Chatbot for Retail

November 1, 2019
By Botpress Team

A chatbot for retail can be extremely useful but at the same time needs to be used with caution for actual shopping functionality.

As with any product or service, the chatbot is useful for any customer support related functions where it can be used to answer repetitive simple questions and escalate complex questions to customer service agents.

The question is to what extent should it support other customer service functions and how should it do this.


It should be obvious that a chatbot is not the best interface for searching and filtering large numbers of products.  A graphical interface is a much better use case here.  A graphical interface can allow the user to set a large number of criteria to narrow down the search.

A chatbot is even a worse use case if the customer is looking for multiple items, such as for a supermarket.  Again a graphical interface is a better fit.

There may be some use cases where the chatbot is a better channel than a graphical interface for retail.  For example:


  • The customer needs a new item of something and either knows the exact item they need or don’t care about which item is purchased.  
  1. This could be because it is a low-cost commodity item that is not worth the time to search for.
  2. This could be because they are not price-sensitive and want pure convenience and trust the chatbot to select an item of reasonable quality and price (the chatbot could use simple heuristics such as overall ratings and number of verified ratings).
  3. This could be that they view the item as arbitrary, in the case of a present for a kids’ birthday party.
  • Same as the above, but the customer doesn’t have access to a screen at that moment.  For example, they are brushing their teeth and suddenly see they need more toothpaste.
  • The chatbot is backed by a professional shopper for the item in question and the customer has confidence that the chatbot will make the best decision.
  • The chatbot gives the customer a smaller list of highly curated items that fit the customer’s exact needs, like a list of high-end corporate gifts.


Even for a supermarket type use case, it is also true that a chatbot, assuming it performs well enough, could be a useful way of creating the initial list of items you want to buy if we assume the customer is using a voice bot.  

For example, the customer could just state all the items they wanted to buy and the chatbot could create a preliminary list that they could refine as necessary.  They could also speak to the chatbot while they are in front of the screen that shows the items they bought last time and say things like “I don’t need milk, eggs and tomatoes”.

It should be noted that it is also a possibility to have a two-way interaction between the chatbot and the graphical interface, where the chatbot can provide immediate assistance with issues or guide the user to the place they need to be on the graphical interface.

Aside from the shopping functionality, the chatbot could be useful in performing many functions that are related to shopping, such as organizing the delivery, getting information from the customer, tracking orders and managing loyalty points.  

Marketing with chatbots is a powerful way of using bots especially if combined with useful content.  Customers can be given time-sensitive promotions, with the ability to act on the promotions by buying something or taking some other action from directly within the chatbot.

As always, the use case in question needs to be carefully analyzed to work out if the chatbot is the best channel for the use case in question versus a graphical interface, and if it is whether it should be a scripted bot or intelligent bot, and if it is an intelligent bot it should be backed by human agents (normally the case).


It is also possible that a chatbot for retail answers other questions for the customers, such as questions about the product.  There is an argument here that if the chatbot allows support agents to be freed up from the drudgery of answering repetitive, simple questions, perhaps the scope of questions that can be answered by the combination of chatbot and human agent ( what we will call the “chagent”) can be increased dramatically.  For example, the changent could answer questions about a product’s features for online customers in the way that in-store human assistants assist customers right now.

This brings us neatly to the concept of in-store and location-based chatbots.  These are voice bots positioned close to the products in-store that customers can ask questions about the products.  Again, this can only be done if it is possible to offer a seamless service to the customer by providing the bot with instant back up to a human agent and the ability to call an instore human assistant if necessary.  These bots can not only answer product questions, but could answer general questions, such as “where is the milk aisle”?  The information in the bot could be hardcoded, or could be compiled by AI bots online or physically in store scanning the available items.

The chatbots themselves could be available in-store or on people’s smartphones, using beacons to allow customers to find location-based chatbots.  Location-based chatbots could also be used to deliver location-based, time-limited promotions to people nearby a store.  Adverts and promotions backed by chatbots can of course access the full power of the chatbot to respond to customer questions.


There are many use cases for chatbots in retail, from shopping chatbots to location-based chatbots. As usual, it is critically important to consider the use case in question to consider if a chatbot is the best channel for the customer, and if so, what type of chatbot to use.


Disclaimer: We encourage our blog authors to give their personal opinions.  The opinions expressed in this blog are therefore those of the authors. They do not necessarily reflect the opinions or views of Botpress as a company.