When thinking about a user interface, chatbots with artificial intelligence and machine learning are up-and-coming technologies. It is known that many government agencies — federal, state and local — often have longer-than-normal hold rates on the phone and long office wait times. Navigating their websites can be time-consuming. Hence, short and quick interactions with a chatbot can provide users more precise answers to their questions immediately, reducing wait times, phone calls and research into additional questions.
Having to renew my own license in the near future, I began asking “Miles” the chatbot some questions through the Department of Motor Vehicles website. First, let me say, it’s great that the DMV is reaching out to the public with a new and advanced technology. This is a great step in the right direction to balance California residents' lifestyles and acknowledge the value of personal time. It is also pleasing to see that the DMV added some pun into the new system! For example, the name Miles, or the response to unknown questions, “Sorry for spinning my wheels. ...”
However, I believe Miles could be a bit better with some answers. For example, I asked Miles, “Where do I go for a Real ID?” He was again “... spinning ... wheels.” The good news: I was provided suggestion buttons, which took me to more information, like what documents I need. One way to speed up (no pun intended) the learning of such systems is to track unanswered questions — see what people are asking, and then provide the system some feedback as to what the answers should be. If Miles uses machine learning, it will take time to build up better navigation automatically. At the end of my few questions, I did get enough information to know what I need to do next: Get the required documents and schedule an appointment.
Previously, I would have had to research the website, probably using up an hour of personal time. With Miles, I was finished in under 5 minutes.
It is not easy to get the wheels to hit the asphalt with chatbots. Often they are not viewed as providing value because users do not see immediate returns. It would be nice to see Miles extend into other mobile systems, such as Google Assistant or Alexa, which would allow customers to use mobile and home devices without the requirement of a personal computer. It is not easy to design a back-end architecture to support all platforms — however, this is possible. At Placer County government, we have been able to deploy a single back-end architecture to support multiple chatbot resources, such as Google Assistant with Dialogflow, Amazon Alexa and a web user interface much like the Miles solution. From this experience, I think it would be an attainable task to extend Miles out into the other scenarios.
Expanding into other platforms also enables better user interaction, both directly with residents and indirectly with corporations requiring information. Someone recently asked a question that made me think more about how a chatbot would really support a resident. After some thought, I understand the simplicity of it now: Simply put, reduce the customer's wait time — as Miles is designed to do — by providing faster and more direct information.
It becomes a pay-it-forward concept with contractors, trucking companies, and many others by reducing rates, labor costs and time spent dealing with government agencies. In theory, these reductions would be realized by lower cost for services to a resident.