Chatbots in Ecommerce
Categorized as : Ecommerce
Ecommerce chatbots are gaining popularity and considered to replace humans in the area of shopping consulting just in several years supposedly.
There are two reasons for this phenomenon, economic and technological. The chance to build and implement human-like digital consultants appeared with the development of the corresponding technology and the process of users migration from applications to messengers.
The messenger format of communication gave rise to the idea of chatbots and later their factual appearance. With the development of technologies chatbots will become smarter and be able to provide high-level consulting and selling services.
The second reason is a trivial opportunity to save on salaries and offer customers the 24/7 service format. And one more thing, chatbots are much simpler than applications and cost less, much less.
media/uploads/2017/08/Audio_Version_Chatbots_in_Ecommerce.mp3
Not all of them, but mostly modern chatbots are not so smart yet and work often as programmed context models able to recognize only several predefined answers or guide customers through predefined question/answer trees. Still, even with this limited functionality they are able to run their tasks successfully.
Currently chatbots are helping customers to choose clothes, order pizzas, book tickets, select music, etc., but they are not only shopping consultants. They can also schedule a meeting, find weather forecasts and proper news, plan your finances, and even be your personal assistant and friend.
Chatbots live where they can serve the largest number of people and now they mostly reside in messengers, but they can be incorporated into almost any software applications.
Perhaps you’ll be surprised, I am anyway, there are about 1200 bots for Facebook Messenger, according to ChatBottle, and even more for all other social media and platforms. All those chatbots differ from each other and knowing their types is very useful for every online merchant. You need to definitely understand all the functionality opportunities and technical aspects in case you want to use or even build own ecommerce chatbots.
By the power of intelligence we can distinguish script bots, smart bots, and intelligent conversational agents.
Script Bots
Script bots are the simplest chatbots operating according to particular scripts and made for certain platforms usually, e.g. Facebook, Twitter, Slack, Telegram, Kik, etc. Some of them exploit Natural Language Processing in order to make interactions and conversations more natural, but usually can recognize only a few words closely related to their main field of activity.
Currently script bots are the most numerous and perhaps will remain the main type of chatbots for ecommerce in the nearest future.
Smart Bots
Smart bots rely on Artificial Intelligence and provide much better user experience since the conversation they offer is much similar to human talks. But, the whole process involving smart bots is not fully automated and may require some human interventions.
For example, if you need to order some food, the bot will ask you about your needs and preferences, offer you particular places, menus, price options and will confirm your order. However, a human agent still needs to make a phone call to the restaurant or food service and make a reservation or place your order.
Intelligent Agents
This type of really clever boots are fully self-sufficient and based on Machine Learning and Artificial Intelligence technologies. Siri, IBM Watson, and Amazon’s Alexa are good examples of such intelligent agents.
Despite intelligent agents are not yet able to handle any types of customer experience due to existing technology limitations, but they certainly look as the most promising cohort of bots in the next decade.
By the number of available functions we can distinguish general and special chatbots. Most Intelligent agents are general bots performing multiple functions, while most script and smart bots are designed to provide or optimize particular operations.
The above types of bots are two sides of the same coin, in fact. For example, general bots can greatly understand and provide you with some general advice, but lack the functionality necessary to complete particular actions.
As far as chatbots are quite complex software they have a lot of different aspects able to become a basis for consideration and comparison. For example we can classify message, voice, and combined bots, or you can identify financial or travelling assistants by the field of implementation.
As for me, most interesting chatbots are the ones that have some skills to understand natural speech. Without this option chatbots look silly and their advantages over standard applications are not obvious.
The main goal of chatbots for ecommerce is to sell products and services, but they can also be useful for other related activities, including customers services, marketing, PR, and branding.
However, chatbots are not suitable for all online stores and customer audiences. As far as most people are traditionalists they prefer well-known ways of purchasing. According to Sumo Heavy, only 10% of the US customers aged between 18 and 54 ever used chatbots and half of them replied that would never do this again. The only group, which is quite comfortable with chatbots is Millennials. 67% of them would likely purchase from chatbots, according to eMarketer.com.
So, you should certainly consider those age limitations in case you are going to take advantage of chatbots for sales.
If you still hope to use a chatbot for your daily sales routines, you need to evaluate all the pros and cons and plan the whole procedure carefully. There are several questions to be answered before you start building your first chatbot.
We’ve already told that some customers do not like chatbots and prefer more traditional ways to make their purchases. But, if a certain portion of them are Millennials you can surely proceed to the next questionnaire item.
Chatbots may appear on any platform, and for now you can discover them in Facebook Messenger, Telegram Twitter, Kik, Slack, Skype, Viber and other messengers and services. So, make sure that your customers use one or several of these or other platforms and go ahead.
