Community Update: [Bangkok AI] #6 “AI in Retail”
[Bangkok AI] Meetup#6 “AI in Retail” was successfully organized at Rise Academy, Bangkok, Thailand on March 29, 2018. The event brought more than 100 people together to learn innovative retail technologies from the domain experts and AI professionals.
The event consists of 3 interesting topics and a panel talk. In the first topic, Mr.Parin Songpracha, CEO and founder at Nasket Retail shared viewpoints for the next generation of retail channels. He said “There a lot of good mobile apps developed but very few apps or online platforms are successful”. This is a reason why Nasket develops its own hardware installing in the condominium rooms. “It is a simplest way to order a grocery product by scanning its barcode.” Nasket’s CEO said. Nasket has developed artificial intelligence for enhancing operations and services, e.g. new product category classifications to reduce human errors and labor costs, product recommendations to offer proper products and promotion, and advertising optimizations to maximize advertising efficiency etc.
The second talk was given by Dr.Asama Kulvanitchaiyanunt, CEO and co-founder at Coraline on AI and big data opportunities in retail. She shared the audience the data science experiences on the retail sector and challenges in various real cases. “Some companies have large amount of data but they have no idea how to use it” Dr.Asama said. This is an important role of a data scientist to give them right advice and to work with them. Proper data architecture and process designs are very crucial in working with retail big data to develop an automated data product. Dr.Asama shown us real cases of retail product design using big data analytics to choose optimal features and colors. Also, promotion management for modern retail shops are based on locations and purchasing behaviors. Demand forecast, inventory and membership management are challenging issues in the retail industry and machine learning algorithms are practically used for the optimal management.
The third talk was shared by the TESCO Lotus hackathon 2018 winner team, Tomato Chatbot. The team spent 48 hours in developing the visual search engine for product identification. “Just a snap to by a product you want” highlighted by the Tomato team. For example, when the customer sends a product photo via Facebook messenger connecting with the chatbot, the AI engine will reply the available products within seconds and customer can make the order through the chat platform. This could help senior customers more convenient in online ordering goods.
Lastly, the panel talk was about needs of AI, Data science, ML and used cases in retail. There are 3 panelists discussing about how AI drive inventory management, logistic and customer insights.
Miss Atiporn Jaroensri, Tesco Asia Technology Director, said that Tesco has massive customer data and she is looking for the day that no more inventory warehouse, while the right products can delivery to stores in the right place and time. It means Tesco can save a lot of money from the warehouse and inventory management costs. And the technology can enhance retention and extra services recommended by machine learning and AI systems.
“In an advanced country, the vendor-managed inventory (VMI) system is very useful for the retail since it can significantly reduce the inventory management costs” said by Mr.Gling Kanchanasuwa, a retail and pricing expert.
Dr.Asama Kulvanitchaiyanunt, CEO and co-founder at Coraline mentioned on the UPS case in using data analytic to reduce significantly the delivery cost by constraining the driver to only turn right on the carry routes. Furthermore, she gave another example from Grab that Grab’s algorithm attempts to find the highest possibility of the taxi driver who will pick up the user, not the nearest driver. This makes a big value for taxi users who usually face the service problem from the regular taxis, especially in Thailand.
At the end of the panel session, there was an interesting discussion on how enterprises begin implementing big data analytics, especially the company that has not well prepared for data architecture. Dr.Asama suggested the company may start with collect the entire data into a data lake and then the data scientist can pick up relevant data to set up new data warehouses prepared for data analytics. This is a practical approach for modern data management and can avoid problems from modifying the existing data architecture.
Many thanks to Bangkok AI volunteers who helped organizing the event and special thank to Mr.Parintorn Kunatatorn for capturing the event photos and drafting the meetup summary.
