When Everlane announced plans to ditch its Facebook Messenger chatbot in March, it was one of the first fashion retailers making use of a chatbot to strategically move away from the ubiquitous tool and return to traditional email communication.

“We’re writing to let you know that we’ll no longer be using Facebook Messenger as a notification service. It was a good couple of years, but we’ve decided to stick with what we do best — email,” Everlane wrote in an email to customers.

Over the last few months, fashion brands have largely remained quiet when it comes to chatbots, a technology that has been particularly fraught for retailers struggling to successfully engage consumers. Though there was a flurry of new bots when Facebook first rolled out the offering in April 2016 — Burberry and Tommy Hilfiger were early adopters, both creating Facebook bots to promote their fashion week shows that fall — few luxury fashion brands have since joined the platform.

According to research from GPShopper, a retail mobile app developer, only 19 percent of consumers are familiar with chatbots, and just 9 percent think the technology will improve their shopping experience. Maya Mikhailov, CMO and co-founder of GPShopper, said the diminishing interest from consumers across all industries comes from poor product development.

“When you look at the natural technology hype cycle, chatbots are in the trough of disillusion. There was original excitement that they would solve all these problems and increase communication. Then, as chatbots were actually implemented, there were limitations on what they could actually can do,” she said.

As a result, brands that do choose to develop bots are moving away from platforms like Facebook, Kik and WeChat, and instead developing integrated Apple iOS programs. Saks Fifth Avenue, for example, launched an update to its iOS app earlier this month that incorporates a messaging functionality allowing consumers to ask questions and make purchases directly through the app.

In addition to tapping into a captive audience — with iPhone accounting for 43.5 percent of the smartphone market — it serves as a trial run to Apple’s forthcoming “business chat” capability, which will be integrated into the next iOS 11 update. Mikhailov said, in essence, iMessage “creates a uniform bot paradigm for businesses.”

Apple will need to be savvy about handling the major problem of existing bots to date — namely, that they aren’t properly developed or trained to handle communicating with humans, Mikhailov said. She noted missteps by brands like Asos, which used bots pretending to act as real-life customer service assistants on its Facebook page in May 2016. Users discovered that responses attributed to human employees were garbled and nonsensical, exposing that Asos had used bots to handle queries.

“[Brands are] pushing the limitations of the bot technology too early into complex decision-making and complex replies,” Mikhailov said. “Bots needs to be trained, meaning there’s a lot of human interaction that they need to understand, and that takes a while.”

She said the best use case of the chatbot is very simple customer service requests, like locating a shipped item or checking on inventory. GPShopper research has found that bots can help with upward of 80 percent of customer service inquiries, which in turn, allows human customer service associates to tackle more difficult queries.

“With chatbots, a retailer can focus its customer service on higher-level problems, which will ultimately lead to better customer service,” she said. “It can empower them to make decisions that make a customer happy and lead to better interaction.”