AI startup DigitMaster uses virtual consumers to simulate product launches, cutting trial costs for brands
创业邦 · wechat · 2026-09-09
Startup DigitMaster, founded by 00s entrepreneur Ding Tian, offers BrandMaster, an AI agent system that simulates consumer feedback to help brands test new products before launch.
Key numbers:
- Only 15% of FMCG innovations succeed; in 2023, only 6 of 3000+ beverage launches exceeded ¥10M monthly sales.
- BrandMaster cuts insight generation from 4-8 weeks to 4 hours, R&D cycle from 2 months to 1 week, boosts marketing ROI by 80% on average, and lifts new product success rate by 65%.
- Annual subscription costs ¥200K-500K; clients include Estée Lauder, Hema, 7-Eleven; 2026 revenue in millions of yuan.
How it works:
- Builds tens of thousands of virtual consumers based on real data to simulate reactions to concepts, pricing, channels, and marketing.
- Three types of agents: consumer insight, product R&D, and brand marketing, covering the full lifecycle.
- Proprietary "industry ontology" reduces hallucinations.
Business model:
- Starts with top international brands for validation.
- Partners with ad groups for internal efficiency and joint client service.
- Also testing a personal assistant "Pebble" for creators and individual operators.
Vision: Build an AGI-era content data infrastructure with a B2B2C data flywheel.
More from Companies & People
- DHH's Omarchy hires Quickshell creator full-time as plugin catalog nears 3,000 — talkaboutdesign · 2026-09-09
- Fei-Fei Li: R&D is forking into token-abundant and token-starved research — drfeifei · 2026-09-09
- Codex Community Meetup lands in Oxford with a Bayesian ML vs. coding agents talk — paw_lean · 2026-09-09
- CS professor: campus is not the best place to learn software engineering — lemire · 2026-09-09
- Anthropic researcher puts >10% odds on AI 'killing all humans' after colleague quits — KoseteBamse · 2026-09-09
- Stealth mode is dying: founders stop sharing ideas because they're too easy to copy — deliprao · 2026-09-09