Chicago-based Sprout Social, a social media management platform, just introduced a feature to make Twitter message recommendations. Called Suggested Replies, it uses machine learning to suggest answers to the most asked questions.
The machine learning algorithm analyzes incoming messages to see if you’ve received similar messages before. It also searches recurring responses to those messages, and then suggests replies you’ve used in the past.
“We created a feature that helps relieve the stress of constant social replies while still staying true to our core principal that social should be about authentic and not automated communication,” said Andrew Caravella, VP of marketing at Sprout Social.
Users will still be able to edit those suggested responses if they need to add any information or tailor the message so they don’t sound automated. The idea is saving time on frequent questions and their frequent responses while still sounding authentic, personal and consistent.
Companies can also use the feature to cover their bases during nights and weekends with no staff, Caravella said.
Increasing Social Media Responses
Suggested Replies is a part of Sprout Social’s enterprise plan, created after a Sprout Hack Week project.
As more consumers turn to social media for questions and support, some companies struggle to keep up with responding to all queries within a reasonable time. Sprout Social believes customer service teams can use this tool to improve response rates and times.
That was the basis of Sprout Social’s last quarterly report in March, which concluded media and entertainment companies fall short in responding to social media messages. These brands receive three times more social media messages than other industries, yet they only achieved a 6 percent response rate.
Get Faster on Social Media
But in general, the report found every industry could speed up their responses and reply to more messages from customers on social media. According to the study, brands only respond to an average of 11 percent of inbound social media messages. Eighty-three percent of social messages go ignored, Caravella said.
“Buy-in needs to come from the top and brands need to employ staff dedicated to monitoring and responding to inquiries on social. Understanding the life cycle of a social message and the interdepartmental needs of customer service on social is critical, and it requires nearly constant attention.”
Sprout Social claims Suggested Replies trumps other canned responses functionalities on other platforms, citing better personalization and faster performance. According to Caravella, Suggested Replies will expand to networks other than Twitter in the future.
“The social manager chooses a relevant reply and then personalizes the response to the specific customer or community member. The feature also teaches and improves itself based on prior responses, so the more it’s used the better it gets at offering suggestions,” he said.
“Engagement is not only about responding to customers, it’s about interacting in a considerate and timely manner.”