Chat for Support: Some Things to Consider

Chat, a.k.a. live chat or web chat is still gaining in popularity. We hear of organizations adding it every day, and are asked about chat good practices on a regular basis.

Chat is really quite a remarkable channel in a lot of ways:

• It’s a quiet channel—you can use it from your local coffee shop or co-working space (or a meeting—we won’t tell) without disturbing others or having your conversation overheard
• It’s both synchronous and asynchronous—you can have both parties chatting in real time or responding when they can
• It’s flexible—you can send links and/or graphics and even files, depending on the chat client software you use
• Experienced analysts can handle more than one chat at the same time

In 2017, it is becoming commonplace to put a chatbot in place for basic triage before a user gets to a live analyst. This method mirrors the basic choices presented by most phone systems: “Press or say 1 for….” Instead of pressing or saying 1, the bot may present the user with three to five choices that help determine whether it’s an urgent need, whether it’s an incident or a service request (“Is something broken, or do you need something new?”), and, in these days of burgeoning Enterprise Service Management, whether it’s IT, HR, facilities, or someone else the user needs to work with. The chatbot can also answer basic, repetitive questions very fast—much faster than a skilled analyst can copy and paste from a canned response—and hand off to a live analyst seamlessly when it doesn’t have the answer.

It is becoming commonplace to put a chatbot in place for basic triage before a user gets to a live analyst.

There is, however, one very interesting statistic about chat, according to the HDI 2017 Technical Support Practices & Salary Report: One-third of the respondent organizations move about half (between 41 and 70 percent) of chat conversations to another channel before resolution. (That compares with none of the organizations moving email tickets at the 21–30 percent level, and one-third moving less than 10 percent.) This stat indicates that chat is certainly not perfect for every type of interaction—in fact, it’s apparently not perfect for quite a few.

This points us to number 2 in Donald Hasson’s 10 Steps to Implementing Chat Support blog: Identify the ideal issue types for chat. Hasson says, “Review types of issues and identify how the features of chat could hurt or improve efficiency and resolution times for customers.”

That, in turn, brings us to the inconvenient conclusion that—like any other channel—chat requires planning and work, and those should begin with conversations with your customers.

Before you invest time, effort, and money in implementing chat, you’ll want to know:

• What do your customers think of chat? They’ve doubtless had experience with chat in the consumer world. Do they like it? Hate it? Just because chat has a high customer satisfaction rate in other organizations doesn’t mean it will work in yours. Ask.
• What problem will chat solve? If you’re thinking it will miraculously lower costs, you are likely to be disappointed; it may or may not. But it may solve another problem, such as your customers’ unwillingness to call you because they wind up waiting on hold.
• Do you have the right skill set on your team? Talking on the phone and writing in chat windows is not the same. Do you have staff who can read quickly, understand, and respond accurately via chat? Can they handle more than one conversation at a time?

There will also be work associated with any automation you wish to implement. Chatbot scripts need to be written. Procedures and processes need to be mapped out so that the machines know what they are supposed to do.

By all means, explore the flexible abilities of chat as a support channel, but remember that there is no “silver bullet” and there is some hard work associated with this, as with any implementation.