Agents Need Virtual Assistants Too

    

How to Use Bots Inside Your Contact Center to Decrease Handle Times and Improve Performance

 

Some customer and product support teams at iQor have a new colleague that seems like a perfect fit—friendly, helpful, eager to improve, and get this: available to everyone, at the same time, 24/7. 

 

That’s better than perfect. It’s virtually perfect. The new hire is a chatbot, powered by a knowledge base of answers to typical customer questions. The data was compiled by experts at iQor and two client companies.

Businesses have deployed chatbots and other self-service tools with varying degrees of success. However, a virtual assistant/coach for customer care and product support professionals is something new.

 

It’s also something badly needed. Customers accustomed to instant responses from Alexa and Siri expect a fast fix to even a complex service issue. Agents need precisely targeted information at their fingertips so they can craft creative solutions quickly—before the customer hangs up and switches to another brand.

 

Born to Deliver Speedy Service

iQor’s virtual assistant, dubbed Q-Bot, was born in late September 2017, a brainchild of iQor’s Experience Innovation Lab. “We wanted to develop digital innovation that would automate shared knowledge and bring it up close to users quickly,” says Lukasz Lubas, a key member of the Bot development team based in Bydgoszcz, Poland. “It’s unique—artificial intelligence they can talk to.”

To use Q-Bot, agents type in keywords as they speak with the customer on the line. In seconds, Q-Bot can zip through thousands of pages of information, follow algorithms to find a solution and text it back to the agent.

 

How does a company go about creating a crackerjack silent partner for customer-facing support pros? Lukasz Lubas and his team learned some important lessons during Q-Bot’s beta testing.

 

Give Your Bot a Friendly Personality

Q-Bot is brainy, but it’s no smart mouth. Its responses are friendly and respectful. “There are some very nasty bots in the movies, but we built ours with a nice personality,” says Lubas with a laugh. “Our content philosophy is that everything Q-Bot says should feel like it comes from a friendly supervisor or helpful colleague.”

 

Choose the Content Team Carefully

That means the bot must use language and phrases that sound natural, not canned or mechanical. The problem is, bots aren’t born knowing how to speak human. They can only process information and apply algorithms. So a content team that understands technical knowledge and can also write conversationally is essential to transform what the bot “knows” into crisp, concise, actionable answers.

 

“Agents and techs are using this on the fly,” explains Lubas. “They have to be able to grasp the meaning of the words immediately, almost without thinking, and use the information to craft a solution for the customer.”

 

Teach Users How to Teach the Bot

Inevitably, and sometimes comically, bots will misfire the wrong answer. iQor’s Q-Bot has self-correcting tools to help. Every time it serves up an answer it asks for feedback: “Was that helpful?” Through the feedback Q-Bot “learns” which answers are right, growing its knowledge base/brain. When answers are labeled unhelpful, the content team works to improve Q-Bot’s knowledge, prioritizing the most frequently asked questions.

 

Keep Building a Bigger Brain

There are lots of plans for Q-Bot. “Eventually we could connect the bot to every knowledge base,” says Lubas. “We could eliminate a lot of information gaps, not only for agents and technicians but for all our iQorians. Engineers could search through countless internal and external databases using only one search engine. Anyone in any department could get data on demand.”

 

Keep the Path Open to Machine Learning

Q-Bot has many ways to grow. It could eventually be deployed externally as well as internally, and it could use voice as well as text. Lubas says, “At some point soon, we will augment basic process knowledge we have uploaded to its memory with Machine Learning techniques so Q-Bot can learn on its own. But first we want to show it how we want it to behave, apply rules and algorithms, and check if it is making an impact in iQor’s business.”

 

The possibilities are virtually endless.

Bots can be brilliant additions to customer and product support, but Artificial Intelligence can never replace human empathy and problem-solving skills. Learn how the entry of chatbots is changing the world of customer care, and how to integrate human and digital ecosystems for the best, most customer-centric solutions. Check out this related blog post, Becoming Digital: Rise of the Augmented Agent