Q&A with TechStyle: Bots and the Digital Advantage

Two days. That’s all it took TechStyle to set up an Amazon Lex bot, one of many innovations the company has made to offer a better experience for customers and agents. The bot has been instrumental in freeing human agents from dealing with non-revenue-generating calls like billing inquiries, order checks and similar routine tasks. Now those agents focus more on up-selling and cross-selling.

TechStyle Fashion Group is the largest digitally native eCommerce platform. It’s known for its brands — Fabletics, Savage X Fenty, JustFab, ShoeDazzle and FabKids. The company handles nearly 6 million phone calls and 3 million chats per year from its 5 million online shoppers across 12 countries.

I spoke with Aarde Cosseboom, Senior Director of GMS Technology, Analytics, and Product for TechStyle, about the company’s experience with chatbots and the results.

You replaced your existing chatbots using Amazon Lex through the Genesys Cloud™ platform. What was the driver?

Aarde Cosseboom: Our goal was to increase the number of countries in which we could deploy bots, and increase the scope and self-service offering of our chatbot — without sacrificing customer service. We also knew it would help us offset the large spikes in volume and needing the staff for those large spikes. And we’re able to modify the chatbot experience as we learn from it, including doing A/B testing on different bot workflows.

You’ve written about what you call the call center “service onion.” Can you explain that and how it fits with your bot strategy?

Cosseboom: In call centers, there are different layers of support and customer experience, and different ways to tackle how you offer that support. When you look closely at the layers, each one has different uses cases and ROI. The outer layer is self-service where customers don’t want a live agent. The next layer is self-serve automation. It’s similar but might use chat or voice IVR to update personal information. The inside layer includes traditional channels, phone support, SMS and live, one-on-one interactions. While these can be costly, they can also be the best experiences.

With our bot at TechStyle, we’re giving customers more choice on where they want to go in to engage with us — whether self-service, bot or traditional.

What were some of your specific results?

Cosseboom: We’ve doubled self-service use and the deflection of chats that go to agents. That alone has saved us about $250,000 annually. We’re also reducing the same amount in licensing costs. These cost reductions were in addition to lower agent average handle time, inbound contact volume and hold time in our queues. We also increased customer satisfaction scores.

How did employees respond to this new way of working with bots?

Cosseboom: Six months after the cutover, we surveyed our agents and they were positive about the switch. In fact, we’ve seen a reduction in the number of technical complaints from our workforce and even encouraging feedback about the tool. And we’re still working on the agent experience. For example, we want to fully use the skills of our existing employees to improve their job satisfaction and overall business results.

Will digital lead to the end of your brick-and-mortar stores?

Cosseboom: No. They’re playing an important new role as marketing fronts for our brands. People can go into a store, try on clothing and have it shipped to their home. There’s no check-out line. We’re part of a shift in reimagining fashion and the way people shop and interact with brands — all eCommerce-based. Putting our systems in the cloud is what enabled it.

 Is there anything you wish you had known when starting with bots?

Cosseboom: Definitely. We didn’t realize how hard it could be if you don’t already know the intents, which are the reasons why customers contact businesses. And intent capture and mapping was hard to do without some external help.

What’s your advice for those just getting started?

Cosseboom: Start with a couple of intents first. Don’t try to release a bot to handle every type of inquiry. We started with four of our highest volume intents and we were able to self-serve about 40% of the inbound contact volume.

What has been the main impact of bots on your organization?

Cosseboom: There were actually three. First, we reduced inbound contact volume which gave some relief to our large queues during busy times. Second, agents are able to see what happened with the bot before being routed to them. Having that full context has enabled quicker resolution and less repeat questioning. And third, we have new data in our reporting around intents that allows us to categorize the types of intents our members or customers have when they contact us. With this, we’re identifying harder or longer conversations based on intent type.

What do you consider your biggest accomplishment?

Cosseboom: Giving customers a consistent and positive experience. By using bots, we’ve made it easy for them to engage — even as we add new digital channels. We’re also making it easy for our agents to treat customers with empathy.

Listen to the podcast on Aarde Cosseboom’s quest to improve customer service.