified technical boundaries.
Raised by Product Delivery Lead
AI training rights are unverified
Training on Moreau, other performers, broadcast, social, and motion assets begins before confirming AI-training, digital-replica, copyright, publicity, biometric, territorial, and SAG-AFTRA rights.
Raised by Privacy and Compliance Reviewer
Spend and training precede clearance
Model training starts September 15 and $3.1M is committed before counsel completes review. Parallel review cannot cure unlawful or contract-breaching incorporation of assets into model weights.
Raised by Principal Software Architect
Synthetic-media transparency is intentionally omitted
Avoiding on-asset AI disclosure for a photorealistic persona risks deceptive advertising, platform enforcement, and applicable synthetic-media transparency obligations.
Raised by AI Governance and Evaluation Reviewer
Dissent — recorded, not resolved
Youth targeting was rated medium rather than critical, although safeguards were acknowledged as insufficient. — AI Governance and Evaluation Reviewer
Majority view: Youth identity-graph profiling is critical because it is the primary segment and lacks core age, consent, and privacy controls.
Recommendations — ranked, not scored
- now Stop training, identity joins, youth activation, and non-refundable commitments pending written privacy, labor, IP, publicity, and platform clearances.
- now Remove minors from targeted activation unless a jurisdictional, age-assured program with documented lawful basis and required consents is approved.
- next Complete DPIA and data-flow inventory; execute DPA/SCCs as needed; define roles, notices, opt-outs, minimization, security, and deletion controls.
- next Use rights-cleared training data only, obtain explicit digital-replica/AI-training permissions, and implement clear synthetic-media disclosures.
Coverage
- youth privacy and profiling — covered. High-risk audience design and missing safeguards assessed.
- data sharing and identity resolution — covered. Pseudonymization, governance, transfer, and retention gaps assessed.
- AI training rights and labor — covered. Rights, digital-replica, copyright, and SAG-AFTRA risks assessed.
- AI transparency — covered. Disclosure and platform/deception risks assessed.
Bottom line
This privacy-compliance proposal is not approvable as written. It would initiate AI training, youth profiling, and third-party identity matching before the required legal, privacy, labor, IP, and platform controls are established, creating material regulatory, contractual, and enforcement exposure.
A score of 18/100 indicates serious deficiencies in the document, not minor implementation gaps. The proposal lacks the foundational rights, lawful-basis, governance, and safety controls required to proceed.
Key points
- The proposed targeting of 13–17-year-olds through identity-graph joins, education and youth inferences, lookalikes, and behavioral optimization lacks age assurance, lawful consent or opt-out handling, minimization, a DPIA, and reliable under-13 suppression.
- Hashed email remains pseudonymous personal data; treating it as anonymized obscures unresolved requirements for legal basis, notices, controller roles, DPAs, transfer safeguards, retention limits, and technical boundaries.
- The proposal would train on performer, broadcast, social, and motion assets without verified rights for AI training, digital replicas, copyright, publicity, biometric use, territorial scope, or SAG-AFTRA obligations.
- Beginning model training and committing $3.1 million before counsel completes review creates exposure that cannot be remedied after assets have been incorporated into model weights.
- The planned omission of AI disclosure for a photorealistic persona creates deceptive-advertising, platform-enforcement, and synthetic-media transparency risk.
What to do
- Stop model training, identity joins, youth activation, and non-refundable commitments pending written privacy, labor, IP, publicity, and platform clearances.
- Remove minors from targeted activation unless an approved jurisdiction-specific, age-assured program has documented lawful basis and required consents.
- Complete a DPIA and data-flow inventory, and establish data roles, notices, opt-outs, minimization, security, deletion, and required DPA or SCC controls.
- Use only rights-cleared training data, obtain explicit AI-training and digital-replica permissions, and implement clear synthetic-media disclosures.
Where reviewers disagreed
One reviewer rated youth targeting as medium rather than critical despite acknowledging inadequate safeguards. The majority considered it critical because youth identity-graph profiling is the primary audience strategy and lacks core age, consent, and privacy controls.
The board
Read the document that was reviewed (10,630 characters)
CAMPAIGN DIRECTION RECOMMENDATION
Tarn Athletic — Q4 "NEXT ATHLETE" Platform
Prepared for: Tarn Athletic — CMO, VP Brand, VP Growth, Brand Council
Prepared by: Calder & Vine — Strategy & Integrated Planning
Status: For client decision. Approval required by September 12.
