Having weak product-market fit before the AI wave was an advantage. Gamma had built foundational tech but wasn't tied to a legacy product or user base, allowing a rapid and complete pivot to AI without the friction that a more successful pre-AI company would have faced.
Rejecting the common 'niche down' advice, Gamma intentionally built a horizontal tool inspired by giants like Notion and Slack. Despite constant pushback from investors, they targeted a broad persona ('external presenters') which proved successful for wide user acquisition.
Obsessed with their innovative new format, Gamma's team ignored early user signals and delayed building a crucial PowerPoint export feature. This mistake stemmed from prioritizing their own vision and building for a future 'moat' before first winning adoption within existing user workflows.
Gamma's enterprise business didn't grow incrementally. It was jump-started by CEOs issuing top-down 'AI mandates,' forcing their companies to adopt AI tools. When these companies surveyed employees for use cases, presentations consistently ranked high, creating a sudden inbound demand that Gamma had to sprint to meet.
Gamma's initial AI product succeeded by adding 'guardrails' to constrain weak LLMs and ensure quality. As models improved, these same guardrails became limitations, holding back the AI's capabilities. The team's focus shifted from adding features to removing these constraints.
Gamma's product team operates on a key principle: use data (quantitative metrics and qualitative feedback) to precisely identify and understand user problems. However, they believe solutions don't come from data, but from creative design and engineering leaps that imagine a better way.
Gamma used AI to fix its low activation rate by solving the 'blank page' onboarding problem. They discovered this wasn't just an onboarding fix; it was the core value proposition customers wanted, solving the user's primary 'job to be done' and unlocking true product-market fit.
The rapid evolution of AI models forces Gamma into a new engineering paradigm. They build systems that must handle over 100 million users but are also designed to be completely replaced within 6-12 months. This means optimizing for usage scale, not time scale, with a culture of planned obsolescence.
The startup journey isn't a linear path to relief. The pre-PMF phase was characterized by creative freedom, followed by intense anxiety as runway dwindled. Achieving PMF didn't bring calm, but a different, equally intense stress of constantly trying to keep up with overwhelming demand.