The Keeper of the Flame: Figma's Mihika Kapoor on Vision, Conviction, and Building Zero to One Inside a Company

Host Lenny Rachitsky read out a note from Figma VP of product Sho Kuwamoto, who said Kapoor is really great at creating a vision and getting people to see what she sees. Kapoor's response was that vision is everything: the product development cycle is messy and chaotic, with extreme highs, extreme lows, and directions abandoned the moment users say they are wrong. Being anchored on a single vision turns those moments into learning opportunities rather than discouragement, so that any step along the way feels like forward progress, and it starts from one question — if you and your team do your job correctly, what does the world look like?

That question cannot be answered in a vacuum. Kapoor argues a PM has to be fundamentally inseparable from users and from their own team, and that a strong vision comes from cross-pollination: research insights that let you feel what a user is feeling, beautiful designs and prototypes that communicate what the future world looks like, and engineering grounding that keeps the target achievable. Visioning is not a linear relay from research to design to engineering, and she points to the book The Medici Effect, about how innovation emerges when people from different places converge, as her model for how the work should be organized.

Communicating that vision is a separate craft, and Figma is an especially demanding audience because it is a deeply detail-oriented culture that practices what it preaches about visual communication. Words will only get you so far, so Kapoor builds pitches as a repeating loop of pain point, solution, proof point rather than the traditional sequence of pain points, solutions, timeline and costing. Describing a product idea in words is simply not as compelling as seeing a user testimonial sitting on top of a prototype, which lets the audience feel the pain and the promise at the same time.

On the FigJam team, Kapoor focused on the meetings experience during the pandemic, when whiteboarding took off because people could no longer jam together in a room. The starting insight was that the most common meeting taking place in a FigJam file was a brainstorm. The gap was that Figma itself ran everything in FigJam — product reviews, Gantt charts, bug bashes — while customers outside the company were only using it for that one activity.

Digging into why, she found that brainstorms are an incredibly democratizing process: ideas can come from anywhere, the loudest or most senior person isn't doing all the talking, and quieter people or people who prefer to ideate alone first can still contribute. Most other meetings are one-way. So the vision became making the generative nature of a brainstorm the norm everywhere — team kickoffs where everyone leaves stickies and comments rather than a PM and designer handing mocks to an engineer, rituals like Kudos boards, templates that teach people to make any meeting more democratic, and features like music and voting that help people get into flow together.

Kapoor's rule for presenting this kind of vision is that the team should produce one artifact together, rather than a research readout followed by a design crit followed by a product review, which leaves each owner pouring energy into their own deliverable instead of a shared one. Decks are built in Figma, and the team leans heavily into designing and prototyping before a project gets a green light. Where most companies size the market and then decide to invest, Figma prefers a see-to-believe, see-to-feel emotional pull toward the thing being worth investing in.

For PMs without design or engineering skills, Kapoor offers two answers. The first is that AI is lowering the floor: she cites Cognition's Devin taking Twitter by storm, and David Huang of Replit describing Replit as a technical co-founder, with Figma aspiring to be a designer co-founder in the same spirit. The second is to simply ask people. She pitched her current project verbally for months before using Maker Week, Figma's company-wide hackathon where the whole company pauses for a week, to walk around the New York office asking everyone, "Will you work on this thing with me?" until someone said yes.

Figma's chief product officer, Yuhki, told the host that Kapoor gets extremely strong conviction extremely quickly, which lets her navigate the messy zero-to-one journey and rally the team. Asked how much of that is deep conviction versus an instinct she rides, she frames the job as creating momentum, and cites Jeff Bezos's one-way versus two-way doors: in a software company most doors are two-way, so it is important to have an opinion that people can anchor around and react to.

At Meta, she says, the PM role was distilled into product sense and execution. Product sense is a vague term that ultimately means good intuition, and intuition is built through insatiable curiosity and talking to users at every chance — she would grill friends at dinners about how they use Figma and FigJam, because a live conversation makes an anecdote stick. Enough conversations, A, B and C, build into a repository she can draw from when making a product decision.

Younger PMs often assume their opinion must be less valid than a tenured colleague's or a user's, so they start from zero and gather insights before saying anything. Kapoor argues the opposite: putting out an idea, even a totally wrong one, is a better catalyst for getting to a good solution, because people react to an idea much more readily than to nothing. If it's right they say so and it accelerates; if it's wrong they steer you somewhere more opinionated. She pairs that with Yuhki's advice to enter research with at least an A- idea, since learning something takes you to an A+ and at worst you are not starting from a B, and with the discipline of killing your darlings when external signal tells you otherwise.

Strong conviction has a cost, and Kapoor is blunt about it: her designer Kean asked who this girl was and why she had so many opinions, and for most of a year they had an hour-long one-on-one every single day. To keep conviction from steamrolling the room, she now labels it out loud — "medium confidence," "this is my hypothesis," "I do not have an opinion here, I defer to you" — and declares her direct communication style up front so people know they are expected to disagree with her. Directness only works if it's two-way, which is why she starts feedback conversations by asking for feedback on herself first, then acts on it, a practice she traces to Meta's "feedback as a gift."

