Wynston Loading Icon

Aug 26, 2020 | By Nina Lalkovic | Share this article

Challenge

One of Canada’s largest home improvement retailers was looking to improve their e-commerce and digital customer experiences at scale by providing fast and accurate answers to their customers, especially during peak volumes. For large organizations such as this one, that have existing chat implementations, gaining valuable insights from real customer conversations is challenging, especially if here are limited resources or expertise available to mine the data. As an early adopter of conversational AI, this organization was looking for ways to better understand and leverage insights derived from real customer behaviour to prioritize and improve business processes and support a more consistent customer service environment across channels. 

 

Solution

Beginning with a 30-day discovery session, Wysdom worked closely alongside the home improvement retailer’s contact centers to uncover the best candidates for automation and train the underlying proprietary deep learning system to understand the language of the organization’s customers. By analyzing historical chat logs and shadowing call centre agents, the Wysdom team was able to identify three key priorities that could quickly and easily benefit from a daily optimized conversational AI system that keeps on getting smarter. 

 

Those key areas were:

 

  • Unified customer support: providing instant, 24/7 support to customer inquiries for frequently asked questions and gracefully escalating to an agent for more complex inquiries 
  • Business insights: using real-customer behaviour to drive business process improvements and decisions 

This sort of deep chat analysis is an easy way for organizations to gain better insight into the types of questions customers are asking, helping to identify any content gaps or areas requiring more attention. 

 

Results

After the initial 30 days, Wysdom performed an A/B test against the incumbent live-chat vendor to really understand the current online customer experience. The test unveiled a number of process improvements and helped to identify several new use cases for automation. Wysdom’s virtual assistant also saw a double-digit increase in customer satisfaction in comparison to the incumbent live-chat solution, learning 25 new intents, 1,100 new utterances, and 12 new customer journeys. It was also directly responsible for driving sales, contributing over $300k in net new orders over the course of 3 months. 

 

Uncovering deeper insights 

Today, Wysdom’s conversational AI powers the home improvement retailer’s entire online chat traffic with over 10,000 customer interactions a week. The bot is capable of handling a variety of customer journeys related to order tracking, finding relevant product information, supporting customer issues, and answering frequently asked questions – with the ability to escalate to a live agent for complex inquiries. Through ongoing training and optimization from the Wysdom team, the virtual assistant has also been able to successfully identify content gaps and create new intents based on customer need. 

 

Ongoing optimization is key to success

Agility and ongoing collaboration has been a key component for the success of this implementation. Organizations looking to leverage conversational AI to gain actionable insights from real-customer behaviour need to be willing to work closely with their vendor to drive results. It is only through this ongoing collaboration that organizations can reap the true benefits of AI-powered conversations and improve their business processes and customer experiences at scale. 

 

If you think your organization can benefit from improvements such as the aforementioned ones, contact us.