Reports, validation studies, and articles — see what Chorus produces and how the method holds up.
Seven UK government ads on teaching, pensions and Stoptober, each watched by 12 synthetic personas. Watch the ads, see the scores and read what they said.
10 Indian personas score Birla Opus's airport launch ad. The 16-year warranty hooks attention; the 'Main bhi' crowd undermines belief. Overall 53/100.
Chorus's 10-persona creative-testing dashboard for Birla Opus's April 2026 'Main bhi' airport ad — emotion arc, per-persona signals, scene heatmap, and timestamped edit suggestions.
Chorus's 10-persona creative-testing dashboard for ITC Mangaldeep's April 2026 'Sixth Sense' ad — emotion arc, per-persona signals, scene heatmap, and timestamped edit suggestions.
10 Indian personas spanning rural UP, Bihar, Rajasthan and metro India rate Mangaldeep's 90-second CSR ad at 79.6/100 — cultural fit strong, brand recall weak.
Chorus's 12-persona creative-testing dashboard for Burger King's 2026 'There's A New King' — emotion arc, per-persona signals, scene heatmap, and timestamped edit suggestions.
A follow-up to our Kumar et al. (2022) food-choice replication. Two moderated focus groups with the same 16 synthetic personas we used in the interview pass — Maharashtra and Kerala panels. The group format recovered all 10 of Kumar's Phase 2 sub-themes and surfaced 4 additional frames on top.
A replication of Kumar et al. (2022) Phase 1 — two focus groups on urban Indian middle-class perceptions of processed foods. 16 synthetic personas matched to Kumar's criteria (age 40-65, college-educated, white-collar, no medical/nutritional background). Recovered all 3 Phase 1 themes and surfaced 5 additional frames on top.
Chorus validation study: on Burger King's 2026 'There's A New King', our 12-persona panel reached the same broad positive direction and dip-then-rise mechanism as System1's 3.8★ review. Independently, the panel's strongest consensus (0.98) landed on the 0:32 self-critical admission. The two scales are different instruments and are not directly comparable.
We replicated the Phase 2 interview protocol of Kumar, Kulkarni & Rathi (2022) with 22 synthetic personas. Matched 9 of the paper's 10 Phase 2 sub-themes and surfaced 6 additional frames the paper's analysis didn't cover.
Audience panel on Gillette's 'We Believe' toxic masculinity ad. 9 perspectives, 35 questions, 6.4/10 overall. Strong message clarity, weak conversion pathway.
Anthropic researchers discovered a neural 'axis' that controls AI personality. Here's what it means for anyone building with LLMs.
Read
Audience panel on SVEDKA's Super Bowl AI robot ad. 8 perspectives, 27 questions, 5.7/10 overall. Uncanny valley undermines trust despite near-perfect brand recall.
Conformity bias distorts focus group feedback in predictable ways. Here's what five peer-reviewed studies reveal about the problem—and what alternatives exist.
Read
New research reveals that advanced reasoning models succeed by simulating internal debates among diverse perspectives, not by simply computing more steps.
Read
Audience panel on Hims & Hers' 2025 Super Bowl ad 'Sick of the System.' 12 perspectives, 29 questions, 6.7/10 overall. Strong emotional impact but trust concerns.
Audience panel research on Coinbase's musical theater crypto ad. 7 perspectives, 32 questions, 7.1/10 overall score. See what resonated and what fell flat.