The Key to E-Commerce Success: Blending Artificial Intelligence with Intelligent (Human) Agents

COVID-19 will forever change retailing, and its initial impact on e-commerce is creating challenges to online selling and service that no one imagined back in January.

As the COVID-19 pandemic continues to reshape our world, more consumers have begun shopping online in greater numbers and frequency. According to new data from IBM’s U.S. Retail Index, the pandemic has accelerated the shift away from physical stores to digital shopping by roughly five years.

Department stores, as a result, are seeing significant declines. In the first quarter of 2020, department store sales and those from other “non-essential” retailers declined by 25 percent. This grew to a 75 percent decline in the second quarter.

The report indicates that department stores are expected to decline by more than 60 percent for the full year. Meanwhile, e-commerce is projected to grow by nearly 20 percent in 2020.

In spite of the plummet, many retail brands (particularly omnichannel retailers) are finding growth through other means, such as online retail and e-commerce. “The pandemic has accelerated the amount of online retail,” said real estate mogul Sam Zell, “and I don’t think that’s ever going to change.” 

Data from Insights confirm this, with online retail revenue as of August 23, 2020, up 71 percent over just a year ago in North America alone. And just as e-commerce surges, so do customer challenges.

COVID-19 Impact on Customer Service

The results of a study earlier this year paint a troubling picture for customer experience and customer-service leaders as well. It covered roughly 1 million customer-service calls. They involved more than 20 companies, representing a broad cross-section of industries.

The findings show:

• In just two weeks, the average company in the study saw the percentage of calls scored as “difficult” more than double from a typical level of 10 percent to more than 20 percent.
• At one company, financial hardship-related calls increased 2.5 times in the span of a week.
• Some businesses had hold times balloon by as much as 34 percent and escalations (calls sent up the chain of command) skyrocket more than 68 percent.

Such a dramatic rise in high-effort interactions are far more likely to lead to customer churn and far less likely to result in any sort of new sales. 

Combining Artificial Intelligence and Intelligent Agents for E-Commerce

The merging of machines and humans occurs as retailers reinvent and reinvest to win over consumers in a “digital-first” environment. Artificial intelligence (AI) and the intelligent agent (IA) must complement each other—with automation segueing into customer care by human agents.

Especially in e-commerce, where synergies flow both ways. Smart retailers mine their data. They analyze consumer purchases, identify patterns, capitalize on buyer preferences, and use that data to scale their numbers of agents to fit the buying season. 

Rather than replace agents, savvy use of AI can drive decision making that dictates the types of service reps to bring on board, their skill sets, the timing of hires, and the numbers of agents required to handle seasonal demands.

Planning for the holiday rush is the obvious example, where retailers’ entire annual profits depend on the final weeks of each calendar year. And with e-commerce soaring 71 percent year-over-year in North America alone in response to the COVID-19 social distancing demands, retailers’ use of AI to predict seasonal customer-service requirements has become even more complex.

A Most Taxing Tax Season

Perhaps less obvious is the scramble that tax-preparation firms experience in Q3 and Q4 each year as April filing deadlines loom and demand for customer assistance skyrockets. Even less predictable was the 2020 season, when filing dates slid into July and the need for additional agents extended well beyond April.

Working Solutions and its tax industry client meshed historical data, forecasting trends, market research, and pilot projects to determine agent needs for peak volume months, well in advance of an otherwise unforeseeable demand. 

The result? The tax-prep firm was able to scale twofold its network of remote customer service contractors and sustain that support throughout one of history’s wackiest tax seasons ever.

Firsthand Agent Knowledge Sells

Just as AI can rescue IA, so can the reverse be true. Agents, who are often customers of the clients they represent, provide insight from their own experiences. They know the products being sold firsthand—such as the fit and wear of a particular garment or the performance of a certain device.

This human touch often influences a buying decision—a sale that’s initiated through e-commerce and cinched through the experience of a knowledgeable agent.  

Take the case of Courtney, a college student who had a lot of extra time on her hands when her job as a server at Denny’s ended and her internship for a graphic design firm was put on hold due to COVID-19 quarantine restrictions.

During a call to a major North American hobby store about a routine $25 order, Courtney met Vikki, a remote customer service agent from Canada with a passion for crafting and upcycling.

The two began to connect over their mutual love for making old things new again. Vikki explained how purchasing a Cricut® smart cutting machine could really speed up Courtney’s artistic process. 

As a result of Vikki’s suggestive selling and probing, Courtney is now running a successful side hustle on Etsy while continuing her education this fall. With the use of her new machine, she is selling upcycled china, crystal, and glassware all over campus and to points far beyond.

Chatbots Only Go So Far

Sometimes it’s difficult to tell when AI leaves off and IA takes over. For example, the uptake of chatbots as customer service “agents” continues to gain ground as tech development costs decrease and acceptance for nonhuman interaction increases.

Successful bots handle routine requests, with agents at the ready to continue the conversation if needed. Customers transferred from chatbots to live agents expect to talk with a “thinking agent”—not a rote-reader using scripted responses.

This is especially important in complex or long-lead selling. Lisa, for example, purchased four tickets through on online ticket broker to see the Detroit Lions play on Thanksgiving Day last year. 

When she arrived at Ford Field and began to struggle with the mobile ticketing app, she worried that because of the holiday, she might have no recourse but to forgo the game. Enter Darren, a remote customer service agent and contractor to the ticket hub for the past five years. 

In-the-Stands Customer Service

Lisa and Darren were able to sync up via the broker’s chat application, a welcome option amid the din of a crowded arena full of turkey-stuffed Lions fans. Even through a virtual interaction, Darren saw the issue immediately and swung into resolution mode. 

Just as Lisa, her husband, and sons were about to step out of the turnstile line, Darren helped her complete the final step in accessing her mobile tickets, all via the chat feature on her cell phone’s mobile app. 

While the Lions lost to the Bears and the halftime show had some sound issues, Lisa’s family had a great time at the game, thanks to Darren’s commitment to first-contact resolution and a friction-free customer experience.

COVID-19 and online buying have upped the ante at both ends of end-to-end. By its very nature, artificial intelligence requires intelligent agents to interact with the latest IT—and to do it while empathizing with customers and thinking outside the digital box. In the best of all worlds, the ultimate in customer care.

Kim Houlne is the founder and chief executive of Working Solutions. Working Solutions, founded 24 years ago, is a recognized leader in on-demand, onshore contact center outsourcing. From across the U.S. and Canada, its workforce of sales, service, and tech-support experts elevate customer care and improve business results.