Comment on page
As with any application development, it takes a lot of people and areas of expertise to define, design, build, test and launch a conversational experience. On the people side, two main factors for success are to have the right skillsets available and to have an extended team that has an understanding and awareness around conversational experiences.
It is critical for the success of a conversational assistant that the business and executive teams have a level of understanding and awareness when it comes to conversational AI. There is a lot of hype and a lot of misinformation that can lead to unrealistic goals. When goals are not realistic, the risk of failure increases dramatically.
A product team who doesn't understand conversational experiences and the technology behind it may have a difficult time defining a product that can be successful. Similarly, a qualitative UX researcher who is not familiar with conversational experiences may not be as effective in running moderated studies (asking questions, distilling insights) as someone who has that familiarity.
Conversation design is an interdisciplinary field and requires an understanding of UX principles, including design, strategy, writing and research. A background in linguistics is very helpful and any background in natural language processing (NLP) and computational linguistics will help you understand the bigger picture of building conversational assistants. Knowledge of cognitive science or psychology is useful. Specifically for customer-service assistants, an understanding of customer service and call centers can be helpful, and a deep understanding of the vertical as well. Healthcare and fintech are two main areas for conversational assistants. They are heavily regulated and understanding that subject matter helps with designing appropriate conversations.
In addition to these hard skills, soft skills such as communication, accountability and collaboration are as important as they are in many other jobs.
This may seem like a daunting list of capabilities, and it is understood that a single person won't have all the hard skills. However, be prepared to be curious and learn continuously in this emerging field.
This is a broad category, with at its core the need for natural language understanding. Technology is fast developing in this area, and depending on the type of conversational experience additional technologies such as speech recognition in voice assistants may be required. Suffice it to say that this type of knowledge is required in a successful conversational AI deployment.
Integrations with existing systems require engineering effort, and this can be frontend or backend engineering.
As is the case with other digital products and services, testing is critical to success. In immature technology areas such as conversational AI, this is often "forgotten", and the MVP is used to "fail fast". However, applying testing processes from software development has to become an inherent part of developing conversational experiences.
Note that in this section we are describing skillsets as roles. It is possible that a conversation designers is also responsible for NLU modeling. Or it's possible that a conversation designer only concerns themself with the conversation structure, and that other individuals are responsible for writing and NLU modeling.