What Makes a Good Bot or Not?
Posted by Thomas Staven
Bots are popping up everywhere from Facebook to home personal assistants. Advances in natural language processing, machine learning and other AI technology created the foundation for bots, but the field has a long way to go before it reaches its full potential.
The Alexas and Cortanas of the world do an effective job at accomplishing requested tasks as long as people present them one at a time. A multi-threaded version of these digital personal assistants would allow them to remember multiple situations. This use case is closer to how people actually want to engage with the bots. Multiple conversations could go on at the same time, with each thread at a different stage in the process. This experience would be much better than the current implementation, and it's likely that they will move toward this ideal during advancements in 2017. Users should also gain the ability to choose text and voice interfaces, rather than being limited to one input.
Add More AI Functionality
The AI-part of these bots is limited to natural language processing and basic functionality tied to fixed data. Bots need the ability to become smarter and more proactive to enable more meaningful communication. For example, they learn user preferences and behavior to offer the right services at the right time. Without this, two-way communication can't happen.
Sometimes Generic is the Right Answer
Many vendors personalize bots by giving them human avatars or adding a personality. The reception to this type of personalization is mixed, with the older generation and business users preferring more utilitarian and professional implementations. "Faceless" bots are the better option during this early adoption period, with the exception of bots intended for younger generations.
Customer Awareness Is Important
Some companies try to pass off bots as human staff. The best practice is to be upfront about the fact that a person is talking to a chatbot. In front-end customer service, some people may hesitate to describe their problems if there's a bot on the other end. Approach this situation by giving the customer a choice, so they can make an informed decision.
Clarify Bot Functionality
AI bots are still new to many people. It's important to specify what the bot can help with from the outset. Organizations set the expectations early on to avoid a negative customer experience. They keep trust in place with the client and limit the chances for disappointment.
The bot quality also needs to be excellent in customer service applications. If every response is "I don't understand the question" then it's useless to your audience. Organizations need to carefully consider their chatbot choices and establish an escalation channel to provide a direct route to human employees. The bot also needs the ability to detect negative sentiments. Many customers are dealing with frustrating issues and a lack of empathy would make the situation worse. The bot can start smoothing over the situation and send them to a customer representative.
AI technology is evolving fast. In a few years, bots will be so intelligent that people can't tell the difference between a bot and a human. Widespread adoption will normalize bots being the first contact point with a company, rather than an unexpected situation. Once the technology gets to that point, the best practice will be to show when customers are talking to a human, rather than a bot, and it's likely the bot's capabilities will exceed the client's rep's.