How does e-commerce chatbot work?

by Anna B.
by Anna B.
Business Analyst
20 Jul 2020
Reading Time: 12 minutes

Chatbots have long been integrated into online retailing. With their help, you can shop online, book tickets, get personalized offers – all this around the clock. Smart assistants are a real catch for business owners. Their productivity never drops; they are always fast and never get tired.

If you own an online shop, you should think of building a chatbot. It will help you automate the work of the contact center and win customer loyalty. In this post, we share our expertise on the subject. If you are interested in eCommerce chatbot development, grab a seat.



What is an e-commerce chatbot and how does it work?


A chatbot is a program that communicates with users in a chat simulating human behavior. It can work on any well-known channel, such as Facebook Messenger, Telegram, Viber, etc. Today, chatbots are taking the lion’s share of operators’ work. The entrepreneurs invest in their development because it is more cost-efficient than hiring a human employee.

E-commerce chatbot use cases

Chatbots are used in many business areas. They are:


  • Online stores. Bots help choose the product and make payment.
  • Cafes and restaurants. Chatbots book a table and accept an order.
  • Insurance. Bots guide through the contract terms and apply for a refund.
  • Banking. Chatbots give financial advice and suggest relevant banking products.
  • Healthcare. Bots arrange an appointment with a doctor and conduct an initial survey.
  • Public sector. Chatbots inform about tariff change and handle appeals from tenants.


A chatbot creates an illusion of live communication and attracts consumers to the brand. If you want to expand your customer base and provide enhanced service to existing clients, it is a perfect solution.


JoinSoft can help you grow business by delivering a chatbot technology for customer service. Have questions?



Why can’t you ignore E-commerce chatbot: statistics


Today customers are very demanding and expect their requests to be processed almost instantly. Help Scoup research states 66% of surveyed believe that valuing their time is a key part of the quality service. Yet, only 56% of queries in eCommerce are now processed within 10 minutes.


Response time in online retail business



Chatbots solve the problem with instant responses. Moreover, they can handle as many calls as you like. Being emotionally unaffected, they provide equally good assistance at any time of day and night.


With this in mind, many sellers have already recoursed to robots. As of today, a chatbot technology in retail is no longer a novelty. According to Business Insider, the chatbot market size was $2.6 billion in 2019. This figure is expected to reach $9.4 billion by 2024.


The prospects of chatbot market size



Smart assistants can reduce the cost of retailers by 30%. Also, 80% of entrepreneurs will have created a chatbot for eCommerce by the end of 2020. Such an expansion would not have been possible without a positive response from consumers. To date,


  • 34 % of retail clients would prefer chatbot over the human operator
  • 40% of customers don’t care whether it is a bot or human manager who assists as long as their problem is solved
  • 58% of users admit chatbots have changed their attitudes to customer service for the better


User attitude to chatbots



How do retail bots work?


E-commerce chatbot features depend on business needs. At JoinSoft, we discuss them beforehand. When all the details are agreed with the client, we proceed to development.


There are primary and advanced chatbot features. We recommend starting with a simple version and upgrade to a complex one later.


Provide information

Bots answer questions in the form of text and graphic content. Besides, they can send product catalogs, geo-coordinates, links to third-party resources, and other useful information. Fast response and relevant content contribute to customer satisfaction and increase the chances of returning to your shop.


Identify needs

Bots help identify customer needs by requesting further details and suggesting options. For example, if the user cannot find the thing he is looking for, the bot may ask about the purpose of the purchase. Based on the answer, it may offer alternative products in a given price range.


Segment customers

Some users buy expensive goods; others prefer cheaper products. Nine-to-fivers look for office wear, gym fans – for sports things. Bots will help you segment users. Based on the preferences, you can organize targeted mailing and increase sales.


Interact with human agents

If the bot faces a complex issue, it passes it to a human colleague. Such a system frees your employees from routine operations and allows them to concentrate on non-trivial queries. Such a system reduces response time and delivers classy service.



Benefits of e-commerce and retail AI technology


Do calculations

The bot can save the data received in the dialogue to Google spreadsheets or other similar tools. Also, it can do basic calculations. For example, knowing a discount card number, it can estimate the payable amount considering the personal discount.


Accept payments

You can integrate the chatbot with payment systems and accept payments directly in the messenger. The payment details are set in advance in the bot, so there is a bare chance to face an error. Note, you will have to pay a fee to the payment service.


