Blue Apron | Conversational AI for Culinary Experiences
Chef is a multi-platform conversational assistant for Blue Apron designed to provide a streamlined and personalized cooking experience.
The use of AI continues to grow as it offers efficient solutions to problems facing people and businesses. Our team sought to research a current brand that could benefit from this type of assistance. Our design involved the development of form, motion, and style of a digital ecosystem for a conversational assistant.
Blue Apron is an existing meal-kit service that delivers pre-portioned ingredients and recipe guidelines for subscribers to cook at home. We identified a need for a conversational assistant in the kitchen as it can facilitate personalized, hands-free interactions. While guiding the user throughout the cooking process, Chef makes the culinary experience more accessible and enjoyable.
Accessible on Blue Apron’s mobile app and a smart home display, Chef’s capabilities in and out of the kitchen include:
- Making personalized meal recommendations based on user preference.
- Connecting to smart kitchen devices for a streamlined cooking experience.
- Guiding individuals and multiple users through the cooking experience.
- Providing ingredient substitutions in real-time.
- Keeping track of multiple timers.
- Integrating with third-party applications like MyFitnessPal, Spotify, and YouTube.
In developing Chef, we followed an end-to-end design process that involved two cycles of research, ideation, prototyping, and reflection.
Blue Apron - Chef | User Journey
Blue Apron - Chef | Visual Design
Blue Apron - Chef | Motion Design
Blue Apron - Chef | Mobile UI
Blue Apron - Chef | Smart Display UI
Blue Apron - Chef | Personalized Dietary Options
Blue Apron - Chef | Smart Control
Blue Apron - Chef | Multiuser Cooking
Blue Apron - Chef | Food Recognition
Blue Apron - Chef | Tracking Food Intake
Our Point of View
We aimed to design for a situation that would benefit from a conversational assistant. Cooking is task-based, usually requires both hands, and oftentimes a piece of technology. Moreover, it is a communal activity. There is an opportunity for a virtual assistant to anticipate and aid with tasks and guide multiple users through the cooking experience.
Given the growth of technology in the home, we also explored ways our conversational assistant could scale across multiple platforms and integrate within the Internet of Things.
Additionally, we put a great amount of thought into designing user-centered conversations. We researched existing conversational and AI design guidelines to help inform our design and found ways to push these conventions forward, reapplying them to the new context of cooking. After rounds of iteration and discussion, we landed on a conversational anatomy that included short, simple, concise words that prompt users with task-oriented responses.
Finally, when defining the autonomy of Chef, we thought about human values and how they factored into what we were creating. We had informative discussions around user consent, control, autonomy, and communication.
Our process can be divided into two parts: first, designing our conversational assistant and second, designing the ecosystem around it. Here is a high-level overview of these two parts:
Research: We analyzed Blue Apron’s brand identity to help us develop an assistant rooted in the company’s visual language, tone, and target audience. Our team also mapped the customer journey to analyze the steps involved in the cooking process and identify intervention opportunities.
Ideation: We explored Chef’s visual from, choosing to develop an abstract visual form in-line with Blue Apron’s brand language. This form proved to be simple and flexible when morphing across all states of motion. We also began developing motion and expressional states, which are visual reactions and responses to calls of action from users. Our team also designed voice and conversation around task-oriented questions, prompted responses, and concise sentence structures to create a user-centered, AI-guided cooking experience.
Prototyping: During our prototyping phase, we developed the UI for multiple platforms (mobile and smart display), considering the technological and design affordances provided by each, and tested our AI conversations in realtime in the kitchen.
Reflection: We took a value-sensitive design approach when reflecting upon our process. In this age of AI, designers have an imperative to develop products that take into account human values of privacy, autonomy, and ownership. Throughout our iterative design process, we made an effort towards understanding stakeholders and their values to inform aspects of our virtual assistant’s ecosystem.
Chef affects several direct and indirect stakeholders.
Direct stakeholders are Blue Apron customers that engage directly with Chef. This group may include users who own an account and users who participate in cooking. Cooking participants may include partners, children, roommates, or friends.
There are several indirect stakeholders:
- Individuals living in or visiting a customer’s household: This may include a customer’s partner, child, family, or friends who do not directly interact with Chef.
- Blue Apron’s business: Internal teams within the company. Key teams include product development, design, engineering, marketing, finance, and customer support.
- Blue Apron’s meal kit supply chain: The people and infrastructures created around food curation, production, packaging, and distribution.
- Larger systems: food culture, the food system, the environment, and public policy.
Our project addresses paradigm-shifts in designing conversation-based user experiences. Chef serves as a guide, fostering collaboration and communication for multiple users in the kitchen.
Conversational-based technology is developing rapidly. To inform our process, we drew on a number of existing paradigms and patterns (e.g., Google Assistant, Siri, Cortana). While these patterns helped us build a foundation, we needed to reframe and push these paradigms into the new context of communal cooking. A few decisions we talked about were: How anthropomorphic should Chef’s voice be? And, how does data privacy impact the types of conversational interactions users will have?
Throughout our process, we evaluated our design based on human values. We challenged ourselves to design a service that considers different stakeholder values of privacy, autonomy, and ownership. Our team then cultivated a thoughtful approach that considered alternative design decisions that balanced stakeholder needs.
*Note - The Blue Apron logo and branded material is the intellectual property of Blue Apron.