OPENING REMARK
ISRIYA PAIREEPAIRIT – CO-FOUNDER, BLOGNONE
Isriya started by addressing Famous VC Marc Andreessen’s quote “software is eating the world” that has inspired many companies to realize that software is the key to drive industry. He pointed out that top companies have changed from energy or banking industry to technology companies like Apple, Alphabet, Facebook, Google, Amazon.
Furthermore, he also emphasized the idea that every company is a tech company, and provided examples of non-tech industries that applied technology in their business as follows.
Agriculture
DJI, a high-tech manufacturing facility specializing in unmanned aerial vehicles, has automated drones to locate certain farming areas and distribute pesticides into the area. The service aims to increase more agricultural productivity while decreasing manpower costs.
Food
JUST Inc., a food manufacturer, retail and biotech company, made a major investment in automating the process with large industrial robotics, artificial intelligence and machine learning to produce plant-based foods like scrambled eggs, mayonnaise, and other foods that are synthesized from plants.
Construction
MX3D, a technology startup in Amsterdam, built a 3D printing bridge from AI technology that helped optimize design, material and cost.
Isriya closed his speech with the core idea of the Blognone Tomorrow event, about the focus on how to apply technology to business.
WHAT’S NEXT AT LINE
ARIYA BANONYONG – MANAGING DIRECTOR, LINE THAILAND
Ariya had given the background of LINE business and its strategy that LINE Thailand, Vietnam, Taiwan, Indonesia are not a branch office of LINE Headquarter, this therefore allows each office to differently manage it business with their own engineer team and can produce local services while they are flexible enough to adopt global services. LINE Thailand is operating with average employee age at 29 years old who are tech millennials.
He further discussed about LINE market in Thailand that there are 45 million users which is 2/3 of online population. He further explained about payment market that the major payment method is not from credit card, but from cash payment. Figure shows that average income of mobile users is 7 USD which means they could not afford to pay for expensive services, especially subscription services. Subscription business model must be encountering difficulty from this situation.
Ariya moved forward to talk about LINE platforms which are key business of LINE Thailand below
- Games
- Content service – LINE stickers, LINE TV, LINE Today
- O2O – LINE Man
- Business Solutions – LINE Business Connect, LINE Ads, LINE Official Account
- Payment – Rabbit LINE Pay
- New service – LINE Job
LINE believes that in order to become competitive, LINE needs flexibility to produce local services. Therefore, LINE decided to drive its business with hyper localization strategy which is different from other tech companies who focus on centralized strategy. The examples of business localization are mentioned below
LINE Payment
LINE launched kiosk sticker top-up payment solution by partnering with Boonterm and Termsabuy top up machine as majority of user is still unable to access online payment/credit card at this moment and also expanded its payment business through USSD payment channel. LINE realized that it is important for business to approach this massive as pre-paid market.
LINE TV
It is no.1 TV platform that was initiated by LINE Thailand. It took only 3 years for the platform to become successful and was adopted by LINE Taiwan.
LINE Today
It is no. 1 news portal operating with a strength to collect contents from all media into single point.
LINE Man
This platform is 100% locally developed in Thailand, it is providing 5 services; food delivery, taxi service, postal service, messenger service and convenience goods delivery.
Line insisted that LINE is not the first who invented services but their core strength is to understand customer behavior.
Rabbit LINE Pay is another trending e-wallet platform, it is a joint venture by LINE, Rabbit and AIS. Ariya stated that there are many e-wallets in the market recently but many of them may disappear or will be consolidated later on. The key to business success is to build ecosystem of cash-in and cash-out points with partners.
Last, Ariya had talked about future plans LINE is executing.
Redesigned LINE@ platform
Line@ for SMEs and Line Official Account for enterprise, they will be merged into the same platform during the beginning of 2019.
BITBOX by LINE Global
It is a digital token exchange service (not available in Thailand and Japan).
LINE Smart Speaker with Clova AI
LINE is developing AI and voice activation also planning to connect voice to all ecosystems.
