Bot Enthusiasts Series: Akemi Tazaki
Our blog interview in March is with Akemi Tazaki. Akemi is an experienced designer in user centered design for mobile, web and consumer electronics. She is absolutely passionate to learn new cultures and skills. Here are some questions we asked her about chatbots and conversational UX and her answers.
Botanalytics: What kind of challenges do you think conversational bots have nowadays and what will be the challenges in the future? And maybe solutions?
AT: The current implementation of bots in most of the industries, such as in customer service, are not as intelligent as media has covered. IBM Watson commercial has set a very high expectation of what is “bots”, or the artificial intelligence to the mass.
We need first to establish the right expectation on what it does for the users by giving an instruction what it is meant to do instead of being an all purpose chat on any subject.
Is that a real-time? How long do I need to wait to start a chat? Is it a person behind or a machine?
A conversational interface is meant to be real-time. Otherwise it is a disguised email. What about the personality, the styling of the text? It demands to refine the way we answer to the users beyond the way Dialogue text is formulated. We design and develop the “soft skills” of the bots.
The challenges are not just for people developing the interface, but also the organization which needs to define the scope to shift from other media (email, web forum, phone) to chat with all the complexities of organizational structure. When integrating the technology, we discern tasks that are better accomplished by a bot vs. by a human.
Botanalytics: How do you think a bot can create a value or can it change the image of a brand in people’s mind? How can conversational UX help them?
AT: The conversational interface is an agent that represents the company all the more than other digital media through which the company projects their image. Conversation sets an environment where users’ get an impression that their needs are addressed personally. Therefore, a company can leverage to set the right response style through to gain trust from users: the tone of voice, contents, the balance between acknowledgement response, clarifying questions, statement, etc.
Theoretically, a harmonious tone of voice of the agent is achievable as there is one machine agent, instead of training multiple human agents to give a quality “conversation” with the users. I say “theoretically” because it depends much on the quality of artificial intelligence you put behind the bot interface and tasks the AI is supposed to do.
UX as a multi-disciplinary team can assist the company to evaluate the tasks that are done by machine vs. humans, to program meaningful AI, and guide the branded “soft skills” of the bot.
Botanalytics: What makes a bot unique considering user experience and usability?
AT: As I mentioned earlier, the conversational interface gives the feeling of having a personalized interface with the system which is a solid starting point to develop human and machine relationship. We do not look at the screen, so no more button pressing. It can be a huge leap to make information accessible to [disabled people>>> Politically correct way to formulate?].
Since usability relies upon the performance of natural language processing and intelligence of algorithms, the industry challenge is to accept to work with multi-disciplinary design team to produce the “soft skills” of bots.
Botanalytics: Do you think a bot should have a personality? Should they behave as much as a human?
AT: Of course, they should. Behaving like the human? Hmm again what bot does as a job for the human? Human is notorious for making mistakes in repetitive tasks or emotions influencing the performance of mechanical tasks, but once we let machine deals with tasks that are better performed by them, we need to rethink what are the human’s contribution to an organization.
Botanalytics: What are your insights and suggestions for bot makers to enhance a conversational flow for maintaining a successful chatbot?
AT: Clarify from the onset why and what tasks you are delegating to the chatbots interface. Is that the medium solves your real problem for your business? If you do not know the answer to the “why” and goals of the chatbot interface, you are building a “vanity project” of little value to the end users or business.
Try to set a narrow task responsibility at a start for the bot so that conversational flow is in a limited area. Give an example of what kind of conversation you can have with the bots you provide.
In your real life, pay attention how you converse with people, in what contexts the conversation is happening, the response you give and how effective some people are delivering information orally:
Nodding to give queue that you are listening, gesturing to emphasise the importance, etc., the way you introduce yourself, the energy you give to the other person, what is considered to be polite and not polite, etc. All these little things we do unconsciously in a real-time conversation is lost in the conversational interface if you do not deliberately express by texts and icons (animations, emoji, etc.).
Botanalytics is an analytics & engagement tool for bots based in San Francisco. If you would like to feature on our blog interview series, just shoot an email at firstname.lastname@example.org