Do You Need a Cofounder Agreement (and What Goes in It)
Short answer: yes, and the reason is not the one most founders assume. A cofounder agreement is rarely useful as a document you enforce. Its real value is that drafting it forces a set of conversations you would otherwise avoid for four years — what happens if one of us leaves, who decides when we disagree, what counts as full-time. Teams that have those conversations early rarely need the document. Teams that skip them end up having the same conversations under pressure with lawyers present. What follows is what actually matters, and what to raise with counsel; none of this is legal advice, and the specifics vary considerably by jurisdiction.
What’s inside
The agreement’s real job
The legal parts and the working parts
Eight clauses that actually matter
The working agreement nobody writes
Doing this retroactively
What an agreement cannot do
Where Blomma fits
The agreement’s real job
Most founders think of this as insurance: a document that protects them if things go wrong. That is one function and it is the smaller one.
The larger function is discovery. Every clause corresponds to a question the two of you have not answered, and the answers are frequently not what either of you assumed. What happens to equity if one of you leaves in year two? Is either of you allowed to take consulting work on the side? What does full-time mean — is it hours, or output, or availability? Who decides if you deadlock? What happens if one of you wants to sell and the other does not?
Ask a founding pair those five questions cold and you will typically get two or three genuine mismatches. Those mismatches are already present in your company; the document simply surfaces them while they are theoretical and cheap. The same mismatch discovered during an actual departure costs a relationship and possibly a lawyer.
Which means the value is front-loaded, and it is why “we trust each other, we don’t need one” gets the logic backwards. High trust is the ideal condition for this conversation, because nothing is at stake yet. Low trust is when it becomes a negotiation.
The legal parts and the working parts
Two distinct documents live under this heading, and conflating them is why so many founding teams end up with boilerplate that changes nothing.
Thelegal agreementis drafted by a lawyer, covers ownership, vesting, IP, and departure, and exists to be enforceable. You need this, it is not expensive relative to what it prevents, and you should not draft it yourself from a template you found.
Theworking agreementis a page the two of you write, covers how you actually operate day to day, and exists to be referenced rather than enforced. Nobody requires you to have one, almost no founding team does, and it prevents considerably more conflict than the legal document does — because the disputes that actually damage founding teams are about decision rights, communication, and load, none of which appear in a legal agreement.
Do both. The legal one once, properly, with counsel. The working one yourselves, and revisit it annually.
Eight clauses that actually matter
Of everything a lawyer may put in front of you, these are the ones worth spending real time on. Discuss each with counsel rather than deciding it from an article.
Vesting and a cliff.Founder equity that vests over time rather than being owned outright from day one. The single most important provision, and its absence is the root of most serious equity disputes. It protects the founder who stays, which is a point worth making explicitly because it is often framed as distrust.
What happens to unvested equity on departure.And whether departure type matters — resignation, removal, incapacity. Get the definitions tight; ambiguity here is what turns an exit into a dispute.
Acceleration provisions.What happens to vesting on an acquisition, or if a founder is removed without cause. Worth understanding before you need it.
IP assignment.Everything each of you creates for the company belongs to the company, including work done before incorporation. Diligence will surface any gap here, usually at the worst moment.
Roles, and what full-time means.Written concretely. “Full-time commitment” without a definition is a clause that cannot be used. Whether outside work, advising, or other ventures are permitted, and who approves them.
Decision-making and deadlock.Who decides what, and what happens when you cannot agree. Frequently omitted from legal agreements and frequently the thing you most needed.
Transfer restrictions.Whether either of you can sell shares to a third party, and any right of first refusal. Prevents the situation where your cofounder’s stake ends up somewhere unexpected.
Exit and buyout mechanics.How a departing founder’s stake is valued and settled. Agreeing a method in advance is far easier than negotiating one during a departure.
The working agreement nobody writes
This is the part that pays for itself, and it takes one session. Four items.
Decision rights.A list of the decisions that matter — pricing, senior hires, roadmap, spend above a threshold — and for each, who decides, who is consulted, and what escalates to both. This alone eliminates the largest single source of recurring cofounder conflict, because it converts questions of authority into questions of substance.
How you communicate.A standing weekly founder conversation that is explicitly not a status meeting. What gets written down. What one of you needs to hear before the rest of the company does. Unglamorous, and it prevents the information asymmetry that erodes most founding pairs.
What each of us needs.Genuinely useful and almost never discussed. One founder may need to be consulted before decisions in their area even when they do not hold them. One may need disagreement raised privately rather than in front of the team. Stating these is cheap and prevents a category of injury that is otherwise mystifying to the person causing it.
How we handle it when this stops working.How either of you raises a concern about the other, and what the mechanism is for a role change. Leaving this undefined is why cofounder performance problems go unaddressed for years — there is no door, so nobody opens one.
One page, both names on it, reviewed annually. It will feel slightly awkward to write and it is the cheapest conflict prevention available to a founding team.
Doing this retroactively
If you are four years in with nothing in place, the situation is normal and fixable, with one caveat: it is now a negotiation rather than a formality, because there is real value on the table.
A few things make it go better.
Frame it around a forward-looking event. Retroactive agreements land badly when they appear from nowhere and read as a move. They land fine when attached to something legitimate — a fundraise, a board formation, a lawyer’s diligence list. Investors will require most of this anyway, which is often the easiest framing.
Separate the legal from the working agreement, and do the working one first. It is lower stakes, produces immediate value, and establishes that this is about clarity rather than leverage.
Be careful about vesting introduced retroactively. Asking a cofounder to put already-owned equity onto a vesting schedule is a significant request and needs to be handled as a real negotiation with independent advice on both sides — not presented as housekeeping.
And take advice separately. Each of you should be able to talk to someone who is not the other person. Founders who share a lawyer for this discover the limitation exactly when it matters.
What an agreement cannot do
Worth being honest about the limits, because over-relying on the document is its own failure mode.
It cannot make someone want to be there. A founder who has disengaged is a management problem, not a contractual one.
It cannot resolve a vision gap. If one of you is building a venture-scale company and the other is not, no clause fixes that — it needs a choice.
It cannot substitute for the conversation. A signed document that neither of you has thought about is boilerplate, and boilerplate does not prevent conflict. The teams who benefit are the ones who argued about the clauses.
And it will not be your primary defence in a genuine dispute, because reaching for the document usually means the relationship has already gone. “You signed it” is legally sufficient and is not an answer to a claim of unfairness; founders who rely on it tend to get the outcome and lose everything else.
Where Blomma fits
The reason founding teams put this off is not laziness. It is that every clause is a question about failure — what if you leave, what if you underperform, what if we cannot agree — and raising those questions with someone you are building a company with feels like an accusation.
Blomma is an always-on AI career coach with no stake in your company. Use it to answer the eight questions honestly for yourself before you raise any of them, so you know what you actually want rather than negotiating in real time. Use it to work out how to open the conversation without it landing as distrust, which is the specific fear that stops most founders. Use it to draft your working agreement, including the item about what each of you needs, which is the one people find hardest to articulate. And if you are doing this retroactively, use it to think through the framing before you raise it. Where it warrants a human who has been through it, bring one in..
A cofounder agreement is worth having and it is not the point. The point is that drafting it makes you answer, while everything is still easy, the questions that otherwise get answered under pressure.