The AI Effect: Customer Service

Étienne Merineau
October 23, 2018

It’s no secret that Canada has become a hotbed of AI, and it’s a community that Heyday is proud to be a part of. So we’re just a little excited that one of our co-founders, Etienne Mérineau was recently featured on The AI Effect, a podcast series that shines a light on this thriving ecosystem. Hosted by Amber Mac and Accenture’s Managing Director of AI, Jodie Wallis, this podcast series examines the extraordinary pace of AI's technological growth and the rewards, and perils for individuals, businesses, and society.

In this episode of the AI Effect, Jodie, Amber and Etienne dive into the role and possibilities of AI in the world of customer service. Now more than ever, an enjoyable, efficient customer experience is crucial to establishing brand loyalty and building a committed community of customers. Customers expect more from the brands they interact with, and they expect it fast. The benefits of AI span a number of industries, and Etienne presents just a few of the exciting new opportunities that AI can offer brands and customers alike.

Listen to the podcast here or read the interview transcript below

Jodie: With us, we have Étienne Mérineau, co-founder and chief product officer of in Montreal. Can you tell us a little bit how you're using AI at Heyday?

Etienne: Essentially, we help retailers connect with customers on a personal level throughout their journey, from product research, to purchase, to support. We use AI in a variety of ways, but our AI, in particular, is both customer-facing and agent-facing. Our software takes care of the lead qualification triage, and then escalates the conversation to the right human when necessary. Take our client in the sportswear industry, for instance. If someone asks a more complex question about golf and the AI can’t handle it, it will escalate it to one of the brand’s golf experts, and the AI subsequently learns from these interactions for future conversations.

On the agent side, our AI works to give “smart” customer service agents recommendations that help them be more efficient. Our model is based on giving superpowers to sales and customer service teams by having that dual capability.

Jodie: Sometimes chatbots get a pretty bad rap. They don't understand what consumers are saying or they aren't able to use any sort of intuition. What makes Heyday different?

Etienne: You're right, “first-generation” chatbots weren’t particularly smart for a number of reasons, but I think the most glaring issue was that they weren't really connected to or tapping into the available data. In some cases, the data was unavailable because the data was poorly structured within the organization.

At the end of the day, data is the supreme fuel of AI, so for AI to truly deliver value, and for quality data to emerge you need to open up the communication channel with customers.

Our model is to basically open up a hybrid channel of AI and human agents to cover the whole spectrum. That way if the AI cannot process the question, it automatically escalates to a human being who can handle the issue. It's kind of the best of both worlds. Basically, automation is perfect for a customer asking a common question and expecting an instant response, but if the question is too complex, too personal, too sensitive, it’s funneled directly to a customer agent.

Amber: Can you talk a little bit about where you see businesses making mistakes when it comes to creating their own chatbots?

Étienne: Absolutely. First and foremost, I would encourage any business looking into AI to really start with a well-defined use case. For example, if you see your customers are asking a lot of the same questions over and over, you might consider FAQ automation, because even if you have an FAQ section on your website, people can get impatient and might prefer to just ask the question in a live chat to get the answer right away. This example is a specific use case where AI can do a good job.

The number one pitfall though, I would say, is businesses trying to be too broad, or trying to solve too many business problems at once, which results in the AI failing at all of them. Focus on one single thing. That’s my first tip. 

Secondly, from a conversation design standpoint, brands need to be fully transparent. As a customer, if I'm talking to a chatbot or an AI assistant, I want to know that I'm talking to a machine. I don't want to be misled. That’s what created frustrations with early chatbots. Because if I'm chatting with what I believe is a human being, I'm happy to have an instant response. But if there's no real value in the AI exchange, or if the AI doesn't understand me and I thought it was a human being, I really just feel like you're an idiot and the whole experience backfires. It creates a negative effect where I'm not going to be loyal to the brand and I'm going to go shop somewhere else.

Amber: Can you give us an example of a client that was able to use your services to either increase revenue or lead to more customer satisfaction?

Étienne: One of our clients is a local renovation center store chain, Patrick Morin. They have 21 stores in Quebec, similar to Lowe's or Home Depot, and while they’re not as big, they still have 40,000 products, so there’s definitely some complexity there. As a small brand, their goal was to differentiate themselves by providing more personalized service. That said, they only had two people at their headquarters who could work full time on delivering their customers that value through live chat.

Within eight weeks of implementation, we managed to automate 60% of their FAQs. The remaining 40% is now redirected to the two people who are dedicated to live chat full-time. And we continue to train the AI based on these interactions. For them, it saved them from hiring an extra two people while giving them an edge in customer satisfaction since their service was more personalized and timely.

At Heyday, we see an increasing number of brands across all industries becoming more curious about conversational marketing and broader AI opportunities. The possibilities are endless, but it’s more important than ever to tailor these technologies and find a solution that meets both your marketing goals and customers’ expectations. 

Interested in learning a bit more about how Heyday can help boost your customer service team and deliver a more personalized customer experience? Get in touch with us for a free demo.