What is a chatbot? 【English Ver】

2016.08.17. by Victoria Safonova

The opening of Messenger bot lab


Chatbots have suddenly become one of the most discussed things in the tech industry. They unlock the ability to provide interactive communication, similar to communication with human customer service, but for much cheaper than call centers.
Bot on Messenger could redefine how businesses sell products and services. For instance, a chat bot could inquire about your product or service criteria, and offer you relevant options, all at the tap of a keyboard. Additionally, it could also learn customer likes and dislikes, and provide personal suggestions based on them.


Just recently Hana-cupid Co.,Ltd., a popular flower delivery company in Japan, has launched its bot on Facebook Messenger.
To make this possible, Hana-cupid teamed up with GMO Payment Gateway, Inc. GMO Payment Gateway was early to adopt a robust support system via the “lab” approach.



■Facebook Messenger Bot / Hana-cupid for businesses


■The website of Hana-cupid for businesses



The Hana-cupid bot is currently offering its services only to corporations. However, the company is planning on launching a chatbot for individual customers as well.

You can activate the bot via the Hana-cupid Facebook page




What actually is the Bot for Facebook Messenger?

The number of Messenger users



On July 20th, 2016, Facebook announced that the number of monthly active users of Messenger had exceeded one billion. Messenger has become the third application ever that is used by more than one billion users. Currently, the number of Facebook users is more than 1.6 billion people.


The future of Facebook Messenger


Facebook announced a launch of Messenger chatbot at the F8 conference on April of 2016. This is made possible through the Messenger Platfrom’s new Send/Receive API. Bots will support not only sending and receiving text, but also images and interactive buttons containing multiple “call-to-action” functions. Businesses will be able to effectively interact with people by offering products and services that they want to receive.


Facebook Messenger in Japan and overseas: Difference


Due to the buzz around chatbot, bots for Messenger, Line, Twitter, etc., have also been growing with each passing day overseas. However, there are only few companies in Japan that are currently is using it. Thus, we believe if overseas users can make a restaurant reservation, book a haircut, and review an e-commerce order through Messenger bot, Japanese user should be able to do the same. Unfortunately, the Japanses market seems to be very slow in adopting this new trend.


However, GMO Payment Gateway announced the launch of Inbound Marketing (including Facebook advertising) last year. Although it came with challenges, Hinomaru Limousine Co., Ltd. became the first one to make a successful trial.




※Although different from Chatbot, users of Hinomaru [LIMOTAXI] service became able to pre-book the Limotaxi from Narita or Haneda airports, to their hotels or other destinations by using Messenger.


Below is the success results achieved by Hinomaru Limousine and posted to the Facebook Business page: https://www.facebook.com/business/success/hinomaru-limousine


Features of facebook Messenger Bot



As of now (August, 2016), the main feature of Messenger bot is ‘tapping’. The customers are able not only to send a request to chatbot by text, but they can also tap the option in a menu within the chat screen. We believe it can be a revolutionary purchasing experience for customers, which is the motivating factor behind our new lab project.



Conversational Commerce


Using chat, messaging, or other natural languages interfaces (i.e. voice) to interact with people has become the beginning of a new movement in technology, as mentioned by Chris Messina of Uber as “Conversational Commerce”.

Reference: https://filipmaertens.com/2016/04/10/chatbots-from-conversation-to-conversion/


In his article, “The 5 branches of conversational commerce: A guide to the bot world”, Jiff Lawson of Twilio introduced five different trends converging on conversational commerce.



These include:


No.1 Advanced notifications

Every day we receive timely alerts via SMS, push notification, or even email by way of interacting with companies and apps. Whether it is a delivery confirmation, or a message that your flight is delayed, proactive notifications are a simple and effective means of improving our interactions with companies.

However, the nature of these notifications are changing. Notifications can include multimedia, maps, directions, contacts, boarding passes, and more. Essentially, they are becoming an invitation to a conversation.


No.2 Bots

There is a wide variety of ways that bots may evolve to do our bidding:


A_Command line bots

For those of us who love Unix, these bots put a command shell in everything. For instance, in Slack, bots have been built for fetching Digg (content), ordering an Uber (transportation), or even pulling data from Google Analytics (data), just to name a few.



Chatbots are essentially a two-way conversation with a computer — the Turing test in action. Recently, Chatbots have received a massive amount of attention, both positive and negative. This is largely due to giant leaps in deep learning. In business, the vision is that chatbots will enable a company to message one-on-one with customers without increasing human overhead costs.


C_Personal assistants

These are bots that can accomplish tasks for a person — answer questions and respond to commands. These bots know your preferences, integrate with all your information, and act on your behalf — essentially replacing human work. In the consumer world, we’ve already seen some success with personal assistants through Amazon Alexa, Apple Siri, and Google Voice Search.