According to the research by Mindbowser and Chatbots Journal, 61% of respondents use IBM Watson for bot building, 41% of businesses prefer Microsoft Bot Framework and almost 45% of developers like Wit.ai. But, if you have no programming skills you’d better take advantage of the following chatbot platforms: ChattyPeople, Chatfuel, Bosify, and other ones. Some services like Facebook Messenger and Telegram promote their own platforms that help building bots successfully, so the choice is quite wide.
Chatbots are comprised of numerous design and functionality elements, including buttons, menus, cards, forms, etc. The main element of each chatbot is conversation, of course, and you need to take into consideration multiple possible questions or answers coming from customers based on most common interaction patterns. Finally, you need to outfit your chatbots with checkout forms and let customers purchase right here with no redirects.
Of course, this topic is a subject for multiple particular articles or a book, so we are not able to cover it here even a bit completely. And remember that this stage almost entirely determines the success and efficiency of your chatbot :)
Creating a fully functional and efficient chatbot is still not the end. The next step is to allow customers to know about your chatbot and let them discover and take advantage of the new way of shopping. The more different ways of promotion you involve the better is the outcome you get.
First of all do not forget about like and share buttons to be included into the interface of you chatbot. Then promote chatbot on your store as an alternative way to make purchases from you. Share the information about it in your official social accounts, place your chatbot into dedicated catalogues and marketplaces, use the power of content and opinion leaders in this field, exploit different languages for chatbot conversations.
Finally, you need to know how your chatbot performs, what the common user behavior is within conversations, what the problems they face and when, what the conversation sentiment is, etc. So, the available analytics platforms are not yet so powerful as Google Analytics, but they cover most urgent needs and sometimes even more. The following products are among the best examples in this sphere: DashBot, BotAnalytics, Facebook Analytics, BotMetrics, etc.
The ultimate goal of chatbots is to look like a human and provide the same quality and value of conversation. You may think that it sounds like a global conspiracy of machines. Well, we don’t know exactly, but as usually it’s promoted under the slogan ‘For the good of mankind’.
Of course, you will never trust a bot to the same extent as a human operator and will hardly ask him to ‘share own impressions’. However, just soon bots will become indistinguishable from us, so soon you will probably never be completely sure about the talker on the other side...
There are two reasons for this phenomenon, economic and technological. The chance to build and implement human-like digital consultants appeared with the development of the corresponding technology and the process of users migration from applications to messengers.
The messenger format of communication gave rise to the idea of chatbots and later their factual appearance. With the development of technologies chatbots will become smarter and be able to provide high-level consulting and selling services.
The second reason is a trivial opportunity to save on salaries and offer customers the 24/7 service format. And one more thing, chatbots are much simpler than applications and cost less, much less.
media/uploads/2017/08/Audio_Version_Chatbots_in_Ecommerce.mp3
Modern Chatbots
Not all of them, but mostly modern chatbots are not so smart yet and work often as programmed context models able to recognize only several predefined answers or guide customers through predefined question/answer trees. Still, even with this limited functionality they are able to run their tasks successfully.
Activity Fields
Currently chatbots are helping customers to choose clothes, order pizzas, book tickets, select music, etc., but they are not only shopping consultants. They can also schedule a meeting, find weather forecasts and proper news, plan your finances, and even be your personal assistant and friend.
Places of Residence
Chatbots live where they can serve the largest number of people and now they mostly reside in messengers, but they can be incorporated into almost any software applications.
Types of Chatbots
Perhaps you’ll be surprised, I am anyway, there are about 1200 bots for Facebook Messenger, according to ChatBottle, and even more for all other social media and platforms. All those chatbots differ from each other and knowing their types is very useful for every online merchant. You need to definitely understand all the functionality opportunities and technical aspects in case you want to use or even build own ecommerce chatbots.
By the Level of Intelligence
By the power of intelligence we can distinguish script bots, smart bots, and intelligent conversational agents.
Script Bots
Script bots are the simplest chatbots operating according to particular scripts and made for certain platforms usually, e.g. Facebook, Twitter, Slack, Telegram, Kik, etc. Some of them exploit Natural Language Processing in order to make interactions and conversations more natural, but usually can recognize only a few words closely related to their main field of activity.
Currently script bots are the most numerous and perhaps will remain the main type of chatbots for ecommerce in the nearest future.
Smart Bots
Smart bots rely on Artificial Intelligence and provide much better user experience since the conversation they offer is much similar to human talks. But, the whole process involving smart bots is not fully automated and may require some human interventions.
For example, if you need to order some food, the bot will ask you about your needs and preferences, offer you particular places, menus, price options and will confirm your order. However, a human agent still needs to make a phone call to the restaurant or food service and make a reservation or place your order.
