Meet the redesigned People + AI Guidebook
The second edition of PAIR’s AI design guidance adds action-oriented materials to help product teams put theory into practice.
The PAIR team just launched an updated version of the People + AI Guidebook. We sat down with the project leads, Rebecca Salois and Gabe Clapper to learn more about what’s new with the Guidebook.
Let’s start from the beginning: what is the People + AI Guidebook?
Rebecca: The People + AI Guidebook is a set of best practices and methods for designing products that use AI.
It’s something that grew out of our own work on PAIR designing AI products. A few years ago, we realized it would be really helpful to have a compendium of best practices for AI design that we could reference. When we couldn’t find anything out there that met our needs, we decided to make one and share it for people to use.
We published our first version in May of 2019. Since then, over a quarter of a million people in 200+ countries have used it. It’s been exciting to see the wide range of people who find it useful for their work.
Gabe: But the Guidebook is not a crash course on machine learning — our focus is on AI design guidance for product teams. We try to give enough context on key AI/ML topics so that people new to the space can use our guidance. For a deeper intro to the technology, we recommend checking out resources like Google’s Machine Learning Crash Course or this fun explanatory visualization on predictive modeling and wine.
And who made it?
Rebecca: Gosh, so many people contributed to this! We built the first edition with input from hundreds of people working in AI design inside and outside of Google, in industry and academia.
Over the past two years, we’ve continued to gather feedback and insights from an array of people who are using our guidance or designing with AI. They’re from all over the world and have a wide array of backgrounds — developers, designers, students, data scientists, and more.
Why this update?
Rebecca: We heard again and again that people want more materials to help them put these best practices into, well, practice — things like product examples, case studies and exercises.
At the same time, the field has continued to evolve since we first launched in 2019 and our own research has moved forward. We had new insights to share.
So, we decided to update the guidebook with more action-oriented materials to help put the guidance into practice, plus with new insights from our research.
Gabe: And don’t worry, the original Guidebook material is still there! We’ve even added new insights to some of the original chapters. All of our new additions are intended to add to and bring out the best from the original material, not replace it.
Tell us more about what’s in this update.
Gabe: We built this new edition for product creators who are in the day-to-day of building and shipping — UXers and PMs, primarily. We’re product creators too. We know how hard it is and how there’s never enough time to make the most optimal decisions, so we really wanted to take as much work as possible out of the process of learning and applying these best practices.
The Guidebook is now centered around 23 action-oriented design patterns. They’re meant to help you make decisions quickly as you build your product. We also added case studies of teams working on AI design challenges and a workshop kit.
What did you learn from the first edition?
Rebecca: Although we initially designed this as a resource for UXers and product managers, we’ve also found that folks in a ton of different roles use this guidance — developers, students, data scientists and policy makers across organizations big and small, in small businesses, in tech companies, in NGOs…the list goes on.
Gabe: With that in mind, we created a set of case studies based on real-world usage of the material and there are a ton of takeaways there to apply to a wide variety of products and organizations. People told us they were craving clear and actionable guidance — just tell me what to do! So we took that to heart and set out to make something that teams could easily consume and share.
What has changed in the way people design with AI since the last edition?
Rebecca: One major change is that, due to the COVID-19 pandemic, more people are working remotely. We’ve put a lot of thought into making sure these materials work for folks who aren’t physically together.
More people and more organizations are using AI, and they’re more comfortable working with it. We’ve seen increasingly sophisticated designs and practices across the industry, which is really exciting.
What about this workshop thing?
Gabe: The workshop kit is to help teams collaboratively work through the real world choices they need to make to apply our design guidance to their products. It’s something we’ve used with dozens of teams inside and outside of Google, and we wanted to share it more broadly.
It’s also designed to be run virtually. We were planning our first pilot of the workshop kit in March 2020, just when the COVID-19 pandemic hit. Like a lot of folks, we had to figure out on the fly how to do everything virtually. We learned pretty quickly that virtual workshops can be hard to set up and really draining to participate in. To make it as easy as possible to run these sessions — and avoid video conference burnout! — we avoided complicated workshop tools and just used a simple, familiar slide deck with a range of agenda options to fit your team’s bandwidth and focus.
The default is a 2-day workshop that covers the full set of activities, but it’s easy to pull apart and run whatever activities are right for you. All you need to do is download the material and follow the checklist in the facilitator guide to get going.
What do you hope people will do with this new version?
Rebecca: I hope people take this guidance and make it their own, adapt it and refine it for their needs. We love seeing examples of work that grew out of our guidance that took it in new or unexpected directions.
Gabe: Yes, we can’t wait to see how people use these materials in their own work!
