The AnyMind Group announced on Thursday (24) a strategic partnership with Alibaba Cloud to incorporate the models Qwen into AnyLive, its live commerce platform with artificial intelligence presenters. The integration seeks to improve the understanding of consumer questions and the generation of real-time responses during sales broadcasts.
Qwen will be added to AnyLive's multi-model architecture and will operate through the AI avatars used in the broadcasts. According to the company, the model should better interpret conversational or ambiguous questions and generate responses more suited to the context in different Asian languages.
The partnership also targets localization aspects that go beyond translation, such as the use of forms of address and adaptation to the cultural particularities of each market. AnyMind states that Qwen is already used in large-scale e-commerce operations, including broadcasts with digital presenters on Taobao.
AnyLive combines AI avatars with human operation
The integration expands an AnyMind strategy aimed at keeping commercial broadcasts active for more hours, including during periods when human presenters are not available. AnyLive allows operating multilingual livestreams with AI avatars and tracking metrics of broadcasts conducted by both people and virtual presenters.
An example of the scale sought by the platform came in April, when Samsung announced the use of AnyLive in eight markets, including Southeast Asian countries, Australia and New Zealand. The project foresees 4,450 additional hours of broadcasts per month, with AI avatars operating on different e-commerce channels.
AnyMind, however, maintains infrastructure for broadcasts conducted by humans and presents automation as complementary. The proposal is to use avatars to expand availability, language coverage and broadcast volume, while human hosts remain employed in formats and interactions in which their participation is considered more appropriate.
