I read a recent WSJ Article, "Hate Chatbots? You Aren't the Only One" and it moved me to write this post in defense of chatbots. Apologies to those who don't have access to the WSJ, but I'll explain enough so it shouldn't matter.
Yes, those chirpy, intrusive customer service chatbots drive us crazy. Yes, they often put you into a communication doom loop. But is that the fault of the bot? Largely not.
More often, it's the bad policies or poor implementation of the companies who buy chatbot tech, hoping to save money on live customer service. The WSJ article listed several chatbot weaknesses, which I counter below.
#1: Chatbots Just Send Me to Pages of Dense/Unreadable Content
This is not the bot. The chatbot would LOVE to give you a short, clear answer to your question, but the company you're dealing with hasn't created one. Instead, they've taught their bot to link to the tome of blather it sends you to. There may be a good reason for this, but more likely either the culture or experience level at the organization hasn't thought through how to best serve you, the customer, with the bot it bought to do so. Most bots can summarize massive amounts of text into a clear response using AI, and even allow for human editing to fine-tune the details. But the bot is going to do what it's taught to do.
#2: Chatbots Take Too Long to Connect Me with a Human
Not the bot. From a customer perspective, it seems that the bot is asking endless questions without offering any help. But bots are only as good as the policy that drives them. Example: refund requests. Too often, corporate policies dictate that every refund, discount or replacement request must escalate to a human. But because the same company is trying to save money by using a chatbot, they try to weed out every possible reason why the customer shouldn't speak to a human, via the chatbot. My advice to companies threading this particular need? Experiment with simple, standard refund policies the chatbot could handle. Check cost versus customer satisfaction metrics to see if it worked.
#3: Chatbots Ask for Too Much Personal Information
How many times do I have to give you my account number, order number and great-aunt's middle name before you know who I am? While data security and system interoperability are real challenges in chatbot implementation, the majority of bot tech allows for back-office system integrations that could streamline these irritating aspects of customer and order identification. So - usually not the bot.
#4: Chatbots Don't Understand Common Industry Language
You guessed it - not the bot. Chatbots understand common language out of the box but can be easily taught industry-specific terms. Chatbots are now using AI to learn those terms themselves, but a bot will perform better out of the gate with a little proactive teaching by the customer success team.
How Can You Make YOUR Chatbot Better?
The good news? Chatbots remember everything. The key to a good chatbot is to train it thoroughly, keep it updated and build policies that maximize the value of having it - for you and your customers. Bots can actually eliminate customer service problems, but only if the humans do the back-office work and craft customer centric policies. Chatbots aren't perfect, but a little care and feeding make sure that they are assets and not liabilities.
About the Author
Diane Pierson is the Founder and Chief Market Strategist of Innovate on Purpose, a consultancy enabling successful product commercialization for B2B tech companies. Order her book, How to Innovate on Purpose or contact Diane at dpierson@innovateonpurpose.com.
Comments