About This Paper
Building AI applications for e-commerce has several usecases ranging from providing inventory intelligence to pre/post sales customer service.
This paper speaks of one such usecase, where a virtual assistant guides a potential/existing customer to chose the right product amongst hundreds of products in the organization’s portfolio – based on the specific requirements of the customer as gathered during the chat interactions with the virtual assistant.
Ecommerce companies are early adopters of Artificial Intelligence technology. Specifically, Natural Language Processing (NLP) technology (language part of AI) is being used by these businesses to facilitate personalized shopping experience for customers.
Most of the personalized shopping experiences (apart from recommendation engines) are facilitated with the help of automated conversational interfaces like chatbots or virtual assistants.
According to a recent study report, it was projected that by 2020, 85% of all the customer service requests of businesses will be served by chatbots or virtual assistants.
Traditional Product Catalog Search in E-Commerce
Online product catalog search is the heart of any eCommerce business. Most of these eCommerce businesses implement a search feature in their respective online stores by creating and displaying a product catalog with a portfolio of products that are categorized based on labels like brand, price and product specific features like size, color, pattern, length etc..
Potential or existing customers typically search across these categories and labels. Some of limitations in doing traditional search using these labels are:
- Spelling mistakes in product search: For example, the customer wants to buy a “15-inch laptop”. If the customer typed like “15 inh laptop” in the search bar of an e-Commerce site, it may not display relevant products to the customer. Neither are any hints shown to correct the spelling mistake – which further results in poor conversion rates.
- Contextual product search: Contextual search is the process of understanding the intent of the user search and showing precise results according to the intent. Most of the online product catalog searches are keyword based and do not understand the meaning of each keyword to infer the context or intent of the user search.
- Availability of physical agents 24×7: In case of any queries during the entire search process, support from a physical agent is required to make a final purchase. Providing 24×7 support to potential or existing customers is a huge challenge for businesses that could only be addressed by making considerable & continuos investments in training and deploying human resources.
AI driven natural language search or smart search delivers more than just accurate results. It helps in streamlining the sales process, satisfying customers – which eventually results in better conversion rates.
Natural Language Search in E-Commerce
In a brick-and-mortar store, a physical agent – salesperson helps the customer in suggesting right product. This suggestion is made possible because the customer can ask as many number of queries to the salesperson (apart from seeing a demo of the product) before making a purchase.
In other words, a natural conversation using language and media is playing a vital role between the physical agent (salesperson) and the customer.
If a similar user experience has to be facilitated in an eCommerce store, it needs advanced technologies like NLP, which is language part of AI to build – what we call as Virtual Assistants – akin to the sales person in a brick-and-mortar store.
By applying NLP to facilitate an online product catalog search in an eCommerce store generates good conversation rates. The virtual assistant can be trained to suggest right products to the site visitors by engaging in a conversation through a natural language with the customer.
We have trained a virtual assistant for a large organization to facilitate an online product catalog search.
For a demo of how a trained Virtual Assistant could be used to facilitate online product catalog search, pl click here.