Hi, I'm Jose Luis from LINE Development Center 1, a lead technical consultant and a product manager at LINE.

On May 18th, I participated in one of the Engage sessions at The Next Web conference in Amsterdam as an invited speaker. I'd like to share about the Engage session and briefly cover the event itself, but first, I would like to express my gratitude to The Next Web for the invitation.

Opening

The opening event was fantastic. It was filled with music and visual effects, which most IT events do not have much acquaintance with. Can you feel the excitement from the pictures?

Exhibition

The exhibition was mostly filled with startups with great ideas. Sponsoring companies were looking for investment opportunities. It was great to see such venue set up to help startups. I personally hope to see LINE get more involved in helping out startup companies in near future.

My session

The topic of the Engage session I participated as a speaker was, 'Why You Should Be Building Bots for SNS Services'. In case you are unfamiliar, an Engage session is like an opening session, but the session speaker engages the listeners to participate in the session.

I noticed a few points as we introduced each other:

• The participants knew LINE but haven't actually used it.
• The participants were business-oriented.
• Most of the participants came from outside Amsterdam to join the conference.

About chatbot

What exactly is a chatbot? When I threw this question, the participants seemed to have an understanding that a chatbot is something that is capable of mimicking human conversation, which is not wrong. In a technical point of a view, a chatbot is a program that can simulate human conversation. Actually, chatbot has been around for quite a while. I have seen one when I used to use Microsoft Excel on Windows XP, where a dolphin popped up and asked me if I wanted assistance. There could even be older chatbots that I am not aware of.

So what's the benefit of a chatbot?

Since chatbots are programmed, no human being is needed to respond; chatbots will do the responding for you. You can program a chatbot to fit your service. For example, you can make a chatbot to take an order for pizza without requiring a human to ask for the pizza type or where to deliver the pizza to.

Nowadays, there are AI services for you to utilize to make chatbots smarter. Despite the benefits chatbots offer, chatbots are yet to replace human; one of the participants pointed out that teaching an AI (Artificial Intelligence) is not an easy job, and that AI cannot handle all the possible cases. Indeed, teaching an AI service is not easy, even with the high-tech tools that come with the AI service, and you still cannot predict what the user will request. So you need to constantly teach the AI service, which takes a lot of work. Fortunately, teaching AI is becoming easier with the help of machine learning. Hopefully, adopting AI will eventually become easy for startups in the future.

What services fit chatbot?

The most common services that fit chatbot are:

• Ordering: Taking orders for pizza, tickets and general orders on E-commerce sites.
• Assistance: Assisting users with travel information and service information.
• Searching: Searching for images, places to go and food.
• Supporting: Providing customer support to help service users. At LINE, we have enabled switching our assistance bot to a human operator, in case the bot cannot fulfill the customer's request.

Chatbot and social media services

As my topic was "Why you want to build a chatbot for social media services", I moved on the social media services. LINE, as a provider of messaging service, builds chatbots for the following reasons:

• Social media services have a big pool of users who are likely to use your chatbot.
• Social media service users can easily spread the words about your service.
• You do not need to make your own site.

A number of social media services do provide chatbot APIs. LINE is one of them and is a major player in Asia among other players such as Facebook, WeChat and Kik. If you ask me which social media service you should use, my answer would be LINE, without a hesitation. However, I'd also recommend you to consider where you want to provide your service at. If you are considering Japan or other South Asia region, I will surely recommend LINE, since it has a strong user base in those regions. Plus, LINE provides promotions for companies in order to increase the chatbot user base. For other parts of the world, Facebook could be an easier option.

Chatbots are still new on social media services, leaving advertising company products and finding business models as homework. However, chatbots on social media services will mature sooner or later, and most social media services will come up with the solution.

Why do companies want to build chatbots on social media services? The first reason is to advertise their products. The companies that have created chatbots on LINE, have informed that the CTR (Click-Through Rate) is much higher than their other advertisements and that more users get to see their products. Meaning that, companies have higher possibility to monetize their products by using chatbots. Other motivations for using chatbots on social media services include keeping loyal customers and providing customer support. The industry is shifting from using telephone to chatbot.

For startups, using chatbots on social media services for promoting services will be cheaper than putting up advertisements and waiting for people to spread the words about their service.

Q&A

Here are a few of the questions I have received from the participants.

What do companies request to social media service providers?

I can only talk about LINE's story. Most companies request to learn how to integrate chatbot with their service. Since they are new to chatbot, they don’t know what type of solution is best for them. LINE provides a consultation service for customers to implement chatbots for their services on our platform.

Is AI taking over chatbot?

Yes and No. Developing AI to do everything is extremely difficult. Plus, you'll need to teach the AI service to respond to questions that you might not even have expected. Of course, machine learning is advancing, and in the future, adopting AI could become easier. But AI is yet to respond to every question. Currently, companies with data and resources do use AI on their chatbots. At LINE, we provide a hybrid solution in which an AI agent responds to the user first, and if the user feels that the response was inadequate, the user can switch the respondent from the AI agent to a human operator.

Does chatbot work for B2B solution?

It will depend on what solution you are trying to provide through a chatbot. We've seen chatbots for B2B, to manage store inventory and to manage rosters. LINE's practice is more of a B2B2C than B2B; companies use LINE to provide customer service to their users.

As I close

I wished I had more time for Q&A to listen more from the participants, but overall, it was great having an opportunity to listen to other people's thoughts. I hope to have such session again in the future.