Teammate Karl Jiang told the host that no PM has ever gotten him so hyped about a feature, and Yuhki described Kapoor as building hype to overcome people's doubts. Her framing is that a PM needs a pulse on how the rest of the company feels about the product, and that different products need different levels of excitement to make it out the door. When leadership has strong conviction, leadership does the amping; when the idea is a bottoms-up zero-to-one push, the onus falls even more heavily on the PM to keep propping the project up.

Coming back from winter break, she says, everyone has forgotten what happened last year, and people trickle into the office on different dates. Figma's SKO, the annual sales kickoff, is the first company-wide forum where the year's priorities are declared. Her product at that point was barely built, buggy every day, and used by roughly ten people outside the team, but she deeply insisted it get a demo in the keynote given by CTO Kris and CPO Yuhki. A demo nobody was expecting, she argues, does a huge amount to drive hype and help people see what she sees in the future.

Her claim is that you can't create hype for something you don't believe in; the only mechanism is getting people to see what you see, which is why she deliberately leverages large forums — Maker Week, SKO, and Config, Figma's annual showcase to the world. She is deliberately scrappy about pushing a product into visibility beyond what its current stage merits, because putting it in front of more people yields more signal on how it is trending. What could have looked like a distraction to the team instead gave them product insights and company-wide excitement about getting the thing shipped.

External hype is a separate muscle, and Figma has had a symbiotic relationship with design Twitter, where launches get celebrated alongside the company. For FigJam's one-year anniversary in April 2022, the team reasoned that anniversaries are champagne-and-retrospective moments, but FigJam has a personality — cheeky, fun, skeuomorphically your friend — so you throw it a birthday party instead. With marketing and engineering she ran a mini launch inside the product, Easter-egging each new feature under a birthday present and sending users on a hunt, because hype is tied to the feeling that something was built for me. Each product has its own brand of delight: for Figma it might be a niche capability that unlocks a years-old pain point, for Apple the unboxing.

Kapoor's approach to immersion depends on the product. At Meta everyone and their mother had an opinion, so any conversation doubled as research; at Figma the audience is more niche, though it broadened as the company went from one product to Figma, FigJam and Dev Mode. Her practice is to ask any tech-adjacent friend about Figma whenever she catches up with them, because the product's reach keeps expanding in ways that are easy to miss from inside.

The sharper move is asking non-users why they aren't using it, which she calls the most insightful conversations, since adoption is as much about marketing and perception as about product quality. She remembers asking former teammates, "Are you guys using FigJam?" and getting "maybe sometimes" back from people the product was literally built for. That conversation chain led FigJam to invest in making it easier for non-designers to get started, including a placeholder experience that previewed use cases on the canvas instead of offering traditional templates.

Not every product allows that: for infrastructure or security products, walking up to a stranger and asking why they don't use your tool makes no sense. Here the stage of the company matters. At an early-stage company the PM has to go find users through whatever channels exist, and founder-market fit lets you use yourself as a litmus test for what users need. At a larger company, the key relationship is with sales: being on calls lets customer pain points cross-pollinate into the roadmap and gives sales visibility into what's coming.

She operationalizes what she hears rather than letting it float in her head. On FigJam she set up a recurring cadence with the sales team to trade signal, walking them through how she uses FigJam in every kind of meeting; that turned into a Loom of her weekend FigJam that sales later distributed to prospects as proof of use cases beyond brainstorming. Feedback that isn't immediately actionable still gets captured, through a Slack-to-Asana emoji integration that turns any comment into a backlog task, followed by a weekly grooming session.

Karl's observation that her deep care makes him care deeply leads Kapoor back to a talk by Julie Zhu, Meta's first-ever intern and VP of design, during her rotational PM program. Zhu's point was that when two people disagree about product strategy, it is because they hold different assumptions. That is why Kapoor refuses to build a roadmap just because it was handed to her: for top-down strategy she digs for the assumptions that led people there, and for bottoms-up ideas she makes sure everyone else shares the assumptions in her head.

For people who don't love what they're working on, her advice is to stop treating whatever is in flight as your scope. Her take is that your scope is the world: step back, understand the company's vision and its users, and work out whether the idea you're passionate about falls inside or outside the company, which should guide your next career step. She also rejects the idea that founding is only for capital-F founders: founding inside a company trades risk for the ability to exploit a distribution or platform advantage, at the cost of never being able to make an executive decision and having to earn buy-in for everything.

Her culture examples are deliberately bottom-up. Hot seat began at the first post-pandemic PM offsite in March 2022, when the team was under fifteen people around a dinner table: each person gets two minutes on the clock and everyone else can ask them anything, with the option to decline. That morning the team had run the Strengths Finder test, which gave the game a foundation for digging into what about people's backgrounds made them think the way they do. It has since spread across the company — PMs play it with their teams, Yuhki and Sho played it with the exec team, and new hires and significant others get put in the hot seat.

