Words by Tim Warren, CEO, Ambit
As consumers seek out more personalised experiences online, retailers must find strategies for technology to take the volume while people take care of the value. Forrester predicts more than 25% of all US retail sales to occur online by 2024, a figure which poses a real challenge to retailers to change their business model to deliver world class experiences at that scale.
This speed of change makes it imperative now to look at how to bring humans and tech together in the online retail environment, using AI-powered chatbots to augment human agents. I like to think of this as the capability value of people multiplied by the capacity volume of technology, providing customised service while actually reducing cost to serve.
Gartner says the top three most significant challenges companies face when considering the implementation of AI are staff skills (56%), the fear of the unknown (42%), and finding a starting point (26%).
This was endorsed by attendees at the iMedia Retail Summit in Rotorua, in May, where the most pressing questions posed by retail leaders about AI were ‘where and when to start?’.
I believe that in New Zealand there are sufficient in-house and vendor skills to address the gap identified by Gartner. In most cases, employing small fast starts with rapid iterative improvements is the smart strategy and overcomes ‘we don’t know what we don’t know’. For some retailers, like Noel Leeming, lockdown provided the catalyst for implementing AI when its stores were closed but experiencing significant demand.
Launched in just twelve days in May last year, the ‘Nola’ chatbot (really a ‘digital employee’) helped Noel Leeming customers with click & collect requirements, locating branches, returns and exchanges, and taking the load off customer support.
Conversations were designed using a supervised machine-learning model to ensure that Nola remained on-brand with her tone of voice, protocols and manner of engagement.
Nola now handles 89% of conversations with a current cost of 47 cents per conversation and falling. While Nola was implemented as a solution to handle Noel Leeming’s growing customer service volumes, it was not long before they realised the power she had when it came to increasing their upsell capacity.
It turns out Noel Leeming consultants combined with Nola are great at upselling extended warranties, achieving close to the instore rates.
Nola’s experience shows that rather than being impersonal, the AI, by its ubiquitous nature, becomes familiar and is reliable, ready for your customers day and night.
It doesn’t mean having to forfeit what is good about human interactions: the AI is intuitive, and in offering a guided selling experience it is, in effect, listening to your customers. Funnelling customers to a live agent to provide expert product advice and assistance when required shows how both can work together.
Brands don’t need something as catastrophic as a national retail lockdown to begin their AI journey. The availability of conversational AI platforms that require no code means it’s easy for in-house teams to begin small and build their experience and business case.
Natural language processing is evolving and chatbots are improving their understanding of the sentiment and emotions felt by today’s customers, to provide ever more sophisticated levels of service.
Over 148,000 customers opted to interact with Nola online in the first seven months and 39% of these conversations occurred after hours. This means four out of 10 conversations addressed customer needs and brought in new revenue outside of live agent hours, a figure which more than anything else shows why conversational AI should be implemented now to address this opportunity cost.
Established in 2016, Ambit is now used by some of New Zealand’s largest retailers including Noel Leeming, Hallensteins and Glassons as well as smaller, online only retail brands.