This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
Good Bots, Bad Bots, No Bots
Earlier this week at Enterprise Connect 2017 in Orlando, Fla., I entered the wrong session room -- but stayed anyways as I was immediately drawn in by the presentation, "Turning Features and Functions into Customer Experience Best Practices," by JR Simmons, president, COMgroup.
His first important comment, in the form of a question, was: "Are you making decisions to benefit the customer, the agents, or the organization?" He presented a number of recommendations for and actions to avoid in the contact center. I was particularly interested in his discussion on bots.
The introduction of artificial intelligence in the contact center leads to chat and voice bots. Bots, like Apple's Siri and Microsoft's Cortana, are software applications (Web robots) that run automated tasks (scripts). Bots perform simple repetitive tasks much faster than would be possible for a human alone. Bots simulate human interactions. The dialog feels like you are conversing/chatting with a human -- or at least in the ideal.
In IVR systems, the traditional contact center has had the equivalent of a simplified bot idea. The IVR is aimed at providing information to guide the caller to the right person or function within the organization. Sometimes the phone trees are quick and easy, other times reaching the appropriate resource takes forever. And sometimes you end up never being able to get to an operator for help solving your problem.
Bots are an alternative to the traditional IVR system, which is built on the concept of self service. Bots are adding AI improvements to help change how we interact with self-service systems.
Bots, however, can encounter the same problems we have with IVR systems. I had to deal with a government agency that had such a long IVR menu that I had to access the system four times, for 30 minutes, and take notes just so I could finally get to the point where I could use the self-service prompts. Even so, in the end I still had a talk to an operator.
Bots can encounter the same problems. Bots, combined with AI, can learn to work better. However, this is not always true. A famous mistake with bots was Tay, the teenage Microsoft chatbot that wound up tweeting racist and offensive messages. Tay lasted online for less than 17 hours.
In his presentation, Simmons detailed the attributes of a good bot and bot best practices. They are:
- Contact centers should notify callers when they are dealing with bots. This should alleviate frustrations callers may encounter during interactions, and give them an understanding of how replies are being generated.
- A bot should solve problems and provide answers promptly.
- Abot should be able to take proper action for standard requests and avoid the caller having to deal with an agent.
- The calls should be routed accurately and rapidly to the proper activity or agent.
- A bot can automate the authentication process and save time for the agents.
- You can track the caller's journey, displaying the actions of the caller to the agent when needed to better assist agents in their responses.
- Bot use should be optional, not forced.
- The bot interaction should be easy to exit.
- The bot should not keep trying when not being successful with a caller.
Any technology can have its problems. Bots are no exception. Companies implementing bot systems are still learning how to use and refine them. To prevent offering up a bad bot, Simmons had this advice:
- Do not try to pretend your bot is human.
- Not everyone communicating with a bot will articulate requests or requirements in the same manne. Understand that bots can misunderstand a request.
- Don't try making the bots do too much. You still require agents for problems a bot can't solve.
- Don't frustrate the caller by continuing to force actions with a bot and preventing the caller from getting live help.
The bot is there to help the customer, not the business. The idea is to make the experience for the caller easy, useful, and successful.
Simmons left some final pithy comments for the audience, and I'll do the same by sharing them here:
- "Bad customer experiences produced by misguided use of technology is worse than no change at all."
- "Buying technology will not improve customer service."
- "Properly implementing technology based on customer needs is the key."