AI and E-commerce

It’s unbelievable that ChatGPT was released to the public early this year and less than a year later, we are now in the midst of the AI boom. Every day when I open up my reading list of blogs, news sites, and newsletters, I read of companies either pivoting to AI, adding AI to their product mix, or new companies created with the mantra of AI first.

In all my years of being in the e-commerce industry working on online to offline, omnichannel, and other various initiatives. I think generative AI has the potential of being a game-changer for those working in the e-commerce industry. The e-commerce industry may not be as hot compared to a few years ago when Covid19 left us stuck at home with nowhere to go, and with only online shopping as the next best alternative.

But e-commerce is still a competitive sector and new competitors are coming in every day. E-commerce in South East Asia is poised to grow exponentially with big companies such as TikTok, Temu, Shein, etc eyeing the big market here. But there is still room for other players to find their edge through competitive pricing, and targeted market segmentation which are just some of the tools in their toolbag. With the explosion of LLM usage among the general public, using this new tool wisely is a good step to ensure that you get a competitive edge over your other competitors. A few good potential use cases for generative AI are

  1. Chatbots that actually do things useful. I am sure many of us would have dealt with chatbots that could not answer simple questions or gave us seriously wrong answers to simple questions. Feeding information about your e-commerce websites such as operating hours, inventory data, terms, and service, etc into the LLMs. Your chatbot is able to digest this information and use them to provide accurate answers to your customers 24/7 while leaving your customer support team to work on higher-value tasks.

  2. Product recommendations Recommending products is a very obvious feature which I am surprised that not a lot of e-commerce stores are doing. But the products that your e-commerce store recommends must be of some relevance. For example, if I have a history of purchasing dog food from your site. Showing me product recommendations for catnip treats is basically useless to me if I do not have cats. So imagine a recommendation engine that takes in a lot of data points from its various customer purchase histories such as - Price point of past purchases - Product Type - Products that other customers bought that are similar to what I have in the cart

    among other information such as the age of the customer etc. All these can be fed into an AI application built on something like LangChain and with a few carefully crafted prompts, you can have a working product recommendation system

  3. Search engines I am sure that you would have encountered search engines in e-commerce sites where you search for a word and the results either miss out on the context or returns irrelevant results. For example, you search for ‘Poodle toy’ and you get zero results in a pet store e-commerce store. So feeding in your product name, description, and tags. etc will enable LLMs to return more relevant results. This will reduce your bounce rate from customers coming to your site as they are getting better search results.

  4. Production descriptions Writing a good, insightful, and yet memorable product description is hard enough. Imagine writing it for the hundreds of SKUs that you have in your e-commerce store. With a few carefully crafted prompts, you can use them to produce and tweak product descriptions easily and quickly. No wonder prompt engineering is now 1 of the highest-paying jobs out there right now, even in recessionary times when a lot of companies are doing tech layoffs.

Running an e-commerce store is not easy and there are other well-established tried and tested technological tools such as improving page loading speed, SEO/SEM tweaks, customer journey UX enhancements, etc. And LLM should be considered as just 1 more tool and a groundbreaking tool at that. These are just some of the ideas of how LLMs can assist you in getting a step up over the rest. In a market where there are a lot of competitors, you will need every tool or help you can get to keep ahead of the competition.