Founder Fridays: How second-time founders build with Frank Greeff, co-founder of Kinso

Kinso’s core bet is that business communication shouldn’t be treated as separate silos—email, Slack, WhatsApp, LinkedIn, Instagram—because the same people and the same relationships drive decisions across all of them. The product aims to ingest those channels into a single app, then use AI to infer what matters most to a user at any moment and stack-rank messages accordingly. The technical centerpiece is not just “better search,” but building AI context that connects related threads across tools, so an AI draft can use the richer, cross-channel history rather than a single-message context window.

That approach is tied to how Frank Greeff and his brother decided to build Kinso after selling Realbase for 180 million. Post-exit reflection reframed “freedom” as the ability to choose which battles to fight, not to stop building. Over a two- or three-month exploration phase, they sampled many business ideas and watched their own enthusiasm rise or fall as they uncovered more data. Kinso won because deeper investigation made the problem feel more exciting—not less—an internal signal they treat as a prerequisite for committing to a long, high-effort journey.

Conviction also came from early traction: a waitlist that reached about 13.5 thousand people, with growth approaching 100 new signups per day. The momentum served as proof that the communication pain point was real and that the market was willing to line up before the full product existed.

Realbase shaped Kinso’s operating philosophy more than its product mechanics. Realbase scaled across Australia and New Zealand to a 47% market share, but it also demonstrated what gets lost as organizations grow: individual accountability and “human” visibility fade. Kinso’s counter-design is a smaller, high-performing team—17 people, all with ESOP/shares—so everyone has a stake and can “win together.” The goal is to recreate the intensity of a small team while still building toward scale.

Hiring and team-building follow the same logic. Greeff criticizes the common practice of judging candidates only against a tiny pool; without enough volume, “best” is just “least worst.” Kinso’s early hiring required about 120 interviews to place eight engineers, creating a clearer measuring stick for what “excellent” looks like. The company also tries to hire for “unfair advantages” rather than generic credentials: ground-level ownership, GitHub autonomy/agency, and fewer layers of middle management.

On the AI side, Kinso leans on agentic workflows and heavy documentation to improve output quality. Engineers use AI coding in parallel—one engineer reportedly runs nine instances of Claude Code simultaneously—while Notion functions as the “brain” of decision-making. Principles for PRDs and other artifacts are documented so AI agents can transform customer feedback into product requirements, then into execution plans for code generation.

Even the founder advice is operational: don’t equate failure with shutting down an ABN; track the individual and whether someone stays in the game. In Greeff’s view, the real determinant is persistence, not a single binary outcome.