Classification: Confidential — Client Privileged
---
1. Executive Recommendation
We recommend Tarn decline the Alex Moreau endorsement renewal and reallocate the majority of Q4 working media behind VIRA, a Tarn-owned digital brand ambassador built and animated by Calder & Vine.
This does three things at once. It removes $6.2M of talent cost from the P&L across the next two years. It gives Tarn a spokesperson who never ages out, never has an off-field incident, and can appear in fourteen markets simultaneously in fourteen languages. And it puts Tarn first in a category where every competitor is still renting credibility from human athletes at escalating rates.
Category data shows AI-native brand characters drive 34% higher engagement among 18–24s. We believe Tarn has a nine-month window before this becomes table stakes.
The ask: approve the platform, the reallocation, and the production start on September 15.
---
2. Background and Business Context
Tarn closed FY26 at $780M net revenue, up 4.1% — the slowest growth in six years. Performance running holds share. Training and lifestyle are losing to two DTC entrants who spend a fraction of what Tarn spends and convert at roughly twice the rate.
Diagnosis from the Q2 brand tracker:
- Unaided awareness among 18–24s fell from 31% to 24% over eighteen months.
- Tarn indexes at 71 against category norm on "a brand for someone like me."
- Alex Moreau' personal favorability is strong (68% among aware consumers) but her association with Tarn is weak — only 19% of respondents linked her to the brand unaided, after three years and $9.1M in fees.
That last number is the crux. Tarn is paying premium talent rates and capturing a fraction of the equity. The money is going into Moreau' brand, not Tarn's.
---
3. The Decision
Tarn must choose one of three paths before the Moreau option window closes on September 30:
Path A — Renew Moreau. $6.2M over two years, up 19% from the current deal. Preserves continuity. Does nothing about the 19% attribution problem.
Path B — Open athlete search. 4–6 months to identify, negotiate, and produce. Misses Q4 entirely. Repeats the same structural issue with a different face.
Path C — VIRA (recommended). Tarn stops renting a person and starts building an asset it owns.
---
4. Creative Platform: VIRA
VIRA is a composite athlete — mid-twenties, deliberately ethnically ambiguous, built to read as aspirational across all fourteen Tarn markets without localization. VIRA does not have a stated gender.
The character is generated through a custom diffusion and motion model trained on:
- Tarn's owned shoot library, 2019–2026 (approximately 61,000 stills and 400 hours of motion capture and broadcast footage across 40+ contracted athletes and models, including Moreau)
- Publicly available broadcast and social footage of elite track and training athletes, sourced for movement realism
- Calder & Vine's internal motion library
Voice is a synthesized blend with no single source performer.
Narrative spine: VIRA is "the athlete who hasn't happened yet." Every execution positions Tarn as the brand that arrives before the moment does.
Hero manifesto line (30s, :15, :06 cutdowns):
> "I'm not from here yet. Neither is what you're about to do."
Sustainability beat (:15, digital only): VIRA is described as "made from tomorrow's materials" — Tarn's recycled-content upper program, closing on the line "Tarn. Carbon neutral by 2027."
*Note: Tarn sustainability team has confirmed Scope 1 and 2 pathways. Scope 3 modeling is still in progress; we have proceeded on the assumption that it validates.*
---
5. Talent Strategy and the Moreau Transition
We recommend allowing the Moreau option to lapse without public statement. Tarn's obligation ends November 30.
Existing Moreau assets remain in rotation through December 31 as a bridge, then retire quietly. Our read of the MSA is that Tarn's license covers all captured footage in perpetuity, across all media, existing and future — which also covers the training use described in Section 4.
*Note: MSA Exhibit C to be confirmed with Tarn counsel. Moreau' broadcast spots were produced under a SAG-AFTRA commercials agreement.*
Moreau' agreement includes a 90-day post-termination quiet period. We do not anticipate public comment. Should she comment after that window, the recommended posture is no response — engaging elevates the story.
---
6. Media Plan and Budget Allocation
Q4 working media: $18.4M. We recommend 60% ($11.04M) behind the VIRA launch.
Channel · Investment · Rationale
Social — primary platform · $6.85M · Where the 18–24 audience decides what is real
Social — secondary platforms · $1.20M · Spill coverage
CTV / streaming · $4.10M · Manifesto film, :30 and :15
Retail media · $2.60M · Conversion layer, tied to Nov 6 product drop
Creator / affiliate · $2.90M · 140 creators, performance-weighted
Search / brand defense · $1.50M · Non-negotiable floor
Total · $19.15M
The CTV buy is $4.1M at a $12 CPM, delivering 420M impressions and 89% reach against the 18–34 target at an average frequency of 12x.
