We had a chat with ChatGPT about the impact it could have on the ecommerce industry. And this is what it told us. 

ChatGPT is taking the world by storm – and with good reason.
(Nevermind the Bing debacle)

The AI language model developed by OpenAI is part of a new wave of AI technology, generative AI. 

Unlike much of the existing AI technology, ChatGPT can generate unique content and ideas, and develop them further via dialogue with humans (and I suppose, with other AI).

Things are changing rapidly and the impact is predicted to be huge.  

Have a read below of our conversation with ChatGPT about its potential impact on the world of ecommerce. 

How may ChatGPT impact the world of ecommerce? 

I had a brief conversation with ChatGPT about how AI language models (like itself) could impact the world of ecommerce, and here’s what it had to say (and it wasn’t hello *emoji*):

This is a good start, because we have already seen ecommerce technology companies using AI to automate workflows and provide a better customer experience in some of these areas. 

For example, Gorgias, the customer service support for ecommerce is already making use of AI models to automate tasks and make better decisions. Gorgias can segment customers by dynamic Magento order data, allowing customer services teams to lean on more valuable customers. 

Klevu is also making use of AI models. They are using it for natural language processing of on-site search queries, and personalisation.

May ChatGPT help us understand the state of our customer experience?

I was interested in the idea that a language model could improve customer experiences, as I tend to see that as something that only humans could do, via human to human empathy. Alas, I think I was wrong.

This shows that a language model could not only provide tactical solutions to generic problems (like product discovery and customer service) but can also automate the design of research methods to understand specific problems in customer experience. 

I asked whether it could take care of a whole lifecycle of quantitative and data-driven research programmes.

It can! It can design surveys. Deploy them. And make sure it collects enough answers, but only the things we need to know. And then it can deliver synthesised answers. The possibilities seem endless. 

May ChatGPT make specific recommendations to improve customer experience pain points, based on data it has collected? 

We have already started to see a trend in platforms such as Google Analytics, where the software provides recommendations based on historical data, such as anomalies or benchmarks. 

But it’s interesting how things could evolve. 

Currently, you can analyse the average speed of the site, however, a more intelligent piece of AI could monitor page speeds and work out the ideal page speed to aim for based on customer exits. 

I was wondering at this point how specific an AI language model can become. It’s not that helpful to say “your conversion rate is low, what about more payment options”, if for example customers aren’t converting for other, more important reasons. 

AI language models could be specific and contextual, by comparing themes in customer experience pains with customer satisfaction scores. You can imagine customer data such as lifetime value being taken into account too 

Imagine this – “Be careful, the satisfaction of your most high-value customers is dipping because of increased delivery times, consider reviewing your delivery partners?”

Great news. The AI model will be able to take into consideration business objectives and other pieces of context, before making any suggestions. For example, AI might not suggest lowering delivery costs if that would result in a loss. 

Interestingly, the AI model suggests at this point A/B testing. In a humbling Socratic moment, the AI model understands it can’t ultimately predict the future, and suggests a test-and-learn framework that, wait for it, will improve its recommendations over time and the intimate knowledge it has of the problem. 

It’s amazing to think that AI language models can help people in charge of ecommerce strategy and customer experience make the best possible decisions.

May ChatPGT become the holy grail of ecommerce? 

The AI language model can help ecommerce teams make smarter decisions which can help them build a solid strategy and enhance customer experience. 

Let it be that invaluable rubber duck that can help ecommerce teams with research, consider new initiatives and validate them through testing and learning. 

However, remember. ChatGPT is not always right and is often based on assumptions built on the information it has learned from. 

Like humans.  

Would you like to know more about how you can use data and insight to improve your customer experience and consider new initiatives? Contact us for a friendly chat about all things ecommerce, strategy and robots.