
Asana’s Kimi Chuang on what growth marketers learn by playing with AI

Kimi Chuang, who recently left Square as Head of Growth Marketing, had a to-do list for her first stretch of time off since middle school – and it included getting her tires rotated and reading a book in her backyard. As she told Knotch co-founder and CEO Anda Gansca in the second episode of her “Reserve the Right” podcast, even a break felt like something worth planning.
So it’s telling that one of Chuang’s biggest learnings about AI came from watching people try things with no particular outcome in mind. Chuang, now head of growth marketing at Asana, drew on her prior time at Square to talk about how marketers get comfortable with AI, what makes an internal tool worth using, and why measuring AI visibility gets more complicated once you ask what it means for revenue.
1. Getting good at AI starts with having room to play
Chuang remembers feeling intimidated when she first started building with AI. At Square, employees had access to different models and were given the freedom to try them. She noticed that some colleagues seemed to get good at it quickly. When she asked what they were doing, one told her he made a habit of trying to build something random. He got comfortable by putting in more reps.
“I say play purposefully,” Chuang said. The freedom to experiment meant she could try an idea without having to justify the time against a specific deliverable. It also gave her more chances to see where the tools let her down. With enough practice, she could start to spot patterns in which tasks went well and which ones needed more scrutiny.
Gansca recognized that from her own experience. “I actually feel like my greatest moments of truth with AI were very much the moments when it was wrong,” she said. The two talked about how those moments changed what they were willing to hand over to AI, and how often they revisit those decisions as the tools improve.
2. A useful AI tool has a clear stopping point
Chuang detailed how during her time at Square, her team needed to produce more creative assets, and they knew that some requests – but not all of them – could be handled through a self-serve tool. But instead of deciding where to draw the line upfront, they built the tool and tested it against a range of real use cases. Along the way, they got clearer about which requests it could handle and which still needed more time or help from the creative team.
Chuang recalls being “very honest about when it should stop.” That boundary gave people a way to decide when to keep working with the tool and when to bring in someone with the expertise the job called for. The team also assigned several people to keep trying new things with it so they could give feedback and improve it over time.
Gansca presses Chuang on how that worked in practice. Who actually built the tool? What did the marketing team have access to? Their exchange got into the everyday details of making AI part of how a team works, including what people learn once they start using a tool on their own assignments.
3. AI visibility is the start of a measurement question
As Square’s team began tracking how the brand appeared in AI answers, it faced a question many growth marketers are still working through. “Okay, we’re showing up now, what does that mean for bottom line revenue?” Chuang asked.
They looked at whether Square appeared in AI answers, how those answers described it, and where ChatGPT referral traffic was landing on the site. Each signal answered a different question, and a brand mention could tell the team that Square was part of the conversation. Looking at how the brand was described could tell them whether the answer was communicating the value they wanted customers to understand. And referral traffic gave them a way to see where some people went next.
Chuang got candid about how much the team could conclude from putting those signals together. The discussion became especially interesting when Gansca asked why Chuang wanted to connect the visibility data to other things Square knew about its audience. How much confidence do you need in a new metric before you start making decisions around it? And how do you tell when you’re optimizing for something simply because you can measure it?
Listen to the full episode of Reserve the Right on Apple or Spotify to hear Anda Gansca and Kimi Chuang talk through what they’ve built, where AI has surprised them, and which questions they’re still asking about AI discovery.
Published on November 11, 2026
Become a thought leader
Become a thought leader
Trusted by the largest (and now smartest) brands in the world.
“Before Knotch we did not understand what content was driving business results. Now we understand which content moves the needle. Knotch’s cohesive reporting and insights paint a real picture of what’s happening on our website instead of the patchwork quilt that comes from a Google Analytics approach. With Knotch we have been able to re-prioritize ad spend, route better leads to our SDR team, and inform our content development initiatives.”

"The Knotch platform ensures that we deliver high-performing content tailored to young home shoppers, enhancing their experience and driving better business outcomes.”

"Our partnership with Knotch has been highly successful, empowering us to leverage data-driven insights and refine our content strategy.”







.jpg)