Intelligent Agents
This type of really clever boots are fully self-sufficient and based on Machine Learning and Artificial Intelligence technologies. Siri, IBM Watson, and Amazon’s Alexa are good examples of such intelligent agents.
Despite intelligent agents are not yet able to handle any types of customer experience due to existing technology limitations, but they certainly look as the most promising cohort of bots in the next decade.
By the Number of Functions
By the number of available functions we can distinguish general and special chatbots. Most Intelligent agents are general bots performing multiple functions, while most script and smart bots are designed to provide or optimize particular operations.
The above types of bots are two sides of the same coin, in fact. For example, general bots can greatly understand and provide you with some general advice, but lack the functionality necessary to complete particular actions.
Other Types of Bots
As far as chatbots are quite complex software they have a lot of different aspects able to become a basis for consideration and comparison. For example we can classify message, voice, and combined bots, or you can identify financial or travelling assistants by the field of implementation.
Chatbots for Ecommerce
As for me, most interesting chatbots are the ones that have some skills to understand natural speech. Without this option chatbots look silly and their advantages over standard applications are not obvious.
The main goal of chatbots for ecommerce is to sell products and services, but they can also be useful for other related activities, including customers services, marketing, PR, and branding.
However, chatbots are not suitable for all online stores and customer audiences. As far as most people are traditionalists they prefer well-known ways of purchasing. According to Sumo Heavy, only 10% of the US customers aged between 18 and 54 ever used chatbots and half of them replied that would never do this again. The only group, which is quite comfortable with chatbots is Millennials. 67% of them would likely purchase from chatbots, according to eMarketer.com.
So, you should certainly consider those age limitations in case you are going to take advantage of chatbots for sales.
How to Create Own Chatbot
If you still hope to use a chatbot for your daily sales routines, you need to evaluate all the pros and cons and plan the whole procedure carefully. There are several questions to be answered before you start building your first chatbot.
Is it Going to Be Really Useful for Your Current Shopping Audience?
We’ve already told that some customers do not like chatbots and prefer more traditional ways to make their purchases. But, if a certain portion of them are Millennials you can surely proceed to the next questionnaire item.
What are the Most Suitable Places for Your Chatbots?
Chatbots may appear on any platform, and for now you can discover them in Facebook Messenger, Telegram Twitter, Kik, Slack, Skype, Viber and other messengers and services. So, make sure that your customers use one or several of these or other platforms and go ahead.
What are the Best Platforms for Building Chatbots?
According to the research by Mindbowser and Chatbots Journal, 61% of respondents use IBM Watson for bot building, 41% of businesses prefer Microsoft Bot Framework and almost 45% of developers like Wit.ai. But, if you have no programming skills you’d better take advantage of the following chatbot platforms: ChattyPeople, Chatfuel, Bosify, and other ones. Some services like Facebook Messenger and Telegram promote their own platforms that help building bots successfully, so the choice is quite wide.
Designing a Chatbot
Chatbots are comprised of numerous design and functionality elements, including buttons, menus, cards, forms, etc. The main element of each chatbot is conversation, of course, and you need to take into consideration multiple possible questions or answers coming from customers based on most common interaction patterns. Finally, you need to outfit your chatbots with checkout forms and let customers purchase right here with no redirects.
Of course, this topic is a subject for multiple particular articles or a book, so we are not able to cover it here even a bit completely. And remember that this stage almost entirely determines the success and efficiency of your chatbot :)
Promoting a Chatbot
Creating a fully functional and efficient chatbot is still not the end. The next step is to allow customers to know about your chatbot and let them discover and take advantage of the new way of shopping. The more different ways of promotion you involve the better is the outcome you get.
First of all do not forget about like and share buttons to be included into the interface of you chatbot. Then promote chatbot on your store as an alternative way to make purchases from you. Share the information about it in your official social accounts, place your chatbot into dedicated catalogues and marketplaces, use the power of content and opinion leaders in this field, exploit different languages for chatbot conversations.
Tracking Chatbot Analytics
Finally, you need to know how your chatbot performs, what the common user behavior is within conversations, what the problems they face and when, what the conversation sentiment is, etc. So, the available analytics platforms are not yet so powerful as Google Analytics, but they cover most urgent needs and sometimes even more. The following products are among the best examples in this sphere: DashBot, BotAnalytics, Facebook Analytics, BotMetrics, etc.
Conclusion
The ultimate goal of chatbots is to look like a human and provide the same quality and value of conversation. You may think that it sounds like a global conspiracy of machines. Well, we don’t know exactly, but as usually it’s promoted under the slogan ‘For the good of mankind’.
Of course, you will never trust a bot to the same extent as a human operator and will hardly ask him to ‘share own impressions’. However, just soon bots will become indistinguishable from us, so soon you will probably never be completely sure about the talker on the other side...