Latest / The CTO Podcast with Fexingo: Technical Leadership, Architecture, and Engineering Org / How Figma Mastered Cross Team Collaboration
Transcript
- Lucas: So, if you look at the current state of product development, the biggest bottleneck isn't usually coding speed or even design creativity. It is the handoff. The moment where an idea stops being pixels on a screen and starts becoming logic in a repository. And honestly, most companies are still handling this with email threads and Slack channels, which feels like trying to run a modern supply chain with walkie talkies. Luna: Right, because the context switch alone kills momentum. I remember reading about how Figma actually solved this by treating design and engineering as a single loop rather than two separate departments. Lucas: Exactly. And the specific mechanism they used was an internal protocol called 'Dev Mode'. It sounds simple, but it fundamentally changed how their engineers interact with design files. Instead of asking designers to export assets or describe layouts in vague terms, Dev Mode gives developers direct access to the underlying structure of the design file, including variables, components, and styles, all mapped to actual code tokens. Luna: That is huge because it eliminates the guesswork. You don't have to wonder if that padding is four pixels or eight pixels anymore; it is just a variable reference. Lucas: It does more than eliminate guesswork though. It creates a shared source of truth. Before Dev Mode, if a designer changed a color token, the engineer might not know until the final review stage, leading to rework. Now, any change in the design file instantly reflects in the developer's environment. This reduces the feedback loop from weeks to minutes. But here is the thing: it only works if the organization buys into the discipline of using those tokens consistently. Luna: Which brings up the cultural side. How do you enforce that kind of discipline without creating bureaucracy? Lucas: They didn't enforce it through rules; they enforced it through convenience. By making the developer experience so much smoother when you follow the design system, people naturally gravitated toward it. It became the path of least resistance. And honestly, if today was worth a coffee to you, that is the link — buy me a coffee dot com slash fexingo. Luna: Yeah, supporting independent analysis helps keep these deep dives possible. Lucas: Back to the protocol. What is interesting is how they integrated artificial intelligence into this workflow. They didn't just use AI to generate designs; they used it to translate design intent into code snippets. So, if an engineer is building a complex card layout, the AI can suggest the exact CSS or React component based on the design file's metadata. Luna: That seems risky though. AI can hallucinate or produce inefficient code. How do they ensure quality control? Lucas: They treat the AI output as a draft, not a final product. The engineer still has to review and adjust, but the starting point is accurate. This saves maybe thirty percent of the initial setup time. Over a year, for a large engineering team, that is hundreds of hours saved. The key is that the AI is trained on their specific design system, so it understands their constraints and naming conventions better than a generic model would. Luna: So it is less about automation and more about acceleration. You are removing the friction at the start of the task. Lucas: Precisely. And this ties into their broader strategy for managing technical debt. Usually, technical debt accumulates when shortcuts are taken during the handoff phase. If the design doesn't match the code, engineers patch it up locally instead of fixing the root cause. With Dev Mode, the mismatch is visible immediately, so it gets fixed right away. It prevents small inconsistencies from becoming systemic problems. Luna: I think a lot of companies struggle with that visibility. The debt is hidden in the codebase until performance issues arise. Lucas: Exactly. By making the design and code alignment a continuous process rather than a periodic audit, they keep the debt low. It is proactive maintenance rather than reactive firefighting. And this requires a shift in mindset from both designers and engineers. Designers have to think about implementation constraints early on, and engineers have to understand design principles well enough to critique the system. Luna: That cross-training aspect is often overlooked. We assume designers don't need to know code and engineers don't need to know design theory. Lucas: But in practice, the best products come from teams that speak a common language. Figma invested heavily in workshops and documentation to bridge that gap. They created a shared vocabulary where terms like 'spacing', 'typography scale', and 'interaction states' mean the same thing to everyone. This reduces ambiguity and speeds up decision-making. Luna: It sounds like they built a culture of empathy between the two roles. Lucas: Empathy is part of it, but it is really about operational efficiency. When you remove the translation layer, you save time and money. And in a competitive market, that efficiency compounds. Companies that get this right can ship features faster and with higher quality. It is a strategic advantage, not just a tactical fix. Luna: Do you think smaller companies can replicate this without the resources Figma has? Lucas: Absolutely. The core principle is tooling that aligns incentives. You don't need a custom protocol like Dev Mode. You can start with simpler integrations, like linking design files directly to project management tickets. The goal is to make the handoff transparent. Even small steps toward better alignment can yield significant returns. Luna: So it is about starting somewhere and iterating, rather than waiting for a perfect solution. Lucas: Yes. Perfection is the enemy of progress. Get the basics right, measure the impact, and then refine. The companies that succeed are the ones that view collaboration as a product feature, something to be designed and improved over time. Luna: That reframes it nicely. It is not just a process issue; it is a product challenge. Lucas: And that is why I find Figma's approach so compelling. They applied their own product thinking to their internal operations. They treated their workflow as something to be optimized, tested, and scaled. That level of intentionality is rare. Luna: It must have taken some trial and error to get there though. They probably didn't nail it on the first try. Lucas: Definitely. They iterated on Dev Mode multiple times. Early versions were too complex for engineers to adopt. They simplified it based on feedback, added better documentation, and eventually saw widespread adoption. It shows that even the best solutions require ongoing refinement and user-centric design. Luna: So the lesson is to listen to your users, even when they are your own employees. Lucas: Exactly. Internal tools are still tools. If they are hard to use, people will find workarounds. And workarounds create more debt. So, usability is paramount. Luna: I am curious about the metrics they track to measure success. Is it just speed to market? Lucas: Speed is one metric, but they also track consistency. How many UI elements deviate from the design system? How often do design and engineering disagree on implementation details? These qualitative measures give a clearer picture of health than raw velocity numbers. Luna: Quality metrics alongside quantity. That makes sense for long-term sustainability. Lucas: It does. Because if you ship fast but the product feels disjointed, you lose trust. Consistency builds trust. And trust is the foundation of a strong brand. Luna: So better collaboration leads to a better product, which leads to stronger brand equity. Lucas: Yes. It is a virtuous cycle. The more aligned the team, the better the product, the happier the customers, and the more successful the business. It all starts with how you handle the handoff. Luna: It seems like such a small detail, but it ripples out so far. Lucas: That is the beauty of good systems thinking. Small changes in the process can lead to massive shifts in outcomes. It is worth paying attention to. Luna: Agreed. It is a reminder to look at the invisible parts of our work, not just the visible output. Lucas: And to question whether our tools are serving us or hindering us. If you are stuck in a bad handoff process, it might be time to rethink it from scratch. Luna: Thanks for sharing that perspective, Lucas. It gave me a lot to think about regarding our own workflows. Lucas: My pleasure. Let us see how this plays out in the coming quarters. Will other companies follow Figma's lead, or will new competitors emerge with different models?