Games
LINE Thailand is partnering with Woody launching live quiz show called TenTen under reversed gameshow concept; from 10 players million viewer to be million players on LINE platform.
LINE TV
LINE is producing original content for LINE TV only and working on how to draw traffic to this original content.
LINE Job
LINE is launching a new job platform for potential applicants and recruiter who are already on their LINE platform.
OVERVIEW OF IOT DEVELOPMENTS IN THAILAND
HUI WENG CHOENG – PRESIDENT, AIS
Hui started his presentation with the list of digital transformation enablers; mobility devices, digital reality, cloud, big data analysis, machine learning and artificial intelligence.
He pointed that by 2020, the number of connected IoT will increase to 20 billion devices from 7.6 billion world’s population. He gave am examples of technology companies who had shifted into IoT market
- Google launched IoT Core; a cloud services tailored to IoT devices
- Deutache Telekom launched the first NB- IoT service in Germany
- Trusted IoT Alliance launched to advanced based blockchain- IoT
- In 2020 Google nest; device that controls every part of household
- Microsoft announced 5 billion investment on IoT in 4-year plan
Additionally, Hui discussed about the situation of IoT competition in countries across the world. The US is focusing on developing advanced manufacturing system based on industrial IoT application while European Union is focusing on 3 pillars; IoT ecosystem, human centric approach, single market for IoT. China is another big country who has announced 13-year plan indicating that manufacturing is the key area of IoT application. What about IoT in Thailand? In 2020, government aims to become ASEAN’s digital infrastructure hub focusing in Bangkok, Phuket and Chiang Mai. Most of spending will be going to manufacturing and logistics.
Commercial Telco Offerings are another key that Hui talked about, they are a combination Connectivity, Security & Network, Platform, IoT Application and Operations. Hui then moved forward to AIS services starting with AIS IoT network; NB-IoT and eMTC. NB-IoT is the network for sensors and eMTC is the network for mobile devices.
Moreover, AIS also provides IoT platforms namely Milky way, Magellan and AIAP. Milky way (in house development platform) is an IoT connectivity portal. Magellan is an IoT hub and an application development platform for device management. AIAP or AIS IoT alliance program is an IoT program that allows members to participate in ecosystem; technical knowledge, business matching program, IoT development infrastructure, AIS IoT events and marketplace https://aiap.ais.co.th/
AIS has initiated and executed variety of IoT projects, here are some examples
PTT – Real-time monitor gas transmission, analysis on pipeline voltage response for the purpose of predictive maintenance, cost reduction and reliability
SCG – Navigate the right truck to the right dock to ensure precise logistics from loading to delivery
Price of Songkla University – Remotely monitor quality of air, soil and water to improve living condition for residents
EGAT – Monitor generator temperature and arsenic level in water to prevent contamination
Sansiri – Partner with Microsoft Azure to provide mixed reality (MR) technology to allow Sansiri to have new selling approach
SC Asset – Integrate and develop IoT solution on smart living
Property Perfect – Integrate devices together to make smart home and operate all from control center
Before ending the session, Hui introduced the audiences to AIS IoT aaS service which is a playground for developers to simply register and develop their IoT solution by using APIs and other infrastructure to invent IoT products and sell in the marketplace. Last, Hui triggered audience’s mindset that IoT is not just Internet of Things but it is Infinite Opportunity of Things.
AI FOR EVERYONE
MICHAEL JITTIVANICH – HEAD OF MARKETING, GOOGLE THAILAND
Michael began his keynote with 3 key business objectives that Google is focusing on, which are
- Bring benefit of AI into application
- Help others to innovate
- Help researchers to solve big challenges
Bring benefit of AI into application
He first discussed about Google photo which has capability to identify objects in photos for examples animals and sunset view, users can now search their photo by describing objects or scene in the photo. Google’s word lens is another Google feature developed on machine learning technology to understand language characters and be able to perform translation in 38 languages. Google lens is another image recognition feature designed to bring up relevant information using visual analysis.
