Yes. Your self-service bot also requires trained agent support.

3 reasons why well-trained, easily accessible staff are critical for implementing good self-service

The tech-handicap that used to exist once upon a time is now history. Now everyone has a smartphone,  they are tech-literate and there is minimal resistance to the idea of self-service.

Or that’s what some managers are prone to think.

Is there a difference in how managers and customers perceive self-service?

MIT Sloan conducted a survey in 2015 to see how each group differs in their self -service viewpoints. They found that “ Managers significantly underestimated the need for employee interaction during a self-service experience”.  (They also found that while customers want speed and convenience from self-service, managers prioritize accuracy.)

Be it visual navigation/conversational IVRs, or voice bots,  SSTs (self-service technologies) should aim to reduce consumer effort while also providing accurate solutions. This blog post concentrates on the three critical roles that your staff play in ensuring good self-service, especially the kind your bot is designed to deliver:

  1. Easing customers into the new technology
  2. Backing up when technology fails.
  3. Continuously improving the system.

Trained staff help ease customers into new technology

Trust perception is a major driving force for customers to adopt your self-service. Trained agents can help improve this trust perception. For example, when self-service kiosks were first introduced in airports, staff was often at hand to help customers with the new technology.

When your support center starts offering self-service via IVR or voice bots, its useful to remember that not all customers will embrace the new technology with ease. Some may be vehemently opposed to using your bots, while others may slightly hesitant to use the new technology. Managers must provision for more assistance during the early stages of implementation.

  • Regular customers may be directed to self-support systems by your live agent, possibly even a representative they are used to dealing with.  
  • Your agent can remain in “conference” mode during the interaction, offering assistance and guidance throughout.
  • Customers should be allowed to opt out of the self-service at any point. Your system should remember customer preferences, routing customers to their preferred channel. For example, self-service for some customer and live support for other customers based on their preferences.

Trained staff back up when technology fails.

Imagine an SST service that has malfunctioned and there is no way for your consumer to find human help. A recent tweet I read describes this best.  After a visit to Walgreens, this customer got an automated call asking him to rate his experience. The problem? No matter how many times he pressed #1 he would get an “invalid entry” response. Don’t let your customers get stuck in this never-ending loop of customer hell.

Without a recovery option, an SST (Self Service Technology) can be easily perceived by your consumers as a cost-saving measure by your company rather than the customer-centric problem-solving mechanism it should be. As  Forbes contributor Shep Hyken says “ (Self-service) is great … until something goes wrong..Then there has to be a backup plan, and that backup is usually a human on a phone or behind a ticket counter.”

  • The recovery options should be easy to access. Websites or visual IVR solutions should have a ‘click-to-call/chat” option available to your consumers.
  • The onus of ensuring your bot/SST is working should not fall on the customer.
    For instance, during our early experiments, we created a bot that connected callers to the extension number they desired. After callers stated the number, the bot would ask the caller for confirmation, e.g, “did you just say 2345?” This single step would annoy callers so much that they would either incorrectly say “no” (even when the number stated was correct) or hang up.
  • During voice interactions, if a bot fails to understand the customer within 1-2 tries, an inbuilt mechanism should connect the customer—on priority—with a trained human resource who not only provides a solution but also acts as a reassuring factor.
    For example, if callers ask your bot to connect them to some department, the system should connect them immediately. If mistakes occur, the agent who accidentally gets that call should politely (and very quickly!) transfer the caller/ arrange a callback or handle the customer complaint.

Trained staff help systems learn and improve

According to research conducted by scholar Joel E. Collier, the key determinants of consumer satisfaction are accuracy and speed. Improving accuracy, like improving customer experience, is a continuous process.

If self-service is delivered via an AI-driven intelligent voice/chatbot, then over time, it can keep learning and improving. For instance, take the case of an SST that handles FAQs via a bot. Sometimes your bot may have the answer but may not recognize the question. Maybe it is grammatically incorrect or just asked in an unexpected manner. Either way, if your system is not designed to answer this question, it can nonetheless learn from it. This question can be tagged, and a data scientist can investigate and help the bot learn from this experience.

Over time, agents’ inputs will keep improving your self-service bot.  This can be through disposition notes or through their call recordings.

Let’s consider a more complicated self-service system. For example, a system that assists customers while booking holidays. The conversations that can take place are limitless. When the bot can’t answer, it will fall back on an agent. This conversation that takes place between agent and customer can then be fed back to the bot for future reference.

In conclusion

Finding an answer without another person’s help is a pleasure. When people know they are talking to bots, they explore a lot more. For example, when we use search engines, we feel free to ask questions. We know we aren’t wasting anyone’s time, or being judged on dumb questions. Think of the customer’s satisfaction when your bot/SST can answer their limitless questions. With a clear vision of how your SST can supplement your business and a little agent training, you can aim towards creating that level of customer satisfaction.

*MIT Sloan, “ How Customers View Self-Service Technologies”

2019-02-21T05:35:36+00:00

About the Author:

Sudhamay is a project manager at Ozonetel Communications. He is our go-to person for all things AI.

Leave A Comment