If you build it, the bots will come. Just as phones are flooded with robo calls, digital channels are under attack. It can sometimes be extremely difficult to sift through the digital garbage to actually see a customer message that needs to be acted on. And once identified, it can be hard to make sure that conversation goes to the right contact center agent. Fortunately, HelpSocial has improved customer support workflow with our intelligent automation tools that allow you to tease out the signal from the noise.
But HelpSocial’s intelligent automation started with humble beginnings. We faced the problem of how to surface customer conversations over bot-produced spam while working on the social media team at Rackspace. Fortunately, early versions of the HelpSocial customer workflow logic were key in keeping the bots at bay while making legitimate customer conversations visible.
HelpSocial’s Customer Support Workflow Logic
While at Rackspace, we identified several scenarios where our social channel feeds would be flooded with messages that could distract service agents from being helpful to customers. They included:
- Financial news and earnings reports
- Major product announcements hitting the tech/mainstream press
- Internal blog articles that picked up steam
- Hacktivists targeting social issues
If you’ve never worked in customer support, you may be wondering what bot traffic has to do with these scenarios. To put it simply, many bots—or even legitimate users trying to boost their social standing—programmatically cue their social accounts to retweet/rebroadcast messages based on:
- RSS feeds of news organizations
- Popular social media users
- Particular stock symbols
- RSS feeds of company blogs
Even today, when a popular news outlet such as TechCrunch posts an article, it’s immediately reposted by hundreds, if not thousands, of bot accounts. So when Rackspace appeared in the headline of one of those articles, our social support team was completely bombarded with messages. This can make searching for a legitimate customer conversation a bit like seeking out that proverbial needle in a haystack.
This same situation occurred with financial earnings calls, company blog posts that went viral and even when the brand was caught up in “hacktivist” situations (for example, Rackspace was often mentioned as the hosting provider of certain websites that some people took offense to).
Filter the noise
Please don’t misunderstand: from a marketing and public relations standpoint, it’s important to capture all messages. However, most of these mentions are detrimental to a social support team actively trying to assist customers with problems as quickly as possible, so you need a way to filter out that noise.
This is why we introduced our workflow automation. When an event drives a lot of social mentions, a HelpSocial administrator can define different kinds of logic to help route those messages away from the contact center. The logic can be based on Boolean operators, paying attention to terms within the social media post as well as the text in a URL.
Furthermore, administrators can create logic based on attributes of particular users. For example, a user with a lot of followers can be surfaced to the top of the queue and flagged as someone to address immediately. This process can be paired with suppressing users in the support queue with a low follower count or those identified as spammers or detractors.
This kind of customer service workflow management was crucial while we were at Rackspace, but we have taken the HelpSocial platform to the next level with the release of next-generation intelligent automation.
Improving Customer Service Workflow with Intelligent Automation
Today, HelpSocial is using the latest AI technology and contact center software integration to help more customers in less time, at a lower cost. The HelpSocial platform can now be configured to analyze customer intent in inbound messages across digital communication channels. This can include understanding:
- Sentiment of the conversation
- Topic, category or purpose for reaching out (message intent)
- Other relevant contextual data from internal systems, such as previous ticket history
All of this data is packaged around each message in real-time, helping the contact center system decide how the inbound message should be routed and if it should be escalated to the top of the queue. The customer conversation could either be sent to a human agent or to an AI-driven chatbot designed to automatically assist with the question.
If it should go to a human agent, the intelligent automation can accurately predict which person will be the most successful in handling the situation and presents that agent with all the necessary contextual information. If the message goes to a chatbot, our intelligent automation gives that AI entity better context to help tailor the response—instead of starting at the first step of a scripted process, the chatbot could skip to a later step.
HelpSocial intelligent automation is the customer support workflow that we could only dream about while working at Rackspace. It was born out of our personal experiences of being on a social media team and was built with the frontline agents of a contact center in mind.
Want to Read More?
Be sure to check out the rest of this series, in HelpSocial as a Case Study, where you’ll learn more about the real-world use cases that led to the features and tools in our platform, including how our tagging feature changed the social sentiment on Rackspace overnight and why we created an open API.