In-Vehicle AI Agent

Product design for NIO drivers to better plan their days, where the in-car AI assistant would be able to memorize and recommend possible short detours the driver can make to do something else while commuting.

Team

Michael Li, 2 Designer Mentors

Time Frame

7 Weeks

My Role

Designer, Researcher

Tools

Figma

01 CONTEXT

I worked on this project for my internship at NIO. During this internship, one of the main focuses of the team was with the development of assistive driving, drivers will have more attention span to preform more complicated interactions while driving. Thus begins the trend of integrating in-vehicle service into the car that will work better than simply using a phone.

There were two main focused areas of the project

02 THE PROBLEM & SOLUTION

The Problem

NIO drivers often struggle to balance leisure against work, are tired about their repetitive routines , and crave more opportunities for leisure and relaxation outside of work.

The Solution

Users would first

Record Interesting Places

and keep a record of favorite places

Cars will

Recommend Simple Tasks

based on the driver's current location & their favorite places

Users Can then

Schedule Reminder to Visit

their locations of interests

03 FINAL DESIGN

Exploration & Scheduling

Drivers are able to quickly ask for information about a place through the voice agent, and schedule future visits while driving, so the vehicle would be able to plan something for the future

That bakery with the red door looks so cute! What do they sell inside?

Are you talking about MATCH bakery & Co? They mainly sell various matcha products, including cakes, bread, and drinks!

Other user highly recommends their succulent matcha mille crepe.

Reminders

After planning a future trip, the vehicle would remind the user at appropriate scenarios

You previously mentioned wanting to check out Matcha & Co to buy breakfast for tomorrow morning. Do you still want to go?

Recommendations

Using the driver’s collection of favorite and most frequently visited spots to create recommendations for interim stops that allow the user to quickly finish tasks like buying tomorrow’s breakfast, coffee, and other small tasks.

Is there anything I can do right now?

Let me see. On your way home, you could buy breakfast for tomorrow, get dog food for your dog, go to the gym for a workout, or relax at the park.

Buy breakfast

Got it! Do you want to buy breakfast at the MATCHA bakery you saw this morning, or the Yuan Ji Bun Shop near your house?

The one next to my place

Got it! Heading to <YuanJi Bun Shop> about 2 minutes from your destination. The drive would take approximately 10 minutes.

04 RESEARCH

Starting out with such a broad statement, I would need to understand what I am designing and who I’m designing for. I looked into 3 main areas, competitiors, the market in general, and specific users.

Existing Market Trends in EV

GPT Integration
Natural interactions between driver and AI-Agent
Personalization
Drivers expect Vehicles to be catering to specific needs
Assistive Driving
Improved assistive driving abilities, giving drivers more freedom to do other things while driving

Who am I Designing for?

Main User Insights

Users find it

Hard to Balance Professional Demands & Leisure

Users want

To Visit and Explore in The City

but don't have the time to schedule visits

Users hopes that the vehicle can

Better Their Time Management

05 IDEATION

What abilities do current NIO vehicles have?

As aforementioned, one of the main goals of the project is to create a product that would create a better experience than simply using a phone, which means the design would have to integrate these features that currently exists in NIO’s vehicles.

AI GPT Integration

Able to perform complex calculations
Remember driver data and personalize experience

Map & Navigation

Understand driver’s basic schedule through location
Allow location context aware abilities
Guiding drivers to important locations

Assistive Driving

Allow drivers to divert some of their attention to other tasks

Interior & Exterior Sensors

Understand the context within and outside the vehicle
Computer vision, sound sensors

In What Scenarios will the car be better than the phone?

With these insights from our research, we brainstormed a range of potential product ideas that could genuinely improve users' day-to-day experiences.

A big goal of this project is to create something that feels like an upgrade from just using a phone while driving, something more immersive and intuitive.

To achieve this, we focused on building a design that seamlessly integrates the unique features already available in NIO vehicles.

There I ideated multiple possible scenarios that a driver would demand some sort of service to improve their driving experience, these include:

screen shots of ideated scenarios in the document

Defining Service at NIO

After the general exploration, the scnearios are classified into 7 general user needs.

These categories are not just used for this project aline, but would be a guide/inspiration for future service related projects

Further, we created 3 main strategies that we want to follow both based on the researched insights and the general guidelines of NIO’s brand guidelines. 

images from meeting white board discussions

These strategic guidelines would also be informing other service related projects at NIO

05 DESIGN

Scenarios Explored

From the insights from the user and the general user needs we explored, I’ve mainly wanted to explore the concept of accomplishing task during daily commutes. 

