NetEase launches its first AI-native voice agent, offering 100,000 yuan for honest feedback

NetEase has released “Bage Shu” (网易叭哥说), which it describes as its first AI-native voice agent, together with a user co-creation effort called the “Honorary Co-founder Program.” According to a QbitAI report published on October 16, the program sets aside 100,000 yuan to solicit product suggestions and imposes no major or degree requirements on participants.

The “AI-native” label is the substantive part of the positioning: it presents the product as a voice agent built around spoken interaction from the start, rather than a voice feature bolted onto an existing app. By calling it its “first” such product, NetEase places voice agents in a separate category within its AI lineup rather than as an accessory to an existing business.

The rules of the campaign have drawn more attention than the product itself. The money is explicitly budgeted for collecting suggestions rather than for conventional promotion, and NetEase summarizes the ask as finding “real needs” — rendered in coverage as “tell the truth and take 100,000 yuan.” Putting a price directly on user feedback turns a product launch into something closer to requirements research.

Dropping the degree requirement matters more than it sounds. Voice agents are used by people who never sit in a product review and never read a spec sheet, so their complaints tend to sit closer to real usage than the assumptions of industry insiders. Lowering the barrier is an admission that the answer to product definition may not live inside the company.

Handing part of the product definition to users is an increasingly common choice in this round of AI competition: as model capabilities converge, scene selection becomes the main source of differentiation, and scenes hide in ordinary conversation. The cash incentive cuts both ways: it can attract a volume of generic submissions written for the reward, and separating actionable requests from noise becomes an operations problem rather than a budget problem.

The question Bage Shu actually has to answer is not whether it can hold a spoken conversation, but which category of tasks a voice agent should take off a user's hands. Whether the 100,000 yuan yields a feature list and a set of scene priorities, or merely a footnote in a launch post, depends on how the team digests the feedback.

Three things are worth watching: whether the criteria for judging suggestions are published, whether accepted ideas show up in the product's release cadence, and how the 100,000 yuan is actually distributed. Those details will reveal whether this co-creation effort is part of the product mechanism or a publicity move.