AI amplifies human ingenuity with intelligent technology to make innovation faster and more accessible. Its capabilities are endless, and it’s re-defining our future and promises even more staggering changes to come.
Wavelength recently hosted an event where one of the most prominent global thought leaders in the deployment of AI, Dr Wanli Min, Chief Machine Intelligence Scientist and Vice President of Alibaba Cloud, spoke to an audience of business leaders.
Alibaba has evolved from being an e-commerce company into a vast digital economy spanning e-commerce, logistics, financial services, cloud computing, digital entertainment and new retail.
We asked him to share insights.
Dr Min, can you share your top tips for those considering AI to innovate?
Always start with the three A’s accessible, affordable, actionable!
Can we access the data? Can we afford it? What is the best way to take it forward? Start with the business problem. Where are the pain points? Where is it hurting the most?
Alibaba’s hugely diverse business is driven by data intelligence and constantly re-fueled by new data, from digital media and entertainment to core commerce to hospitality to logistics and local services.
And at every point we have taken steps to establish exactly what needs to change in order to serve our customers best.
What sort of data helps build this picture?
Data that helps us have a better understanding of what our customers want.
We have developed metrics to calculate the affinity between a specific merchant or product and a specific buyer. It helps us predict the propensity of a buyer to fall in love with the item and buy it.
There are many factors that are important: the price, the reputation of the seller and the aesthetic and design of the page including elements such as video. From this we build a model to calculate the probability of a match.
What were the results?
A huge uplift in sales. We now have nearly 700 million active users on our platforms. In fact, Alibaba took $1 billion sales in the first 85 seconds of the Global Shopping Festival held on ‘Singles Day’ (the Chinese equivalent of Black Friday) grossing a staggering $30.8 billion in 24 hours.
So, how do you develop the technology?
The next step is to identify the best technology for the problem. In our case it was an AI powered machinery model – but that may not be the case for everyone. Don’t jump into AI too early. You need to ensure that you have enough data. It may be that you don’t need it as other tools such as traditional data science or linear regression modeling may do the job.
Once you have solved the business problem, connected the right data set and the right technology – the rest is straightforward.
Are there any other important considerations?
Consider what kind of talent you need to engage. Do you need a whizz kid with a PHD from Stanford or someone who has worked in your domain for 20 years?
In my opinion neither is enough, and both are imperative.
The answer lies in finding a combination of domain and subject matter expertise and AI technology and bringing them together. That’s when the magic happens.
The AI experts should work with the subject matter experts to translate the process into a data flow.
But a final word of warning is, don’t take a moon shot. Take it step by step, implement measurable KPIs and start with the smaller projects.
What are the other parts of Alibaba’s business where AI has made a huge impact?
Freshippo supermarkets are viewed as the future of retail. They converge online and offline shopping. Part retail theatre, part dining experience, part logistics business. We make grocery shopping a real experience and AI makes this possible.
All shopping is done through the Freshippo app enabling us to build an incredible picture of what our customers want and tailor our offering accordingly. Scanning products with the app reveals details on provenance and shares recipes and ideas. Shoppers can eat in store or place items on a conveyor belt and have them delivered home in half an hour.
In fact, hungry customers don’t even need to leave home, a tap on the app arranges for food to be delivered to the door in the same timeframe.
We have also made huge improvements in our logistics and delivery business. With an unlimited number of consumers, the demand is always dynamic, and we never know where buyers will be based.
We discovered that when people order food their patience won’t stretch past half an hour.
Only a machine learning algorithm can balance the supply and demand and make our logistics networks and food delivery businesses so effective. Food is delivered within 30 minutes for those who live within a 3km radius of a store.
Then there is our futuristic Flyzoo hotel in China. Here guests can check in and gain lift and room access with facial recognition technology (yes that means no more room keys!). They can also enjoy the benefits of a fully voice activated hotel room, requesting changes in everything from temperature to lighting to curtains to music. The Flyzoo restaurant combines human waiters with robots, who can even make cocktails. It’s AI powered technology at its most hospitable and entertaining.
Where do you see AI technology going in the future?
AI technology originated from the Internet and e-commerce, but it’s clear that it’s going to be used beyond that and penetrate the traditional economy. This includes agriculture, manufacturing and the smarter city.
The smarter city has always been associated with hardware (cameras, databases, PCs, etc.) But technology from e-commerce can be translated and injected into the smarter city and combined with its historical and live streamed data. Making the smarter city even smarter.
So, technology that originated from the internet will inevitably shift into other traditional data intensive sectors.
We have also developed technology called City Brain, which helps with the massive issue of traffic congestion in some Chinese cities.
It leverages all the existing available data such as cameras and traffic sensors and combines it with the AI algorithm and computation power. This provides vital insight for the authorities and helps improve traffic flow, reduces congestion and improves the turnover rate of parking spaces. Hospitals are using it as well, with ambulances shaving off chunks of time and therefore saving lives.
AI technology can also be used to improve quality control in manufacturing. Fifty years ago, General Electric devised Six Sigma, a system which streamlines processes to reduce defects and improve quality control. We believe other real time data will give us better ideas and suggestions about what kind of parameters in the process should be adjusted. Again, we can use AI capability with existing data to provide greater detail, reduce product flaws and reduce energy consumption.
The opportunities are endless and will permeate many sectors. It’s very exciting for business. Good luck!