The Figgies came from envy of the sales and marketing org's awards night at SKO. With another PM, Elan, she decked out Figma's boardroom, Bigma, with a red carpet and gold curtains, ordered trophies on Amazon with people's names on them, and voted absurd categories like most likely to name their child Figma or to go an entire career without writing a PRD, with everyone forced into an acceptance speech. She stresses it was low cost and assembled in an hour. Culture matters, she argues, because it establishes trust: passion comes from the vision and from the people you work with, and while roadmaps and products change, connection makes a team durable when times get tough — which is why a company value like Figma's "play" gives cover for these rituals.

Before her product career, Kapoor founded Design Nation, and the story is really about pivoting. In college, design was having a watershed moment in tech, with Airbnb and Pinterest differentiating through the interface layer, yet curricula hadn't caught up — Princeton had nothing resembling product design. When she interned at Facebook she found her class of twenty-five interns were all self-taught product designers, and a documentary about design rewriting the future convinced her to build a conference that would democratize access to design education.

Her original plan was to run it inside an existing Princeton organization that had funding, resources and expertise, but that group was skeptical of the business value of design, so she went from having everything taken care of to having none of it. A piece of college advice stuck with her — don't underestimate the power of a .edu email address — and she cold-emailed executives and designers until people replied that this had been a problem when they were in college and of course they'd help. The conference ran for years, went from national to international, and now draws speakers like Stuart Weitzman and Airbnb co-founder Joe Gebbia, aimed at design-driven students rather than only capital-D designers. The host's label for this pattern is high agency: see a problem, decide you will be the one to solve it, and reallocate your own time as the constraint changes.

Asked what she is not good at, Kapoor names the shadow side of her strengths. High conviction can read as pushing something because she believes it rather than because users need it, especially with people she hasn't worked with long, so she leans on user proof points. Scrappiness means editing the product review deck minutes before presenting, or starting the night before and staying up until 3 a.m., which teammates don't always love. She also gets consumed by details when some decisions should simply be deferred. The host connects this to the idea that every superpower has a shadow, and Kapoor adds the team-level version: be self-aware about blind spots and hire for them, because you want spiky individuals and a well-rounded team.

Kapoor's case for zero-to-one work inside big companies is competitive: if a company isn't constantly thinking about what's next and seeing around the corner from competitors, it will get taken over, so it needs to stay entrepreneurial. Figma has a "run with it" value that treats someone sprinting off in a new direction as a manifestation of company values rather than a distraction. Some of its most monumental launches came from that: the recently shipped Multi-edit, Jambot (an AI plugin inside FigJam from an AI hackathon), and the entire widgets platform, which started as a hackathon project.

Her metaphor for the zero-to-one PM is the keeper of the flame, since flames are destined to die unless someone tends them. In Greek mythology the gods sit on Mount Olympus while Hestia, keeper of the hearth, stays behind to keep the fire burning; the person or team pushing a new idea has to stoke the embers when they risk going out and make the idea spread like wildfire. Concretely that means three things: having the right idea, which comes from empathy and daily user conversations, securing buy-in through the vision, and making it spread so that a new joiner asks what that flame is and how to get involved. The right idea also has to ladder into a company goal — at Figma, the move from designing for designers to covering the whole product development cycle and non-designers — and requires understanding a user's full end-to-end tooling, not just your own product.

On buy-in, she says the key is optimism bordering on delusion, a reality distortion field that doesn't hear "no" or at least translates it into "not yet." Her first pitches of her current product landed as "maybe," including at a PM offsite on next year's strategy; the third attempt, at Maker Week, stuck. She was hosting that hackathon while pitching a brand-new product to be built in a week, which is exactly why scrappiness matters: you choose where to take the hit on quality or believability. To make the product feel real, the team made a two-line change swapping the FigJam icon in the file browser to the new product's icon — adding a whole new entry point was impossible, but changing something small communicated the difference. At demo day she introduced each demo and did a wardrobe change in between, and because the pitch happened in a company-wide forum, her colleagues ended up evangelizing the idea with her.

Spreading the idea is where Figma's multi-month staging and dogfooding process does the work; Kapoor learned it early when she assumed a week on staging meant launch, and was told no. Being vulnerable about the product and treating feedback as a gift matures it, and putting things on staging early also makes people invested: they give feedback, see it implemented, and feel they shaped that part of the product. Her closing advice is that if you have an insight other people aren't seeing, it is even more on you to shout it from the rooftops, because that in itself creates a more entrepreneurial culture. Finally, running a team means understanding motivations individually — some engineers want a solution handed to them, others can't feel passionate about something they didn't contribute to — so she asks directly, "How much do you like being involved in product decisions?" and tries to cater to the individual rather than the average.