The Cofounder Conversations AI Startup Teams Say Prevent Costly Breakups

What Three AI Startup Teams Put in Their Cofounder Prenup

Some startup founders are borrowing a concept from marriage: the prenup. Before writing code, three AI startup teams say they set explicit rules about equity, commitment, ownership and how to handle disagreements. Their framing is practical rather than pessimistic — the point is to make hard conversations happen while everyone is still excited, before money and pressure change the dynamic.

The context is a US startup market that keeps expanding. The Census Bureau counted more than 450,000 additional new business applications in the first half of 2026 compared with the same period last year. With AI enabling smaller teams to move faster, founders say the burden on a two-person partnership has grown, not shrunk.

David Emelianov, who sold the app Trimbox with cofounder Jordan Gaston for $4 million at the end of 2025, says three conversations should precede any code: what kind of company the pair is building, what level of commitment each founder expects, and what happens if someone wants out. Shalini Aggarwal and Andy Ratsirarson, the CEO and CTO of Tenfali after both left Amazon, push for a written cofounder prenup covering equity split, vesting, IP ownership, roles, decision-making and exits. Praneet Dutta and Joe Cheuk of the AI startup Pomo focus on the 30 minutes after a disagreement: who has final say in each domain, when a decision must be a live conversation, and how the agreement gets revisited as the team grows.

Why Founder Alignment Is the Real Risk in Tiny AI Teams

The advice is drawn from first-person accounts Business Insider edited for length and clarity; it is not independently verified and should be read as practitioner experience, not legal guidance.

Emelianov: playing the same game before it starts

Emelianov's key warning is that founders often skip the company-type conversation and only collide later. A venture-backed, winner-take-all startup and a profitable bootstrapped business are different games; if one founder is optimising for a $20 million acquisition and the other for a 30-year company, the partnership has a built-in fault line. He also says commitment needs to be defined rather than assumed: side projects, consulting, angel investing and even starting another company can be acceptable only if both founders explicitly agree on what committed means.

Tenfali: making the implicit explicit

Aggarwal and Ratsirarson frame the prenup as protecting the company and the relationship, not expecting failure. Their documented items — equity split, vesting, IP ownership, roles, major-decision rules and exit triggers — are designed to surface conflicts before emotions and money are involved. They also lay out a decision framework for inevitable moments of disagreement over slow customer conversion, product changes or tight cash: once a decision is made jointly, neither founder gets to play a blame game afterward.

Pomo: owning decisions at speed

Dutta and Cheuk describe the real risk as fake agreement, not conflict. Their system assigns domain ownership so a company does not stall waiting for consensus that will not arrive. Significant decisions require a live call, never a chat thread, because tone and context matter when stakes are high. They also treat their agreement as a living document that a six-person team will eventually need to update, and they allow either founder to ask to sleep on a big decision before the final call.

Conversations Founders Should Have Before the Partnership Is Tested

These are the practical conversations the founders say should happen before the partnership is tested.

• Decide the company type and commitment level before writing code. Emelianov's first question — venture-backed winner-take-all or profitable bootstrapped business — should be answered explicitly, along with whether side projects, consulting or angel investing are acceptable.
• Write down the hard items early. Tenfali's founders list equity split, vesting schedule, IP ownership, roles, how major decisions are made and what happens if someone leaves or an acquisition offer arrives.
• Assign final decision rights by domain. Pomo's rule is to know whose call is final in each area so the company does not wait for a consensus that will not come.
• Keep significant disagreements off chat. Pomo says major decisions should be handled in person or on a call to preserve tone and context.
• Build in a speed mechanism. Cheuk's clause: either founder can ask to sleep on a big decision, each person lays out their full view, and once a call is made both own it and keep moving.
• Revisit the agreement as the team scales. A framework that works for a six-person team may not work later, so the agreement should be treated as a living document.

The common thread is to make these conversations normal before silence gets expensive.