No.3 Chat in apps

Chat in apps provide a conversational functionality embedded within apps. Apps are becoming a smarter and more valuable tool because they inherently contain the context of information in the application you are using. For example, if you’re in the Uber app and want to call your driver, the driver already knows your name, and where you are — which is a huge improvement over calling a taxi dispatcher.


No.4 Apps in chats

This is about bringing application functionality into chat, as we have seen done in China with WeChat. In this model, chat is connecting people and businesses, with a flexible in-chat interface that can be programmed like an app to enable a wide variety of use cases.


No.5 Humans chatting with each other

The last category of messaging is simply humans conversing with one another. In short, humans want to communicate with businesses the same way they communicate with other humans.




Chatbots, from conversation to conversion: [What kind of changes can happen with bots?]


Although many people think that the purpose of using chatbots is to maximize the conversation rate (CVR), the developer’s opinion is that it can maximize the conversion (CV). This was mentioned in an article by Filip Maertens:

“Being able to elicit additional information from a human user, through a friendly conversation, enables algorithms to tap into new sources of data that can better profile said user, hereby greatly enhance propensity models and as a result drive higher conversions.”

Reference: https://filipmaertens.com/2016/04/10/chatbots-from-conversation-to-conversion/


Before CV = building bots to gain more new customers

・Command line bots.  Providing customers with information about products, and inventory conditions, all triggered by conversation with the chatbot.

・Unique Command line bots.  With the ability to analyze and understand customer’s needs and preferences, chatbot can regularly send push notifications.

・Command line bots. A  bot admin can set Lead Generation program to analyze lead campaigns. The model can be customized to meet each organization’s needs.

(After texting such information as「Name、Company name、Male address、Home address、Telephone number tec.」within chatbot, a customer receives an immediate email containing it. ※In this case, this customer must agree on the condition of providing personal information before he or she buys a product, and starts using a service.


After CV = bots optimizing customer retention, cross-sell, up-sell, auto-follow

・Delivering only necessary information to users in an appropriate time and case (delivery status, service guide, extraordinary assistance plan etc).

・Delivering the information about special products or flash sales via push notifications PUSH.


※We also think about possible bots such as Before CV+After CV.



Chatbots, from conversation to conversion : [How can a customer discover bots?]

Since searching in an app store has been transferring into searching within a chat or within a message, different approaches to help users discover their bots can be very useful.

Below is some of them.


01_ Link to Messenger app from e-commerce Facebook page

The e-commerce business can add to its advertising post on Facebook a direct link including「m.me/○○○」.


02_ Link to Messenger bot via DSP, Native AD, Display Network advertisements

This method can’t be implemented in DSP, Display Network, but can be used in Lookalike and retargeting.


03_ The link to the bot can be set on a website as a pop-up window

Pop-up windows can be used to invite customers not only to check out the mail magazine, but also can be invitations to Line, Facebook, and Twitter etc.


04_ Using Messenger code, similar to QR code

It can be printed on a business card or flyer, as well as added to a website or a profile picture of the corporate Facebook page.


Above is Messenger code of Hana-cupid. However the profile picture of Hana-cupid is without it.



Command line bots. How to build a chatbot


Methods of search are continuing to grow. We can already search information not only by typing key words, but we are able to use voice search and image search as well. We can see that the system replies to both our keywords, and to our tap-commands or comment-commands within chat in Messenger.

The manager of listing ads has to group thousands of key words and then consider an appropriate ads slogan and corresponding lending pages. As we have noticed, the listing structure is very similar to a bot structure (Please refer to charts below).




Chatbots can be free to communicate with customers by texting and also sending them various links, pictures, data, maps and etc.


messenger bot structure



We believe that it will be necessary to verify on any occasion if settings and operation of Messenger Bot contributes to CV or not by turning the circle above.


We also believe that, being on the pre-stage of having personal assistant bots, companies and/or people will need professionals who can offer an LTV (Customer lifetime value) scenario considering chatbot settings, push notifications, ways to discover bots, and operate this scenario.


We are quickly approaching the day when we will be able to carry out business transactions like this:


Chat bot_example



※We are currently (August, 2016) taking different approaches as a lab to improve early-stage settings of Messenger bots, and operation assistance. You can contact us if you want to learn more about setting stages before launching a service. Please keep in mind we may not be able to answer to all your questions and requests. Thank you for your consideration.



The following two tabs change content below.
Victoria Safonova

Victoria Safonova

2014年ロシア国立人文大学卒業後、東京大学大学院に入学。卒業後の2016年に新卒としてGMOペイメントゲートウェイ株式会社に入社。経営企画部 新領域促進グループ所属。
Victoria Safonova

最新記事 by Victoria Safonova (全て見る)