So in some of the social apps, we also see our own platform. They start to use this large language model to create specific characters with a defined background personality and even like the of their talking language. And that attracts quite some interest from consumers because that gives customers freedom in sort of having a conversational relationship with someone that, by imagination, is their ideal partner. So that’s an interesting invention just recently. But similarly, those things could happen in customer engagement use cases. But also, people would naturally imagine that there were this smart chat sort of service, including from the past Alexa and Siri. In the past, the conversation somewhat becomes very dry, like boring. We can only do command control kind of conversation.
But now with the large language model, human and machine interaction interface could be totally disruptive. On that side, we could enhance such a conversation or a human machine interface, in a way, to ensure the interaction to be real time. So those are all potentials we could see, but there will be more. We do have a close eye on those possibilities.
Jingbo Wang: I’ll quickly chime in on the take rate and pricing side. So Tony mentioned a few new use cases, like live shopping in the U.S. There, we see the pricing remains very healthy. And digital transformation in China, that’s a very different market with large enterprises, financial institutions. So there, it requires a different pricing model. But we believe what we see is it’s still a very profitable market for us.
Operator: And our next question will come from the line of Allen Li from JPMorgan.
Allen Li: I have two questions. So firstly, could you please give us some color on the revenue breakdown by key categories in 2022, such as entertainment, education and maybe other some rising categories like live video shopping? And my second question is on our business storage. We see non-China business have slowed down quite a bit in past 3 quarters, while China revenue, excluding K12, seemed to hold up well. So do you think this is more like a temporary thing? Or is there any structural reasons behind that? And how are we going to allocate resource between China and the international markets going forward?
Jingbo Wang: Okay. I’ll take both questions. So the first one in terms of the — okay, first of all, we operate in 2 markets, the U.S. and international markets and the China market. Obviously, the vertical breakdown in those 2 markets are slightly different, so I will talk about them separately. So in the U.S. international markets, social is still the largest vertical, followed by what we call future in work. And that will include things like collaboration, online event and live video shopping. The second biggest is education. I think these are the main verticals. In China, obviously, social is dominant because now, after the regulation change, education is a much, much smaller part of the business now. And then digital transformation, so large traditional enterprises previously was almost zero and now has a meaningful contribution last year.
And also IoT, Internet of Things, something like smart doorbells, smart TV, these will also start to contribute revenue in 2022. That’s on the sector breakup. In terms of like China revenue holding up in the past few quarters, I think there are a few factors. First of all, it has to do with the new product, what we call the broadcast streaming product, which had lower GP margin but has been very nice revenue pickup. So that now contributes to about 10% of revenue in China already. So that’s something new in 2022. And secondly, we have been working with customers in China, not just focused on business in China but also, for example, expanding their apps in overseas markets. So that’s another growing business. And lastly, as I mentioned, the digital transformation business, that almost didn’t exist in 2021.
