What happened
DeepZero, stock code 2723.HK, has released an AI agent product called 'AI New Product Innovation.' The product uses multiple agents that take turns working on subtasks, all aimed at converting complex market signals into product directions that can be validated and compared. The launch comes as fast-changing consumer demand and shorter market opportunity windows expose the limits of traditional new product research, which often relies on scattered data, manual judgment, and slow concept testing.
In the opportunity identification phase, an AI demand insight agent studies market performance, consumer voices, social media trends, and the competitive landscape, supported by industry knowledge and a dedicated knowledge base. An AI strategy research agent then weighs market space and competition, while an AI scoring agent compares and ranks candidate directions to help companies set priorities. Later, an AI concept and business evaluation agent turns promising opportunities into concrete product concepts, taking into account target consumers, usage scenarios, core needs, and product value.
The system can also activate an AI consumer agent to simulate target consumers' reactions to different concepts, checking purchase intention, key attractions, and concerns, with results feeding back into concept refinement. In one real business application, a leading tea beverage brand used the product to shorten a new product research cycle that previously lasted three to six weeks down to several hours. The process narrowed the field to two priority concepts, which showed about 20% higher purchase intention than the initial options.
Why it matters
This product signals a move from experience-driven new product development to a more systematic, AI-assisted workflow. When multiple specialized agents carry out different stages of product decision-making, companies can speed up the journey from market signal to validated concept, which is valuable when opportunity windows are short and consumer preferences shift quickly.
Beyond a single tool, the multi-agent design echoes how cross-functional product teams work—combining market insight, strategic judgment, and customer feedback. The tea beverage case demonstrates that such orchestration can be measured against tangible outcomes like purchase intention, suggesting that simulated consumer reactions may become a standard part of the innovation process before real-world resources are committed.
Key facts
DeepZero (stock code 2723.HK) released an AI agent product named 'AI New Product Innovation' for end-to-end decision support in product development.
The multi-agent system spans opportunity identification, direction research, concept generation, and market validation.
In a tea beverage brand application, a new product research cycle dropped from three to six weeks to several hours, and two priority concepts achieved roughly 20% higher purchase intention.
The system includes an AI consumer agent that simulates consumer feedback on concepts, covering purchase intention, main attractions, and concerns.
What to watch next
Watch whether the product expands beyond the beverage example into other fast-moving consumer sectors where shortening research cycles directly affects competitiveness.
Watch whether DeepZero adds agents that track a launch's post-market performance, allowing the same AI pipeline to learn from real-world response and improve future product innovation cycles.