Our conversation with Winter Wei, Design Lead at Kira Systems.

Subscribe to this episode via iTunes, Google Play, YouTube, and Spotify, or stream it and connect with hundreds of AI startups and influencers from our chatbot. A full text transcript can also be found at the end of this post.

Finding a needle in a haystack


One of the simplest, most basic application of artificial intelligence that we see today is for task automation – to remove repetitive and monotonous activity.  
In the legal world, this comes in the form of document review and analysis, like Kira Systems. Instead of sifting through pages of documents to find potential risk, Kira automatically highlights and extracts the most important sections of a contract. Rather than trying to find a “needle in a haystack”, lawyers can more effectively and efficiently review documents.
While it seems like a no-brainer to use this tool to improve productivity, there is still the need for lawyers to ultimately understand the context and semantics of the documents being reviewed.
Especially when the potential risk associated to the task is high, building trust into a machine to do the work is imperative in the human-AI collaboration.

Interpretability vs. explainability

When designing for machine learning products and features like this, understanding mental models is crucial. Users need to feel like they understand how the machine came up with the result. 

For Winter Wei, Design Lead at Kira Systems, this means it’s her responsibility to help users interpret what they are seeing.

“When users say they want transparency, they’re not talking about the technicalities of the feature and how it works. What they actually want to know is if they can predict how this behaves.”

-Winter Wei, Design Lead, Kira Systems, speaking to Ask AI podcast

In this podcast episode, we speak to Winter about how she designs trust into the machine learning features at Kira Systems. She dives into the importance of designing in a way that users can interpret and explain the results they are seeing. 


The Ask AI interview with Winter Wei, Design Lead, Kira Systems


Listeners should also check out the Ask AI interview with Ben Alarie, the CEO of Blue J Legal, another Canadian company using artificial intelligence technology to disrupt the professional services industry.


Share Episode

If you enjoy our podcasts and free events, please take 2 seconds to share this episode with the buttons below. The Ask AI volunteers thank you!

What You’ll Hear

How she applied her background in science and statistics to now designing products.

The “needle in a haystack” challenge Kira Systems is trying to solve in the legal industry with artificial intelligence.

Why trust is so important in machine learning products and how to incorporate it into the user-design process.  

How designing for interpretability and explainability can solve the back box problem in AI.


Episode Links

Kira Systems Home

Kira Systems Twitter

Kira Systems Linkedin

Winter Wei Website

Subscription Links



Google Play

Google Podcasts


Episode Credits

Executive Producer: Chris McLellan

Senior Producer: Mike Letourneau

Host: Jaxson Khan

Assistant Producer: Francois Desmarais

Chatbot Sponsor: Ada

About Ask AI

Ask AI is an educational nonprofit with a mission to increase connections and inclusiveness within Canada’s artificial intelligence sector.

Our volunteer-run projects include hosted events, the Ask AI podcast, our chatbot, and the Ask AI newsletter.

To learn more, and get information on sponsorships and volunteering just ask our bot!