Google also developed machine learning to detect spam and malware on email and to generate smart reply by checking email context and automatically suggest short replies, this is a new feature on Gmail. The last AI application is Google Assistant that detect voice and understand conversation meaning from dialogue context.
Help others to innovate
Michael further introduced audiences with applications that help innovate idea or project starting with TensorFlow, which is an open-source software library for dataflow programming used for machine learning applications. It was implemented by Kewpie to detect child food to ensure safety. Tensor Processing Unit is another AI accelerator application-specific integrated circuit specifically developed and optimized for neural network machine learning. The last platform Google provides to support innovation is Google Cloud API that allow users to access well-trained APIs for image/voice recognition and other Google services.
Help researchers to solve big challenges
Google has been expanding its AI service to solve pain points in various industry. Google AI was adopted by healthcare business to prevent blindness possibility from Diabetic retinopathy. AI accuracy on detecting back retina photos was equivalent to general doctors in 2016 and now leverage to specialist doctor in 2018. It also helps to predict medical possibility by collecting patient signals in every visit and create prediction model. AI also plays a great role in agriculture business, it is able to identify diseases on plants and this application has enabled farmers to evaluate their crops. Apart from that, AI voice recognition was applied to endangered animals tracking and supported wildlife conservation. Michael has closed his session with a great discovery from machine learning that could analyze space telescope data and discover new planet called Kepler90-i.
HARNESSING POWER OF AI FOR BUSINESS
METHIT MUKDASIRI – CLIENT PARTNER, FACEBOOK THAILAND
Facebook, one of the most popular social media services, is developing their AI algorithms towards business approach. Methit has clearly summarized and illustrated Facebook features used in business solutions that Facebook AI help enable business to reach target audiences, deliver right messages and drive sales volume.
Business Solutions
Methit started the session with Facebook Brand Awareness that offers optimization campaigns to enhance targeting precision and provide learning model will collect data and polling from reaction on random ads distribution to identify potential customers for specific products. Facebook Dynamic creative helps create multiple versions of ads and distribute them to each target group to react, resulting in mostly- reacted version among the audiences.
He further showed the combination of 2 interesting services; Facebook Dynamic Ads and Facebook Broad Audience. The Dynamic Ads can detect activity signals of website visitors and display products and related products based on individual interest (as known as retargeting) while the Broad Audience service detects usage signal on social platform like Facebook and Instagram, and co-function with the Dynamic Ads to offer personalized product suggestion on other platforms.
10-Year Roadmap
Facebook is aiming to develop platforms and applications to offer more personalized experiences by creating new products for individual users and business sectors as well as placing more investment in Artificial Intelligence (AI), Augmented Reality (AR) and Virtual Reality (VR).
Facebook Innovation in Thailand Market
Facebook adopted social commerce from inbox purchasing model and developed marketplace platform. Another innovative initiation is partnering with KBank to offer non-credit card purchasing model.
HOW TO MAKE 5G REAL IN THAILAND
DR. AIK CHINDAPOL – VP HEAD OF TECHNOLOGY STRATEGY, DTAC
Dr. Aik opened his session with the revolution of internet. He showed the audiences that the increasing usage of internet is driven by mobile internet; the percentage of internet from mobile devices in Thailand and global are 60% and 50% respectively. His statement has emphasized that mobile internet revolution has developed not only from new technology platform or innovation but also the amount of device usage.
4G Vs 5G
The difference of 4G and 5G is interestingly pointed out, Dr. Aik mentioned that 5G will provide greater performance in terms of speed, latency and connection. 5G speed performance is 10-time faster than 4G. This performance will lead to better online experience of users as 4K YouTube video can be downloaded in only 1 sec with 5G.
As for latency, Dr. Aik stated that responsiveness in communication is 30 time better with 5G. He further gave an example about the idea of the ball balancing table treatment for Parkinson patients that could be virtually created by 5G and virtual reality technology to allow patients to get real-time treatment at home.
