Interviews, insight & analysis on Ecommerce

Enough chat – it’s time to use ChatGPT to offer customers more personalised experiences

By Dan Hartveld, CTO at Red Ant  

With the rapid adoption of AI and evolution of ChatGPT, there’s much industry hubbub around the place of AI tools and “bots” in the retail ecosystem. Although investment in AI is nothing new, IDC predicts that global spending on artificial intelligence will skyrocket to more than $300bn by 2026. In the retail industry, with customer experience and personalised services both priorities for retailers, CIOs, ecommerce and customer success leaders are starting to consider how new tools like ChatGPT could become a source of support and inspiration for store associates.

The challenge for store associates 

Brands are aware that their human expertise is a critical asset to building remarkable customer experiences. If a brand expects their store associates to be doing more and more clienteling, they want them to be sending out messages and creating relationships with customers. But unless they’re a super-premium brand, their store associates are busy with lots of other tasks to do. With hundreds if not thousands of customers walking through their stores each day, it’s not only a challenge to identify which customers to focus on building relationships with but also to gain inspiration about what to talk with these customers about. 

On a practical level, tasks like this are something that a natural language AI such as ChatGPT can help with, because ChatGPT is really effective at generating proactive ideas of what store associates could be talking about, and what customers could be interested in. Its particular forte is creating natural-seeming engaging messages that can rival what the best store associate can do from an English skills perspective. 

The state of ChatGPT in retail 

Exploring the recent hype around ChatGPT, Red Ant asked retail executives their views at the annual Retail Technology Show. It seems that, while brands are keen to experiment with new tools like ChatGPT, 32% of them said they don’t know enough about the potential of AI models for supporting store associates, so this will be a learning curve for 2023. Although 14% of retailers said they felt tools like ChatGPT might help store associates in their roles, a solid half (50%) claimed that it can’t replace the human touch. 

Yet with 44% or brands aiming to increase store associate productivity, exploration of ChatGPT with technology partners can help them to deploy this new technology effectively alongside human expertise. 

Opportunities and limitations for ChatGPT in retail 

Without a doubt, new tools like ChatGPT look set to be an enhancement for retail experience. In retail, they’re going to be a tool for ‘real life’ store associates rather than a replacement for them. In a retail context, ChatGPT is especially good at identifying and inspiring store associates about what customers might be interested in. This makes the store associate best placed to deliver for customers’ expectations and to respond to queries throughout their browsing and buying experiences – to clientele better.  

Rather than store associates creating a message to the customer and sending via WhatsApp, a set of automated message actions can engage further with the customer and maintain a personalised conversation with them based on their history and preferences. Due to this prompting of ideas, it’s strikingly good technology for engaging and motivating store associates. 

ChatGPT is also game changing from an ease of implementation perspective. Traditional approaches would involve a data scientist or investing in a team of experts, so the barrier to entry is far lower. The biggest advantage for today’s retailers is that they don’t have to invest heavily just to try this new tech out. 

And the cons? There are inevitably some dangers in that ChatGPT is bad at understanding the real-life context and therefore knowing its limitations is essential. For instance, it might give a response about a line of products the business isn’t currently promoting, or the store associate knows the customer isn’t interested as they’ve already spoken with them. The key issue is that ChatGPT has a bias to please, so it will often pretend it knows something even if it doesn’t, answering with authority and certainty even if it’s completely wrong. It doesn’t want to disappoint and for this reason alone retailers can’t trust it to interact on its own. 

On balance, these new technologies can help store associates become more productive and create great customer experiences. However, ChatGPT hasn’t been deployed in enough meaningful scenarios for brands to fully understand its potential. The next generation of ChatGPT will see it embedded within software systems like an omnichannel platform, part of search engines, and as bolt-on functionality. 

Key considerations for ChatGPT deployment 

When exploring ChatGPT, there are some key considerations to bear in mind prior to implementation, such as:  

  • The importance of retail trust and quality control 

When brands can’t guarantee what ChatGPT is sending to customers, they need some kind of quality control and store associates are best placed to provide this.  

  • Getting data in order 

Success of recommendation engines all comes down to having good data about customers and products. Although ChatGPT is allowing retailers to do things with limited data, getting data in line is essential for adoption of this kind of technology. Working with an omnichannel platform that harnesses all the retailer’s data is essential. 

  • Human expertise 

Tools like ChatGPT need real people to use it effectively, so retailers need to have a platform to provide that to real life users in a useful, readable, and actionable way. In-store tech is a better place to get started with ChatGPT. Since you can’t guarantee what it will say to people, you need to have store associates in person to navigate any unexpected situations. 

  • Measurement 

With omnichannel platforms like RetailOS you gain essential feedback on positive versus negative interactions and can AB test with store associates who are using it as a prompt.  

Getting started with ChatGPT 

Progressive brands can get started today in exploring the potential of ChatGPT tools. It is surprisingly easy to implement and brands can test it quite quickly as it’s not necessary to train models. The next wave for ChatGPT is going to be software tools using ChatGPT as new functionality within their systems. 

With some retail tech specialists integrating ChatGPT tools into their platforms, retailers who are excited to explore the possibilities should trial the technology to fully understand its potential for creating remarkable personalised customer experiences. 


More posts from ->

Ecommerce Age

How do we solve the issues with lockers?

We’ve all seen the banks of parcel lockers that we’ve all seen outside supermarkets and train stations. They are clearly a sensible idea, and one that I’ve been more than willing to use, but I’m very rarely given the option to do so when ordering goods online. Part of the problem, according to Gary Winter, VP of global strategic initiatives for parcel lockers at Quadient, is that they are invariably linked to a single delivery firm – such as Amazon or InPost – and this limits traction.


General Retail

More posts from ->

Related articles