Behind every great chatbot there’s a great Conversation Designer
Role and responsabilities of whom give a voice to chatbots
Have you ever wondered how chatbots can understand natural language and answer using the same linguistic code?
The answer is simple: behind every chatbot there’s a human being or a team of human beings (often) called Conversation Designers, who give them a human voice.
Conversation Designers: who are they and what do they do?
The bases of any frictionless interaction are the ability to understand unambiguously and the ability to answer accordingly.
Human-Chatbot communication makes no exception.
A chatbot understands messages thanks to NLP technologies that analyze and interpret natural languages using complex Machine Learning algorithms. Conversation Designers (working as Computational Linguists) train these algorithms, feeding them with pre-analyzed set of texts that work as the examples that the machine uses to “learn".
Chatbots, then, have to give the correct answer, i.e. the most relevant one among those stored in its database. Moreover, answers should sound conversationally natural and it must be easy to reach them through guided path or free questions.
In order to do so, conversational interfaces need a voice, a personality and a context awareness similar to humans’. Conversation Designers plan these characteristics in order to create a customized Conversation Experience and a captivating UX.
To sum up, a Conversation Designer is a Linguist who deals with the NLP component of a chatbot, who designs its Conversational Flows and who writes the most appropriate copy for each of these flows.
Let’s see how.
How to give a voice to a chatbot in 3 steps
What does it mean to give a voice to a chatbot? It means dealing with its communicative level, and therefore giving users the answers they are looking for, by engaging them in a pleasant dialogue.
STEP 1 — UNDERSTANDING what the user is saying
Understanding what the user needs is of primary importance if we want our chatbot to be successful. Matching the correct intent of the user, though, is not an easy task. Even if NLU (Natural Language Understanding) technologies can perform a semantic, syntactic, phonetic and morphological analysis of written and spoken utterances, pragmatics still remains a real challenge. Since human languages are formally complex and highly dependent on context, a request can be expressed in so many different and ambiguous ways, that is highly unlikely to map them all.
Conversation Designers, though, try to find patterns in user requests and feed NLU algorithms with examples that will help them associate requests and answers. In order to do so and depending on the NLU system used, different actions might be needed, focusing more on one or another level of linguistic analysis.
STEP 2 — ANSWERING: guiding the user to the answer
The second step in planning a conversation is designing its flows, and so guiding the user towards all answers in store.
First of all, information can be provided in two ways: answering freely asked questions or leading users through specific paths. These two methods are compared through a SWOT analysis in my previous article and can be combined to provide an even better Conversational UX.
In this step, Conversation Designers plan tree-structured paths based on a careful analysis of the domain and of the specific topics dealt by the chatbot.
They can also customize these answers on the basis of external factors, such as the website page the user is visiting or the moment of the day or of the year, and on the basis of users’ personal data, like language, age, gender, number of orders or frequent purchases.
STEP 3— ENTERTAINING: customizing the chatbot’s personality
Is it better to have a formal or an informal virtual assistant? A serious or funny one? To be respectful or ironic? Is it better to use a polished, eloquent language or a simple, straightforward slang? Is it better to give brief, practical solutions or to entertain with jokes and puns?
The tone of voice, of course, mostly depends on the brand you’re representing and on the target of users you’re talking to. Writing conversations for a chatbot, though, doesn’t simply mean paying attention to correct grammar, appropriate tone and accurate lexical choices, but also playing with words, with cultural standards and with shared knowledge, so that the bot’s answers will sound as natural as possible and the conversational experience will be perceived as well targeted.
There is no perfect Conversational UX, only perfectly flexible Conversation Designers
Finally, when you design and write conversations, you need to keep in mind that the perfect Conversational UX does not exist (yet). Each chatbot has to be customized depending on the clients’ and customers’ needs and on the technology used.
Moreover, Conversational UX is still a newborn field, so the only real rule is to learn from experience in order to identify the most satisfying best practices.