Connection of 5G is expected to be 100 time faster. dtac shared the current global situation that there are 7,500 million people connecting to 8,700 million devices and the number will increase to 10,000 million people with 20,000 million devices in 2020. Therefore, the revolution of 5G network connection design must be shifted from individual coverage to device usage coverage and also to support future growth of device usage.
5G is the whole new experience
Dr. Aik further explained that word of mouth can be created from great experience and 5G is the key enabler. One great example is drone network connection, drones were used to bring out 3D damage map during Nepal earthquake in 2015. The map allowed other rescue units to further proceed disaster recovery. 5G network connection will empower more drones to have greater performance for similar activity.
5G Enablers
Key enablers for 5G are connection management, advanced radio technology and virtualization. It is important for network providers to have a good connection management model to control layers of multiple connections. In order to have efficient connectivity, many techniques are applied to 5G development, beamforming is one key technique to filter inference in higher radio frequency. Virtualization or virtualized system can lower latency and improve connection between other processing systems like cloud, analytic engines and AI communication.
Additionally, dtac emphasized that good ecosystem is facilitating 5G development, lower regulation barrier and clear spectrum roadmap will expedite the deployment while industry collaboration is another in need issue. Last, Dr. Aik mentioned that dtac is developing its foundation for 5G through products and services; Beamforming Athena, IoT solutions, Core Virtualization and Digital ecosystem (dtac Accelerate).
PANEL DISCUSSION: HOW TO BUILD YOUR OWN AI AND DATA ANALYTIC
TANAPONG ITTISAKULCHAI – ENTERPRISE & PARTNER GROUP LEAD, MICROSOFT THAILAND
DR. NATTH BEJRABURNIN – SENIOR DATA SCIENTIST, SCB ABACUS
AI: Empowering us all
Tanapong started the panel discussion with the greater role AI is becoming in current situation as how people are afraid of its capability. However, Tanapong expressed his thought that AI is not replacing human but will amplify ability of human to perform and achieve more. He gave an example of how technology help reduce time and resource from detecting bridge corrosion, drones and machine learning are used to examine bridge by taking and analyzing bridge photos. Another example of AI tool is chatbot, Thai airways and Revenue Department both adopted AI and ML to help respond customers/client inquiries.
AI Ethic
Tanapong further discussed about issues that Microsoft is concerning in AI platform development, there are values that AI needs to respect; fairness, reliability & safety, privacy & security and inclusiveness. He stated that we have to define that it’s not about what technology can do, but what it should and should not do.
What’s SCB Abacus business?
Dr. Natth first introduced the audience with the business model SCB Abacus is developing, here are their solutions
- Credit scoring platform – Clickcash platform for e-marketplace
- Robo advising platform – Puekhom Chatbot on messenger and Keb Hom application for saving
- Personalization engines – My deal feature in SCB easy app, ML will rank user preferences and offer personalized deals
5 lessons to start building AI & ML product
Dr. Natth shared data scientist practice from real experience that business should never start with AI but begin with people’s pain point or expected outcome. He further pointed out that AI and Ml requires data, therefore it is important for a business to use labeled data (user feedback/interaction on platform like, delete, dislike) in analysis or put some investment in data labelling service. He advised that complex AI project should be built incrementally, it should be able to first create simple debugging model and develop analytics once stable. Success metric should be created to monitor developing status in order to evaluate and predict success and failure stage. If the project is going to fail, fail fast and do not waste time and resources. And final element in developing AI project is team component, only data scientist is not sufficient, it also requires software engineer and software developer. However, if data analytic is not a core business, it is better not to have in-house analytic team but engaging outsourcing service instead.
Isriya (Moderator) has questioned both speaker about what is the key feature, service or interface for AI development, both panelists agreed that good customer data, data lake and data warehouse are the starting point of AI project development. Learning culture in a team and ready-to-build platform are key to success.
Source: Blognone Tomorrow