Working at a startup like Typeface, where AI and design intersect, has made me really think about what separates a good product from a lovable one.
In our recent team discussion with the CPO, we talked about how Typeface is being mentioned in the same room as Google Cloud, Adobe, and Anthropic — with C-suite meetings that include companies like Coca-Cola. That kind of validation is rare for a company our size. But it also comes with a massive responsibility: to make a product that doesn't just work, but that people want to use.
The Real Opportunity
AI is everywhere right now — every company wants to integrate it, automate workflows, and accelerate creativity. But despite the excitement, almost everyone faces the same problem: change management.
Executives love the idea of AI. They see the potential, the efficiency, the ROI. But the people actually using these tools day to day? They're often skeptical, frustrated, or overwhelmed. Adoption doesn't fail because the technology isn't powerful enough — it fails because the experience isn't intuitive, and the tools don't feel like they were built for the user.
It's a familiar pattern: leadership gets excited about a vision, but the end users on the ground struggle to extract real value from it. They hit bugs, slowdowns, confusing flows — and eventually lose trust. At Typeface, we've seen firsthand how little frustrations — the "thousand paper cuts" — can add up and prevent real adoption.
Our CPO shared the story of Microsoft Office's ASHA score — a metric that measured the percentage of sessions with no veto moments, or no frustrating interruptions. Microsoft started with 0% "happy sessions" and improved over time simply by focusing on eliminating small annoyances. That level of empathy — seeing every bug or delay as a piece of user frustration — is what separates a usable product from a delightful one.
Lessons from Airbnb
One of my favorite examples of this kind of thinking is Airbnb's early story. Back in 2009, they were close to shutting down. Revenue had flatlined, and the founders were maxing out credit cards.
During a session with Paul Graham, they realized something simple but game-changing. All their New York listings had terrible photos. Paul told them to fly to New York, rent a camera, and take better pictures themselves.
It wasn't scalable. It wasn't even technical. But it worked — revenue doubled almost instantly.
That decision, to do things that don't scale, became the turning point for Airbnb. It taught them that understanding users sometimes means stepping away from your keyboard and into their world. As co-founder Joe Gebbia later said, "You have to become the patient."
In design school, Gebbia's team would actually test medical devices on themselves, in order to feel exactly what it was like to be the patient before designing solutions. That mindset shaped Airbnb's entire culture: every new hire takes a trip as a user in their first week to experience the product firsthand.
Bringing It Back to Typeface
At Typeface, I've realized the same principle applies. As engineers, we're often deep in the weeds — writing code, fixing bugs, and optimizing performance. But real progress comes from stepping back and asking:
Does this experience actually feel good for the user?
When you think in those terms, the priorities shift. Fixing a "minor" usability bug can suddenly feel as important as shipping a new feature, because you understand the friction it removes. Improving latency by 300ms might not make a marketing headline, but it makes the product feel faster, smoother, and more human.
There's a saying I keep coming back to: go slow to move fast. Taking time to deeply understand users — to think holistically, not just tactically — often prevents rework later and accelerates long-term growth. The north star shouldn't just be "more features," but rather "a better experience."
The Path Forward
Typeface's product has come a long way, but the next step is about refinement, or the craft. It's about fixing the thousand paper cuts, from Figma/PSD import flows to image generation quality, and polishing every micro-interaction until it feels effortless.
The companies that win in this next era of AI creativity won't just have the most powerful models or largest datasets. They'll be the ones that make people feel something when they use the product.
That's the kind of product I want to help build.