Small retailers see big results with AI and machine learning


When Soleil Guerrero launched ecommerce site Bebonia in March, she didn’t realize how much time was needed to respond to customer inquiries. She thought she and her sister could answer the emails themselves.

“We were incredibly overwhelmed,” she says. “There was no way to track to see if our answers solved the customers’ questions and our response times were generally more than 24 hours. We lost a lot of potential sales because we couldn’t answer customer questions in a timely manner, and a lot of questions fell through the cracks.”

A similar customer service problem was happening at Sonja by Sonja Morgan. The online fashion company had two full-time agents answering emails, but as the business expanded it became clear more were needed, at least during busy seasons and special offerings. Agents were working 9 a.m. to midnight with even more limited hours on weekends, yet questions were coming in at all hours.

Both Bebonia and Sonja Morgan needed 24/7 customer service that could expand when needed. Both found the answer in Simplr, a service provider that uses artificial intelligence to assist in answering emails and providing online chat to help ecommerce customers.


Bebonia, which sells hair extensions, was looking for Simplr to provide nearly all its customer support. Most questions sent to Bebonia deal with customers needing recommendations about the different styles, color and curl patterns that would be best for their hair. With information provided by Bebonia, Simplr agents can do that.

Guerrero estimates that Simplr can answer about 90 percent of the emails and questions her company receives; only a few more complicated questions are referred to her.

“Since using Simplr, we’ve seen an uptick in orders, which has justified our investment. Before, customers who had immediate questions got responses back too late. Now we can reply to questions while customers are still shopping at our site at any hour,” Guerrero says. “Our biggest increase in conversions have been off hours when I was not available to answer emails.”

Having outside agents handle routine questions lets Guerrero focus on other aspects of her business. “This is my baby,” she says. “It brings down my morale if my customers are not being served properly.”

Simplr does more than just answer questions. Using artificial intelligence and machine learning to analyze data on customer trends, questions and interest, Simplr can improve responses to questions and spot trends so it can recommend changes to retailers.

By analyzing the questions sent to Bebonia, Simplr has been able to change language on the website and reduce customer inquiries. “We’ve learned how to improve our website content so customers get their questions answered without having to send us questions,” Guerrero says.

At Sonja Morgan, meanwhile, the issue was scale and access to information during off hours, says Laura Cantor, founder of Eleven Commerce, a firm that provides digital marketing and ecommerce support for Sonja Morgan. The site’s fashions are often featured on popular television shows, which generates many immediate email questions that were overwhelming internal service staff. Simplr provides additional assistance during these peak hours, after midnight and during weekend hours.

Sonja Morgan had looked at adding more customer service staff to handle the growth, but found it was more cost effective to use Simpler.

Most email questions Sonja Morgan receives relate to product information, shipment schedules or return policies. But Cantor says the company also receives emails from other companies that want to collaborate with Sonja Morgan or charities that want to co-sponsor events or promotions.


Simplr’s employees are U.S.-based and work as contract employees from their homes. Many have other jobs and are supplementing their income; others are stay-at-home parents who earn extra income.

That’s an important feature, both Guerrero and Cantor say. “I didn’t want to turn my customers over to agents who were not based in the U.S.,” Guerrero says. “These are top-notch agents.”

“They are given tests to make sure they have the empathy skills we want,” Cantor says. “And there is no difference in the responses customers receive from Simplr compared with our own agents.”

Simplr charges retailers a flat $2.25 per resolution; retailers are charged only if customers receive the answer they are looking for. Retailers are not charged for unresolved questions and there are no minimums or long-term contracts.

Vincent Phamvan, head of growth for Simplr, says that can be a big savings for many retailers. “If we are covering for a retailer from Friday night to Monday morning, a retailer might get 50 questions and we would charge about $105. To have 24/7 coverage, a retailer would have to pay about $12 an hour for 48 hours,” he says.

“A lot of the companies we looked at had a minimum volume requirement,” Guerrero says. “We didn’t have enough volume to justify the minimum fees. Simplr was more flexible.”

“I didn’t even know something like this existed before. This has helped us reduce our service costs and not have to hire more people, which would have been more expensive,” Cantor says.

Simplr founders had been working on the service for two years before launching with 10 clients earlier this year. The service opened to other retailers in June and Phamvan expects to have 100 retail partners by end of 2018, including some that use Simplr to handle all email and live chat questions as well as those who use it to supplement their own service during off hours or special promotions or seasons.

Phamvan says a lot of retailers will increase their use of Simplr during the fourth quarter to handle additional holiday-related business.

He estimates about half the questions his firm deals with are related to pre-sale questions about products. The remainder deal with return questions or inquiries about order and shipment status.

“We use our machine language technology to analyze a company’s products and policies and then we are able to respond to questions using the same language the retailer would have used,” Phamvan says.

Knowing so much about a company and what kind of questions customers ask allows Simplr to make recommendations to retailers. “We’ve been able to update the FAQs for a number of retailers we’ve worked with based on the frequent questions we’ve received. We also had one company that was using inadequate packaging for its product, based on the number of questions and comments we received relating to the packaging. We were able to recommend a change in packaging quality,” he says.

Lauri Giesen is a Libertyville, Ill.-based business writer with extensive experience in covering payment and finance issues.


Comments are closed.