The author once heard a vice president of Baidu Cloud say:

Why is voice user interface？

To throw a question first, try to recall, how do you operate the navigation while driving?

There are two broad categories of answers:

1. interact with mobile phones;

2. Interact with in-car navigation.

So, what about the specific interaction method? Use hands, or voice? Which one do you think is more convenient?

Which is better and which is worse is not yet conclusive, let's try to analyze the reasons for the emergence of voice interaction.

The author once heard a vice president of Baidu Cloud say:

Assume "people are lazy after all", temporarily b2b data set this as theorem 1.

Based on this rule, theorem 2 can be deduced: the method that makes people lazy can get likes.

So, how can one be lazy?

1. Reduce human learning costs;

2. Improve human operation efficiency.

Now look at VUI (Voice user interface), voice interaction has the characteristics of extremely low learning cost, as long as the purpose can be clearly expressed in language. However, VUI cannot completely guarantee the improvement of human operation efficiency.

Therefore, the author believes that the most efficient application scenario of VUI is "the user clearly knows the task goal, and the speed of voice interaction is faster than that of contact interaction".

Above, let's go back to the driving scene.

Humans have five senses - shape, sound, smell, taste and touch. From the point of view of safe driving, the line of sight needs to observe the road conditions for a long time, and it is not suitable to frequently hit the wild. In terms of information input, the "listening" of the ear is more advantageous.

After the brain processes information, it needs to output information. Since the "touching" of the hands is limited by the steering wheel, and from the perspective of safety, the advantage of "speaking" is reflected again.

To sum up, the author understands that the applicable function of voice interaction in driving scenarios is: the driver has a clear goal of the task, such as: playing a certain song, and the efficiency of the task through voice operation is higher than that of contact operation.

How to do VUI design?

General Interaction Flow: speech acquisition -> speech processing -> information presentation.

Similar to GUI design, designers need to consider the collection, processing, and presentation of information, but with the addition of voice as an information carrier.

The design process is as follows:

1. Requirements: clarify the needs of the business and the needs of users, and then determine the needs of the product;

2. Flow: Starting from the requirements, sort out the interaction process;

3. Prompts: Design voice prompts for each process;

4. Grammar: For voice prompts, design grammar to collect voice feedback;

5. User testing: After completing the design, pass the user test mobile phone feedback;

6. Tuning: Based on feedback, further improve the product.

Requirements

The definition of product requirements is the core of determining a good product. Only when limited resources are concentrated on the "just-needed" function, users will finally pay. All subsequent design and development are meaningful.

The author believes that in profit-oriented products, demand design always finds a balance between commercial interests and user interests. When it comes to company politics, business interests also include the interests of the leadership and the interests of the team.

You need to know what is the thinking of each department of the company, what is the thinking of the leadership?

This is the core of ensuring the smooth implementation of the project. Second, the product has to be profitable for the user, otherwise the project itself doesn't make any sense unless you want to do a bureaucratic product.

Designers need to consider all aspects of business, company, personnel, users, etc., and rationally balance various factors to achieve a temporary optimal solution.

Since the level of soft capabilities cannot be generalized, this article will only design functions from the perspective of user experience.

Flow

After the function is determined, the process of the function needs to be sorted out in detail. The flow should reflect the priority of the function, and at the same time, ensure that each step of the user has corresponding feedback.

The author feels that at this step, team members can communicate more and discuss more about the process that can be optimized. Improve the process as much as possible to avoid additional costs caused by subsequent changes.

Secondly, VUI designer and GUI designer should start to cooperate more at this stage. In view of the vehicle's central control screen, Kombi display and HUD (Head-up display), a good GUI interface is crucial to improve the user experience of VUI.

When prompting information, displaying information, and prompting errors, the graphical interface can well assist the user to perform voice operations. The two complement each other and are indispensable.

Prompts

In dialogue design, designers need to have a deep understanding of the language, including but not limited to: pauses, intonation, words, emphasis, etc.

Each language has its own unique Prosody, which requires a lot of extracurricular accumulation. The author consciously has limited knowledge, and here I only talk about how to design dialogue from r enters the voice command again: Audi China Building;

1. The system recognizes it correctly: your company's address is Audi China Building, located at No. 2, Jiuxianqiao Road, 798, please confirm.

GUI prompt

1. User confirms company address: Yes.

2. User voice command input: Navigate to the company.

3. System identification, turn on navigation: The company is 3 kilometers away from the current location, and the estimated driving time is 20 minutes. Whether to turn on the navigation.

GUI prompt

If the error is identified more than 2 times, the help message is displayed: you can say "Help" to consult the manual, and "End" to terminate the operation.