It starts with a lie.
“Hi, I’m Stacy,” says the pop-up avatar in a speech bubble, in the lower right corner of your screen. But we all know that Stacy is a bot.
It often gets worse from there. “Stacy” doesn’t understand what you want or need.
“She” isn’t aware that you’re (choose all that apply) 1. Pissed, 2. Confused, 3. Frustrated, or 4. Not a fan of bots.
This helps explain why people are actually happy when a chatbot fails to help them.
On the other side of the scale is this hard truth for chatbot-haters: They work. Especially at the “first touch” level of customer service – where rudimentary questions are asked and answered – chatbots keep website visitors informed and moving deeper into the sales funnel, lowering abandonment and bounce rates and boosting conversion.
The key to success is setting expectations before investing in a chatbot.
Here are 5 takeaways.
1. Determine if a chatbot is the right tool for the job. Harvard Business Review reminds us that “Virtual agents are most effective in customer service applications in service-heavy industries like financial services, retail, travel, and telecom.”
Example: Butterball’s automated Turkey Talk program fields 20,000+ questions during the holidays. Perfect for a chatbot. So before investing in a chatbot, do a top-down analysis of your customer service workload and see if you have the scale to justify the resources.
2. Craft a job description for the chatbot, just as you would an employee.Ultimately, chatbots are like anyone who works for an organization: They need focus and clarity around their scope of work. If you don’t start with a specific job description for a chatbot, you’re setting up yourself for disappointment. This tees up the next takeaway.
3. Focus on the 10 most-asked questions. Chatbots are at their most effective when they’re delivering targeted FAQ-level responses – although quantum computing will almost certainly change this (keep reading).
4. Hang on to your human experts. The informal 80/20 rule of chatbots is that 80% of customer service inquiries may be handled by a bot, but the other 20% need – and expect- to have access to a human. This opens an opportunity to position your HR assets as wizards, heroes, concierges, and premium-level support with a pulse.
5. Leverage NLP. Natural language processing is an AI program that supports a chatbot’s capability to grasp natural human diction. NLP-based chatbots and able to infer customer emotional states, attitudes, and frustration levels as well as idioms, slang, and sarcasm.
Get ready for quantum chatbots
The promise of quantum computing is that it will enable chatbots to handle significantly more complexity. While a chatbot will never be able to “care,” future versions that are powered by quantum computing and AI may become meaning-aware and able to infer customer emotional states, attitudes, and frustration levels as well as idioms, slang, and sarcasm.
When you’re ready to start using information and creativity to solve marketing problems, contact us.
