JD’s big GenAI initiative is called ChatRhino. Which is a cool name.
And it’s part of a Generative AI strategy that is an interesting contrast to Alibaba.
In this article, I’ll lay out 3 ways JD is different than Alibaba in GenAI.
Point 1: ChatRhino Is a Big Play in Intelligent Industry Solutions
I have written several articles about Alibaba’s GenAI strategy, with is mostly focused on ecommerce and cloud (at this point). My breakdown is located here.
But I have also been looking at JD. And they have been slowly releasing details about their GenAI projects. Such as digital avatars and logistics solutions.
But their biggest project is ChatRhino – which is their LLM model.
JD unveiled ChatRhino in 2023 at their tech summit. And the thing that immediately jumped out at me was this comment in their press release (I added the bold).
“By combining 70% generalized data with 30% native intelligent supply chain data, JD’s latest AI model offers targeted solutions for real industry challenges across sectors such as retail, logistics, finance, health, and city.”
That’s interesting. That sounds a lot more like Baidu AI Cloud than Alibaba and other ecommerce players.
They are using GenAI to create new industry solutions.
And they are trying to leverage logistics and supply chain data and expertise in this strategy.
Interesting.
Today, JD really has two big business:
- Domestic ecommerce
- Global logistics and supply chain solutions
But in their 20 year plan (released in 2023), JD said their “path for sustained growth over the next two decades” is “37511” – which stood for:
- The “3” in 35711 is their goal of establishing three businesses with over 1T RMB in revenue (and 70B RMB in net profits).
- The “5” is their goal of having five JD businesses on the Fortune Global 500 list.
- The “7” is their goal of having 7 publicly listed businesses that started from zero and achieved a market value of 100B RMB.
Basically, they committed to building a 3rd big business over 1T RMB in revenue.
Is this what they are going for in GenAI and industry solutions?
I’m thinking yes.
Point 2: ChatRhino Will Have to Build a Flywheel Between Adoption and Intelligence.
A few years ago, I interviewed Dr Xiaodong He, who was then head of AI at JD. You can find that interview here.
Dr. He is now Director of JD Explore Academy and President of JD Technology’s Intelligent Services and Products Division. And he has released a three-step approach for GenAI for JD (I added the bold):
- Step 1: Build a generative large model based on practices within JD.com’s existing business operations.
- Step 2: Enhance and iterate the model through highly complex industrial scenarios, creating robust services for industry use.
- Step 3: Open up the capabilities of the model for serious commercial applications.
He basically described an industry-specific intelligence flywheel. Here’s my version of what they are doing:
- They are rapidly building general foundation models.
- They are using general and internal data to get these models going right now.
- Over time, they will need to make these models smarter. And they will need to develop industry specific LLMs.
- They key to all of this is getting businesses and developers to start using, customizing, and deploying their models.
- The feedback from adoption with make the models smarter and more industry specific.
Industrial adoption is the key. The resulting data and use cases will create an intelligence flywheel, making the models smarter and smarter. And more industry specific.
This is exactly the strategy that Robin Li has been pursuing at Baidu AI Cloud. I’ve written about that here.
So how is JD going to compete with the major AI Cloud players in China (Baidu, Alibaba, Tencent, Huawei) and abroad (AWS, Google Cloud)?
It sounds like they are focusing on their data, expertise and partnerships in logistics, supply chain and retail.
Point 3: ChatRhino Has Its First Use Cases in Ecommerce and Logistics
If you’re going for an industry intelligence flywheel, then it’s all about model deployment in specific industries. They are clearly starting in ecommerce and logistics.
It means working with companies and developers. And creating lots of tool kits for them.
Businesses need to customize their general generative models into specialized ones. And for that, you need to give them tools so it’s easy, fast and without a lot of resource-intensive procedures.
JD says it now offers “ChatRhino AI Developing and Computing Platform” for industry-specific application development. And that it has “over 100 training and inference optimization tools”.
These are the KPIs I’m paying attention to. Plus, any compelling new use cases.
One use case is the Jinghui Intelligent Supply Chain Data Management Platform. This is ChatRhino applied to “sales forecasting, inventory management and replenishment”. It identifies supply chain problems and offers solutions through its “interactive supply chain control tower”. Here is some of their logistics tech.
In ecommerce, we can see a couple of compelling ChatRhino use cases:
- AIGC content marketing. Merchants and brand can create lots of visuals, marketing posters, and product images using just one product image. They are enabling rapid store setup and product promotion. JD says the production cost of visual sets has been reduced by 90%, while the time required has been cut from seven days to half a day.
- Digital Avatars. This is pretty interesting. Merchants can now create their own virtual representative to chat with customers. I’ve seen this at the JD HQ and it’s pretty great. CEO Richard Liu has been re-created as a digital avatar. Here is one of the avatars.
Note that customer service is actually a long-standing challenge for JD. They differentiate with quality service (plus price and product quality). But JD keeps growing in GMV and the volume of customer service inquiries has become a problem. Doing high quality customer service at large scale is really difficult.
JD’s customer service volume is now far beyond what can be handled with humans. It’s just too many inquiries. And hiring, training, and retaining customer service representatives is really hard. People really don’t enjoy that job and they quit frequently. Digital avatars for merchants and JD itself are an interesting approach to this.
Conclusions
Overall, JD has a solid GenAI strategy.
- They are building their own models and apps internally.
- They are focusing on GenAI use cases in areas where they have better data (ecommerce and supply chain).
- They are applying this to ecommerce but are also trying to leverage this into a big business in GenAI industry solutions.
That’s it. Cheers, Jeff
Here are some photos from a recent visit to a JD logistics facility in Beijing. They have really good logistics tech.
————
Related articles:
- GenAI Playbook (Step 3): How to Build Barriers to Entry with Intelligence Capabilities (9 of 10) (Tech Strategy)
- AutoGPT: The Rise of Digital Agents and Non-Human Platforms & Business Models (Tech Strategy – Podcast 163)
- Why ChatGPT and Generative AI Are a Mortal Threat to Disney, Netflix and Most Hollywood Studios (Tech Strategy – Podcast 150)
From the Concept Library, concepts for this article are:
- Generative AI
- Ecommerce
From the Company Library, companies for this article are:
- JD
———
I write, speak and consult about how to win (and not lose) in digital strategy and transformation.
I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.
My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.
This content (articles, podcasts, website info) is not investment, legal or tax advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. This is not investment advice. Investing is risky. Do your own research.