Connect with third-party services

You can teach the chatbot to use external APIs such as CRM, weather services, your own website database, etc. For example, a bot may save the data directly to your CRM instead of Google spreadsheet. Also, it can pull the product price out of the supplier’s database.


Speak intelligently

Advanced bots use NLP to communicate with clients in many languages. They understand the context, ask related questions, and clarify intentions. Such bots break linear thinking and do proactive research.


e-comerce Chatbot



Best e-commerce chatbot examples


Chatbots in retail have long proved its value. They create a sense of live dialogue, which adds fun to the shopping experience and thereby boosts customer satisfaction. The world-known brands, such as H&M, Tommy Hilfiger, Sephora, Lidl, and Whole Foods are growing their businesses with the help of AI and ML. Let’s see how they use chatbots and why they fit them well.



In 2016, H&M launched a support bot based on Kik messenger. The bot helps browse through the product catalog, share favorite items, and make purchases directly from the chat. In addition to performing basic functions, it picks up things that suit the user’s style and gives advice, acting as a personal stylist. Knowing customer habits the bot creates unique style profiles. Also, it helps compose custom outfits. The users can view each other’s outfits and buy the ones they like.





Tommy Hilfiger

TMY.GRL mainly acts as a guide to brand collections. At the beginning of the chat, it offers three options: find a new look, view the collection, go behind the scenes. The first two involve viewing the products as well as selecting outfits depending on the user’s preferences. For example, TMY.GRL may ask about your favorite colors and usual style. Based on the answers, it will suggest the most suitable looks. The third option means viewing selected items on models from the latest fashion show.


Chatbot Tommy Hilfiger




Sephora has three bots. One of them runs on Kik, another two – on Facebook Messenger. Each bot does specific tasks, which aim to improve the quality of service. The Kik bot helps users pick up goods and make purchases. Once the user enters the messenger, it offers a quiz to learn the customer’s preferences. Further, it provides information on the products and suggests how to combine them with an individual style. Facebook bots help book the meeting with a human beauty expert and get the right colors based on the liked photos. For example, a user may scan a picture of a celebrity, and the bot will detect what tone of lipstick she uses.


Chatbot Sephora, Kiki




Lidl’s Facebook bot is called Margot, and it is a real wine expert. Not only helps it to choose wine, but also advises on the best food pairings. Margot is very communicative. It offers interactive wine quizzes and uses emojis to spice up the chat. Besides, it guides through wine types by country, grape, color, price, and other parameters.


As the Covid19 pandemic started, Lidl launched a WhatsApp bot that determines the best time for shopping. For example, you enter the estimated date and time of shopping, and the bot tells you whether the store will be jam-packed or not. It helps avoid crowds and maintain social distance.


Chatbot Lidle



Whole Foods

Facebook-based bot of Whole Foods searches for nearby stores, pulls locations, and suggests the products to buy. Moreover, it advises recipes and product sets depending on the user’s preferences. For instance, the user may ask for seafood or vegetarian dishes. The bot will give ready recipes with step-by-step guidance, pics, and nutritional information. Besides, it will tell you the nearest stores where you can find necessary products. You may ask the bot to choose dishes for a specific diet. Also, it can send recipes based on emojis.


Chatbots Foods market



How to make a retail bot


Chatbot development for retail is a smart investment that pays off soon. There are two ways to make a bot:

  • use bot-building platforms
  • develop a bot from scratch



Bot-building platforms


The first option is more suitable for small and medium-sized businesses. It requires minimal investments and coding skills. In general, you will build a bot from predefined templates. Usually, you can do it without the help of a developer. The most popular services for making chatbots are:

  • Chatfuel
  • IntelliTicks
  • Wit.ai
  • Motion.ai
  • Mobile Monkey


When you can use bot-building platforms

  • a standard set of options offered by the service is enough for you
  • you don’t want to invest in development and wait until the bot is made
  • you need a fast but simple solution


When you should not use bot-building platforms

  • you prefer a tailor-made solution with unique features
  • you don’t want to pay a regular fee which is often really high


Custom development

This option is the right choice for mid-sized and large scale companies. It allows you to create a bot that 100% meets the business needs and contributes to economic growth. To build a messenger chatbot for eCommerce, you need to hire a team of professionals. Get ready to spend some time discussing the project and wait a few months until it is delivered.


When you should build a custom-made chatbot for eCommerce

  • you need a unique solution with highly customizable features
  • you do not want to depend on third-party services
  • an ongoing monthly fee is not in your cost list


When you should not build a custom-made chatbot for eCommerce

  • ready-made solutions cover the needs of your business
  • you are not ready to make a significant one-time investment


Don't know where to start eCommerce chatbot development? We will advise the best solution for your business.



