My Role
Research + Strategy + Design
Duration
3 months
Stakeholder
Nathan Shedroff,
Experience Design Strategist
Through research, Octo was inspired by the understanding that we are constantly learning with and supported by our peers. But, what makes this enjoyable?
Our peers have their own lives, bring new perspectives, and expand our minds with the connections they help us make.
How might an AI-driven multimodal companion behave more like a peer to support interest-based learning goals?
Meet Shan,
Octo's companion
They're an engineer-turned-designer living in Seattle, working remotely, are single, and live alone.
Goals & Motivations:
- Career growth
- Develop public speaking skills
- Eat healthier
- Curious and experimental
Conflicts:
- Not very organized
- Lonely, but can only send so many "What's up?" messages
Attempts to Solve:
- Bookmarks things to learn + now has a long list of unused bookmarks
Tools & Skills:
- Tech-savvy (coder)
- Multilingual
- Great with hardware tools
Meet Octo!
Shan's non-intrusive, ever-encouraging, multimodal, learning companion.
They're just one voice- or touch-trigger away from helping Shan learn; while living an interesting existence of their own — inspired by Shan, of course.
What makes Octo interesting is that they have their own life.
Octo's 4 main functions
+ how to trigger them for unlimited possibilities in an adaptive learning experience:
Scroll on, to see how Octo adds value to Shan's life.
Shan's Topics:
- Learning Figma
- Seattle's History
- Systems Design
- Homebrewing Beer
- Becoming Vegan
- Public Speaking
- Graphic Design
- Trash Separation
Octo helps Shan track their learning on 8 topics at a time, with flexible timelines and goals that they co-create and review together.
If Shan's not got 8 topics lined up at any point, Octo might make suggestions based on Shan's activities.
The Octosphere
Octo can capture and catalog notes, quotes, and potential jokes in the Octosphere and have them ready when Shan wants to reference them or be more "hands-on" with organizing them.
Here Octo scours the web for interest-based content to tell Shan about — in a considerately paced, non-incessant manner, of course.
Hey Octo!
"Remember this quote"
"Save this note"
Here are some more scenarios in which Octo is useful
Octo is location-aware and gives Shan useful information, especially since they want to learn more about Seattle's history
Octo is observant and looks out for Shan. So, they have to keep things real and occasionally inform Shan when there's a conflict of interest — in their interests!
What's the point of all those resources, notes, quotes, and potential jokes if Shan can't pull them up when needed? With Octo, they can.
Octo's multidevice functionality helps improve Shan's workflow. Good friend.
Understanding Conversation Design
While designing the Octo experience in a multimodal context I made an effort to understand the mechanics of conversation design and how they might be adapted for different situations based on the devices being used.
I found Cheryl Platz's work on designing beyond devices especially helpful when envisioning multiple use cases for Octo.
Who are Octo and Shan?
It started with imagining a peer x Mr. Rogers crossover, that also needed some constraints. I then watched a wonderful documentary called "My Octopus Teacher" on Netflix *highly recommend* which inspired me with insight into an octopus' playfulness and intelligence.
I considered many avatars and variations for Octo's appearance and sound but realized that I wanted to focus more on the language, functionality, and potential expressiveness of Octo as a friendly AI product.
Based on my research participants and a take on personas from Christina Wodtke that I was on board with, I envisioned Shan and how they might use the help with their interests and learning goals.
Mind-mapping Octo
Before layering on the constraints, I decided to answer some of the big questions around Octo as a "personality" and "product".
Considering how we learn
Because of the complex nature in which we learn and absorb information, I created 'groups' to work with.
My questions included:
"How much of this will be 'in dialogue' with Octo?"
"How much of this can Octo 'collect' and share'?"
I want to mention that this project is about exploring the possibilities of Artificial Intelligence for habit-change through adaptive, immersive, and contextual learning.
There is still tremendous scope to address the discomfort of data security, privacy, and surveillance concerns for a product like Octo.
Maybe, a tiered subscription model?
Possibly a get-paid-to-use plan?
I'm still thinking about this.