What A Generative AI Chatbot Can Do - For Growing eCommerce Sales

I have worked with eCommerce brands for over a decade now, and almost without exception, none of them uses an intelligent chatbot.

At least not effectively.

While programming a chatbot requires a significant amount of effort in scripting all possible questions and responses, a lot of them ended up causing more frustration the relief, by taking the users into an endless (and sometimes mindless) loop.

Sure, there are order tracking devices, and some basic MCQ type conversations, none of them options are really a chat unless you connect them directly to a human.

Until...

Enter ChatGPT circa Nov 2022

Given the widespread adoption of ChatGPT, and the kind of conversational help it can provide, it is only natural to assume that humans prefer a chatbot interaction over "delayed" email or support ticket exchanges, as a behavior.

So, let's start with this as our fundamental assumption and build at least 5 use cases that have the potential to double, if not triple, (or even quadruple) the conversion rates on most eCommerce stores.

Standard Product Recommendation

I know, almost every single store that has some customer-focused strategy powering their decisions, is using a "recommended products" widget of some sort.

For the most part, it's an "AI-powered" recommendation engine, which has little to no knowledge of what this specific user wants.

And it doesn't always work :-( as expected.

So, what is the alternative?

I propose the following example that is powered by a large language model and accesses product and inventory information in real (or near-real) time.

The bot responds first with a relevant question based on the keyword identified in the original question posed by the user

Notice how the customer doesn't provide any answers, so the bot continues and asks a follow-up question based on the product itself in order to make a decision for the customer.

Then the bot moves on to ask questions about accessories, but relevant specifically to this product. This is one of the primary reasons why most eCommerce companies do not have chatbots.

In order to program it this way, they have to go through 100's of scripted questions in case they have anything more than a dozen or so SKUs

The opportunities are endless.

Here is another example of a B2B eCommerce chatbot.

Before anything else, the bot starts by asking the quantity, which normally is not the case in a B2C cycle. Since this is a B2B decision for the most part, a large part of decision-making will depend on the quantity and the bot is smart enough to ask this question.

Somewhere in the conversation, the prospect doesn't understand a feature the bot asked about, so it takes a minute to explain the feature before getting back into the conversation.

Let's take context awareness to another level.

In this conversation, the bot starts with a relationship development question to personalize the conversation and build a rapport firsthand.

Then, upon receiving further confirmation on the context, it queries an external system to get a weather report for a particular city at a specific time of the year, and begins building upon the information.

Notice how the user started with warm weather jackets, but the bot smartly steered the conversation towards need-specific recommendations.

It does this by way of providing not only what is asked originally, but also recommends additional products (that aren't even sold in this store)

This solidifies its intentions of being helpful and as the buyer is already making a decision to buy, it throws in another helpful upsell that they DO sell in this store.

Last, but not the least, the bot takes them through options to finance their purchase.

What kind of an intelligent bot would you develop for your business?