Scenario 1:
"Missed Opportunity: Forgetting a Great Restaurant"

The Missed Chance

Two people drive past a vibrant restaurant. User A says, "That place looks great! Let’s come back later."

Problem: They notice the restaurant but don’t save it for later.

Forgotten on the Way Home

On their way back, they’re distracted and forget about the restaurant.

Problem: Without a reminder, the opportunity slips away.

Regret Days Later

User A tries to remember the restaurant but can’t recall its name or location.

Problem: Frustration sets in as they realize they can’t find it again.

Scenario 2:
"After-work Fatigue: End of Workday Decisions"

The Missed Chance

A person at their desk, staring at the clock, drained from the day.

Problem: Too tired to plan their evening, they want to relax but can’t decide what to do.

Getting in Car - Indecision

They sit in the car, staring at their phone, unsure of what to do next

Problem: They can’t decide on plans and feel stuck in indecision.

Missing Out on Relaxation

They drive past coffee shops and parks but keep going, too tired to stop.

Problem: They miss chances to unwind and feel frustrated for not acting.

User Journey

User Journey trying to outline the design opportunites in each stage of the user going through the experience

From creating the user journey, the insights from research are transformed into solid design festures that the product should include

User Needs

Users find it

Hard to Balance Professional Demands & Leisure

Users want

To Visit and Explore in The City

but don't have the time to schedule visits

Users hopes that the vehicle can

Better Their Time Management

Features

Allow drivers to

Explore

Locations of Interests

Schedule

Reminders

For Visiting Locations of Interests

Allow the vehicle to

Recommend

Possible tasjs drivers can do during commutes based on the context

User Flows

Prototyping

screenshots of multiple iterations before the final

06 DESIGN BREAKDOWN

The Final Design is:
An in-vehicle AI Agent capable of memorizing and recommending possible leisure detours the driver can make to do something else while on commuting.

The final design consists of three main sections:

Drivers can

Quickly ask for information about a place

and schedule future visits while driving, so the vehicle would be able to plan something for the future

The vehicle

Reminds the driver at an appropriate timing

with full awareness of the driver's current context

Ex: commuting home after work has more free-time v.s. commuting to work in the morning

The vechile will occasionally

Recommend the driver for interim stops

that allow the user to quickly finish tasks.

Ex: buying tomorrow’s breakfast, coffee, and other small tasks.

When creating the final design, I ensure any design decisions are judged by the following design strategies to ensure the optimal experience for the drivers.

Main Design Strategies

Multi-Modal Interaction

Use visuals to highlight the most important elements, and use the voice agent to provide further detail if necessary

Task-Oriented

Focus the design on Tasks the driver want to accomplish and not by specific location names

Context Aware

The design and interactions depends on the driver’s current location and timing

Multi-modal Interactions

Why?

Although assistive driving makes it easier for drivers to divert their attention while driving, it is still dangerous too display too much information through visual mediums.

Compare to smartphones, computers, and other digital tools, car interfaces shouldn’t be the main focus of the driver’s attention.

How?

The interface would visually present the most important information, while the voice agent NOMI provides additional information verbally to the driver

Notifications that quietly asks for the user choices, and adapts as the user becomes more familiar with the system

Task-Oriented Choices

Why?

Down to the most basic user needs, users want to be able to utilize their commuting times to do “something” rather than go to specific places.

How?

Users are first asked to decided what they hope to do, and from that choice, the system would recommend specific locations to satisfy the user’s need base on specific metrics

Context Aware

Why?

Sensors and navigation is on the biggest advantage of in-vehicle products, and is what provides competitive advantage compare to products based on mobile platforms. Thus, it is essential for the product to be highly context aware and adjust it’s choices and contents based on the context

How?

The choice/user flow differs depending on the current situation

While getting information from the locations outside

While recommending different tasks

07 REFLECTION

Familiarity with Vehicle-Specific User Behaviors and Scenarios:

Before this projects, I was mainly familiar with user scenarios related to mobile/PC devices. It was hard to extract myself from preconceptions built from past experience designing for mobile screens, and become familiar with vehicle interfaces

Formation of the Main Concept:

I went through a long phase of confusion with the project at the start, where it was hard to navigate through such a large project and understand my own subset. Eventually, I was able to extract the main idea of the project and was able to push forwards

Understanding Experience Definition:

During the design phase, one of the biggest challenges was to understand what make a design “good.” This led to me returning back to the drawing board multiple times, until I was finalized with the several design strategies that helps me identify which parts of the design is “good” for the project and while other details need more refinement.

Familiarize with Multi-Modal Interactions:

When designing the prototype itself in Figma, I had to become more familiar with the different types of interaction the vehicle affords, and how to represent these different types of interactions (touch/voice) in the prototype to better communicating my design.