E-commerce chatbot development process

The development process differs from project to project, yet there are typical steps that we usually take:



Backend covers a chatbot business logic. It sets up microservices to process messages and establishes conversational mechanisms. The technology stack of the server-side depends on the project specifics and the developer’s preferences. At JoinSoft, we often opt for Ruby due to its simplicity and intuitiveness of the code. For the bot core, we choose Node-RED. AWS helps us ensure the operational stability of the bot.


The approximate time for the server setup is 4 hours.



Communication channels

Communication channels are the messengers through which the bot communicates with customers. The most common ones are Facebook Messenger, Kik, WhatsApp, Telegram, Slack. You can choose as many channels as you like. Note, each messenger has specific features of integrating a chatbot. Nevertheless, for each of them, the developers have to set up an endpoint to accept and post messages based on security token authorization.


Integration of one channel may take up to 15 hours.




NLP stands for Natural Language Processing. It empowers the bot to understand human speech and keep up the talk naturally. Many companies offer ready-made NLP services. The chatbot developers only need to select a proper service and connect it with the system. At JoinSoft, we work with Kore.io, IBM Watson, Google Dialogflow, and others. They enhance text recognition with advanced technologies, such as sentiment analysis, word associations, topic tagging, and visual context graphs.


Setting up the NLP mechanism may take up to 80 hours.



Conversational intelligence

Speech interpretation is half the battle. A more significant challenge is to teach the bot to speak intelligently. To that end, developers need to create an algorithm for each conversation and clear navigation so that the user can quickly start a chat. It is necessary to develop decision trees, chat workflows, and deep learning algorithms to manage communication flows. It is a complex task that may take the lion’s share of the overall work.


The estimated time is up to 120 hours.




Customer service chatbots for retail help clients choose products, make payments, set up a delivery, process refund, and do other related things. The bot should access various services to complete these tasks. For example, it needs to connect with the product catalog and payment service provider to set up a purchase. Also, it has to reach the shipment company API to offer delivery options. The time for implementation of this option depends on the number of services you need to connect.


In general, it may take up to 90 hours.



Control panel

The control panel helps manage chatbot activity and do analytics. You can view the chats history, trace error logs, and see some users’ data. Such information lets you eliminate possible mistakes and make the right decisions in time. A control panel is a handy tool for building an effective business strategy. Being aware of communication nuances, you can adjust the bot behavior in time.

CP implementation takes up to 40 hours.



How much does it cost to develop a retail chatbot?


The chatbot development price consists of:

  • business analysis and planning
  • direct development costs
  • UI/UX design
  • testing and maintenance


It greatly varies depending on the region of the development team. The most expensive developers work in Australia and North America, the cheapest – in India.


Development cost around the globe


If you are looking for the best price-quality ratio, consider outsourcing a project to Eastern Europe. In this region, IT experts do classy work for a fair price. Let’s see a detailed estimation of eCommerce chatbot development in Eastern Europe.


BA and planning Development UI/UX design Testing and maintenance Total
Hours 20 350 50 50 470
Money $12-24K



Retail bot software with Joinsoft


If you decided to outsource custom development, you might be interested in how it all goes. At JoinSoft, we start with a rough estimation of the project. We usually give such an assessment based on the previous experience.


Further, we do a detailed estimation. It describes:

  • team members
  • technology stack
  • targeted platforms
  • development steps
  • deadlines


After agreement on the above plan with the client, we proceed directly to an implementation process. At JoinSoft, we adhere to Agile development. At the end of each sprint, we report on the work done.


If you are thinking of eCommerce chatbot development, drop us a line. We will suggest the best possible solutions based on the specific nature of your business.



Rate this article!

Average rating 5 / 5. 8

No votes so far! Be the first to rate this post.

You may also be interested in
Main CRM
Ever noticed how easy it is to get lost searching a suitable tool for your firm? You can spend hours browsing the web and end up confused. According to Statista, CRM usage will skyrocket in the near future. Get this....
by Polina S. & JoinSoft Team
14 Aug 2020
Main_How to Integrate Payment Gateway into Wordpress Website?
According to Statista, there are 3 most popular ways to pay for online purchases:...
by Anna B.
21 Apr 2021
Types of Errors in Programming
Programming is a complex process requiring a great attention to details. Making mistakes here is, probably, inevitable. Some of the bugs, the most unpleasant ones, are logical by nature. I...
by Anna B.
21 Jun 2021