FB Pixel no scriptChatbots find fertile ground in Thailand, Southeast Asia's largest social commerce market | KrASIA

Chatbots find fertile ground in Thailand, Southeast Asia’s largest social commerce market

Written by Giovanni Puglisi Published on   5 mins read

The Thai language poses some specific challenges to chatbot developers.

Sealing a deal in a chat conversation on social media is a common practice all over Southeast Asia.

It dates back to the earlier days of the internet, when there were social networks and online forums but few dedicated e-commerce sites in the region. Small businesses used social media this way because it’s an easy way to start and requires little investment, but offers the potential to reach millions of customers thanks to the proliferation of smartphones and mobile internet.

Commerce through social networks also offered an element of trust. Before you follow through with an online purchase you engage in a conversation or even befriend the seller.

Eventually, e-commerce platforms such as Shopee, Lazada and Tokopedia became popular in Southeast Asia, and they are meant to make the online shopping process easier and more efficient, and reduce the amount of time a seller needs to spend with each individual client. But social commerce has not faded into the background.

Quite the opposite–”informal” social commerce on platforms like Facebook and Instagram continue to thrive.

Also, social and conversational elements are increasingly finding their way into the “formal” online shopping environments created by e-commerce platforms. These worlds are colliding, aided by technology.

One example is Lazada’s venture into live streaming, where sellers can personally promote products and interact with customers. Another development that has the potential to “turbocharge” social commerce are chatbots–AI-powered programs that are integrated into messaging apps, where they help online sellers deal with customer requests more efficiently by automating parts of the interaction.

Testing ground Thailand

This is especially the case in Thailand, the market that embraces social commerce the most among its peers.

A study commissioned by PayPal in 2018 found that As many as 95% of merchants in Thailand see social media as a viable business platform. According to a 2019 Forrester Southeast Asia survey, 70% of respondents in Thailand said they use social networks or online social tools to talk about, post or trade products and services on a daily basis, compared to 38% in Singapore and 53% in Malaysia. Sales via social media in Thailand stood at USD 10.9 billion in 2017, according to an estimate from the country’s Electronic Transaction Development Agency.

These figures made Thailand an attractive testing ground for new social commerce features.

Facebook–still the largest social network globally–has been considered a slow mover in this regard. It launched Facebook Marketplace in 2016, a suite of features to simplify peer-to-peer sales, and allegedly plans to launch a standalone shopping app that complements Instagram. But considering Facebook’s scale, its efforts to transform into a commerce ecosystem seem restrained.

Thailand has been one exemption for the US social giant, where it launched an innovative social commerce feature for Thai users last year.

In cooperation with Kasikorn Bank, Facebook now allows card payments and P2P transfers, as well as chatbot support for in-app payments on its messenger interface–one of the few countries in the world where this feature is available. P2P payments through Facebook Messenger are also available in the US, but in the UK and in France, the firm discontinued this feature earlier this year.

Chris Marshall, associate vice president for IDC Asia Pacific told KrASIA that adding these functionalities to Facebook’s messenger is a natural extension for the social network, as people invest in their existing Facebook pages to complement their brands and drive traffic to their website. “It is a logical transition for them to also adopt a chatbot within Facebook’s native messaging app,” he said.

The upsides are obvious. From conversations with online sellers, KrASIA learned that small business owners are constantly asking themselves how to improve sales. Some of the reasons for failing to increase online sales can be that the customer needs more product information; wants to compare your product with other products; or has difficulties with finding products quickly and with making payments in one place. It’s these recurring, repetitive problems where chatbots can step in to support small businesses.

Imagine that the small business owner has a potential customer that needs help finding a pair of trousers to match a jacket, and would be able to simply type in “what trousers will match this jacket?” followed by a photo of what is wearing. An algorithm would instantly analyse and respond with three different options that fit the criteria. With payments integration in the chatbot, the purchase could then be completed without leaving the messaging app at all. In this way, businesses can help their customers save time too, but still have the conversational shopping experience. If a chatbot hits a wall with a request, a human customer relations manager can step in.

Chanakarn Chinchatchawal, Founder of Zwiz.AI, a Bangkok-based startup and bot developer, notes chatbots can help SMEs in two ways. Firstly, they provide automated reply to the frequently asked questions. Secondly and most importantly, they help shortening sales processed by providing a way to let the customer order and pay directly in the chat.

“Most customers in Thailand still prefer to pay via bank transfer. Chatbots would summarize the total price of the order and provide simple payment solutions. For example, chatbots would help capture the payment slip image and forward it directly to the SME owner for verification”, Mr Chinchatchawal said.

Diamond Corner, a small retailer that sells its products both via Facebook and through e-commerce platforms such a Lazada, have registered a number of improvements to their sales by adopting chatbots, says Chinchatchawa. Diamond Corner is a Zwiz.AI customer.

The cost saving factor has been the most significant for the retailer. With 64% of orders placed by customers without human support, the company was able to reduce their service desk staff from 10 employees dedicated to the management of 9 Facebook pages, down to only 4 people dedicated to ‘supervising’ the chatbots, which also allowed them to further save time and to shorten the sales process while providing a 24 hour chat commerce service.

The “get to know your customers better” aspect is also an important feature. Chatbots contribute to SMEs growth by using  data and customer information to market their products and by sending additional notifications to the sales team for commercial purposes. These type of bots are generally known as the “Task bots”, which typically take information or functionality from somewhere else, process it in some way, and then route it back to the user, Chinchatchawa explains.

In Thailand were consumers embrace social commerce, chatbots can potentially provide huge advantages for SMEs, but of course there are also downsides. The most obvious is language. Chatbots must be trained to be able to interact with customers in a way that feels natural. In the case of the Thai language there are several limitations. One particularity is that in its written version, Thai features no spaces in between sentences, but similar sentences can have different meanings depending on context, so chatbots need to process the sentence by looking at previous conversations.

However, it’s possible to partially overcome these shortcomings with advanced chatbots equipped with heavy natural language processing (NLP) capabilities.

Marshall notes that conversation types of bots “present some of the most difficult and intriguing challenges for bot developers as they take unstructured language, turn it into structured data, and then translate that data into a response. This is made even harder by virtue of the Thai language”.


Auto loading next article...