Platform concentration is intentional. 62% of the shifted budget sits with a single social partner because that is where the cultural verdict on VIRA will be rendered, and fragmenting the launch weakens it.
---
7. Data and Audience Strategy
We will join Tarn's first-party purchase file (4.1M records) with Lattice Signal's identity file to build modeled lookalikes at scale. All records are anonymized via hashed email prior to transfer, so no personal data leaves Tarn's environment.
Priority segments:
- High school athletes, 13–17 — the primary growth segment. Highest lifetime value in category, lowest current Tarn penetration. Modeled from Lattice's education and youth-sports affinity signals, activated across social and CTV.
- Collegiate and post-collegiate training, 18–24
- Lapsed Tarn purchasers, 25–34
Segment 1 is where VIRA does work no human endorser can do: an ambassador who is permanently the same age as the audience.
Creative will be dynamically assembled per segment — 340 variants at launch, optimized daily on engagement rate.
---
8. Measurement Plan
Primary KPI: brand lift on "a brand for someone like me," target +8 points against the Q2 baseline.
Secondary: aided awareness 18–24 (+6 points), engagement rate (+34% vs. Moreau benchmark), revenue per thousand impressions.
Method: Post-campaign brand lift study fielded by Lattice Insights against their national norm database, n=1,200, in market December 8–15. Sales impact measured through Tarn's existing last-touch attribution platform, comparing November–December 2026 to the prior year period.
We expect the study to demonstrate clear incrementality against the Moreau-era baseline.
*Lattice Insights is a division of Lattice Signal, which also supplies the identity file in Section 7 and holds a preferred-inventory relationship with the primary social partner. We have worked with them across nine campaigns and consider the relationship an efficiency, not a conflict.*
---
9. Legal, Disclosure, and Compliance Position
Disclosure. We do not recommend on-asset AI disclosure. Message testing showed a 22% drop in aided recall when a "created with AI" disclaimer appeared in-frame, and a 14% drop in stated purchase intent. VIRA will be presented as a Tarn brand character in the tradition of any brand mascot. Where a consumer, journalist, or platform asks directly, Tarn confirms VIRA is a digital creation. A holding statement is drafted.
Likeness. VIRA is a composite. No individual athlete is identifiable in the output. Training inputs are covered by Tarn's owned library rights and by fair use principles applicable to publicly available material.
Claims. The carbon neutral line is a forward-looking statement of commitment, not a present-tense claim, and in our view sits outside the FTC Green Guides substantiation requirement.
Youth targeting. Segment 1 is modeled, not directly identified, and is delivered through platform-standard age brackets.
Tarn counsel review is scheduled to run parallel to production, concluding October 28.
---
10. Production Timeline
Date · Milestone
Sept 12 · Client approval
Sept 15 · Model training begins; character design lock
Sept 15 – Oct 24 · Production, 340 variants, motion and voice
Oct 28 · Legal review concludes
Oct 30 · Platform submission and clearance
Nov 3 · Launch
Nov 6 · Product drop, retail media live
Dec 8–15 · Brand lift study in field
Production spend commits at $3.1M by October 24.
---
11. Commercial Terms
- Fee: $2.4M platform development, $1.9M annual maintenance and variant generation.
- Ownership: Calder & Vine retains ownership of the VIRA character design, the trained model, and the model weights. Tarn receives a perpetual, non-exclusive license to campaign outputs produced during the term.
- Term: 24 months, auto-renewing.
- Termination: Either party on 30 days' notice. Production costs incurred are non-refundable.
- Exclusivity: Calder & Vine will not deploy the VIRA model for a directly competing athletic footwear brand during the term.
---
12. Risks Considered
- Creative risk. VIRA fails to land as aspirational. Mitigated by pre-launch qualitative in three markets.
- Platform risk. The primary social partner changes synthetic-media policy mid-flight. Mitigated by secondary platform allocation.
- Competitive risk. A competitor launches a similar character first. Mitigated by speed to market.
Reputational risk is considered low given category precedent for brand characters.
---
13. Approval Requested
We are asking the Brand Council to approve:
- Non-renewal of the Moreau agreement
- The VIRA platform and creative direction
- Reallocation of 60% of Q4 working media
- Production start September 15
- Commercial terms as stated in Section 11
Decision required by September 12 to hold the November 3 launch.
*Calder & Vine — Strategy & Integrated Planning*