Most startups don’t fail because the tech is bad. They fail because the founding team breaks.
And the break usually starts small. A missed promise. A “we’ll figure it out later” decision. A quiet worry about who owns what. Then one day, you’re not building anymore—you’re arguing.
A founder agreement is how you stop that story early.
Think of it like a seatbelt. You don’t wear it because you plan to crash. You wear it because crashes happen fast. And when they happen, you want the damage to be small.
If you are building robotics, AI, or any hard tech, this matters even more. Because your company is not just “an idea.” It’s code, models, designs, data flows, hardware drawings, and inventions. If that work is not clearly owned by the company, you can lose deals, lose investors, and lose patents.
This article is about what to put in writing early, in plain words, with a founder-first lens. Not legal theory. Not “template talk.” Real moves you can make this week.
And if you want help protecting what you’re building—before you raise your seed—Tran.vc invests up to $50,000 in in-kind patent and IP services for deep tech founders, so you can build a real moat without giving up control too soon. You can apply anytime here: https://www.tran.vc/apply-now-form/
The real job of a founder agreement
A founder agreement is not “paperwork.” It is a decision tool.
It forces you to answer the questions that will show up later, when stress is high:
Who owns the work?
Who decides what?
What happens if someone leaves?
How do we handle money?
What if we disagree?
If you don’t answer these now, you will answer them later. But later, you will answer them in pain.
A strong founder agreement does three things:
It protects the company.
It protects the team.
It protects the future value of your work.
That last part is the one technical founders often miss. In deep tech, value is not only in “traction.” Value is in defensible work—the stuff that can become patents, trade secrets, and real assets. If ownership is fuzzy, your IP is fragile. Investors see that fast. Acquirers see it even faster.
So let’s get tactical.
Start with the hard truth: roles are not titles
Most teams start with titles because it feels clean: CEO, CTO, COO.
But titles are not the real issue. The real issue is: who is accountable for what outcomes.
Two founders can both be “CTO” in their head. One writes core code. The other talks to early customers. Both think they are carrying the company. Then the company grows, pressure rises, and each feels unseen.
So, in writing, define “jobs,” not “status.”
Here’s how to do it without making it weird:
Write down the main work areas of the company. For a deep tech startup, that is usually:
Product (what you build)
Tech (how it works)
Data (what trains it)
Customers (who buys it)
Operations (how you run it)
Money (how you spend it)
Now match founders to areas where they have final responsibility.
Final responsibility means: if it fails, that person is on point. Not alone. But on point.
Then write down what “done” looks like.
Not big dreams. Simple measures.
For example, instead of “Lead product,” write: “Own product plan. Own release dates. Own feature choices after team input.”
Instead of “Lead sales,” write: “Own pricing, pipeline steps, and customer contracts. Own weekly update on deals.”
If you cannot write what “done” looks like, you do not have a role yet. You have a vibe.
One more thing: define time.
Is everyone full-time right now?
If not, write down exactly when each founder becomes full-time and what happens if they don’t.
This is where many teams lie to themselves. They say “Soon.” Soon becomes six months. Then resentment grows.
Put dates. Put conditions. Put consequences.
If you want Tran.vc’s help thinking through this in a founder-friendly way—especially for deep tech teams where IP and execution are tied together—you can apply here: https://www.tran.vc/apply-now-form/
Equity: the part everyone avoids, but everyone remembers
Equity talk is awkward because it feels personal.
But equity is not a reward for friendship. Equity is payment for risk and work.
If you don’t handle it early, you will handle it during a crisis.
The simplest rule is this: equity should match expected long-term contribution, not just who had the original idea.
Ideas matter. But execution is what turns an idea into a company.
Also, in robotics and AI, the “idea” is rarely one moment. It becomes real through thousands of decisions: architecture, training, iteration, testing, and the careful protection of inventions.
So how do you decide equity without turning it into a fight?
You do it in two steps.
Step one: agree on what the company needs to win.
Be honest. If you need a founder who can ship hardware, that matters. If you need a founder who can sell into factories, that matters. If you need a founder who can raise, recruit, and lead, that matters.
Step two: agree on who will carry those needs for the next 2–4 years.
Not this month. Not last year. The next 2–4 years.
Then you assign equity based on that.
Now the key part: use vesting.
Vesting is not punishment. Vesting is fairness.
It says: you earn equity by staying and building.
Most standard setups use a 4-year vest with a 1-year cliff. You don’t need to love the standard. But you should understand what you are trying to protect:
If someone leaves early, they should not keep a giant piece of the company.
Because the company still has to survive.
If your team does not use vesting, you are choosing a future where a founder can quit and still own a huge chunk forever. That can kill the company later.
Put vesting in writing early.
And be clear on what happens if someone is fired versus if they quit.
Be careful here. Not because you expect betrayal, but because confusion is expensive.
Also: decide whether any founder is putting in cash. If one founder pays the first bills, write it down. Is it a loan? Is it buying extra equity? Is it reimbursed later?
Don’t “handshake” money. Handshakes vanish.
Decision making: stop the silent wars
Many founders say, “We decide everything together.”
That sounds fair. It also breaks fast.
Because “together” means “we need agreement,” and agreement disappears when speed matters.
You need a clear way to decide:
Small decisions.
Big decisions.
Tie situations.
In writing, define a few things:
What decisions each founder can make alone in their area.
What decisions need both founders.
What decisions need a formal vote.
And what happens when you’re stuck.
A deadlock clause is not dramatic. It is realistic.
Deadlock happens when you’re 50/50 and disagree.
So decide your tie-break path now.
Some teams choose: CEO breaks ties.
Some choose: bring in a trusted advisor for a binding decision.
Some choose: a third board member later.
But pick something.
Because if you don’t, you are choosing the worst option: endless loops, stalled execution, and private frustration.
One practical way to keep this simple is to write “reserved matters.”
Reserved matters are the few decisions that always need both founders. Things like:
Issuing new shares
Taking on debt
Selling the company
Changing the mission in a big way
Hiring or firing a founder-level role
Signing a big contract above a certain dollar amount
You don’t need a giant list. You need a clear one.
And then, for everything else, let owners own.
That is how you move fast without becoming unfair.
Company IP: if this is unclear, you don’t own your company
This is the part deep tech founders cannot treat as optional.
If your work is not clearly assigned to the company, you can lose the right to patent it. You can lose the right to sell the company cleanly. You can even lose your ability to raise.
So, in writing, you want a clean statement:
All work created for the company is owned by the company.
All inventions related to the company’s business belong to the company.
All founders assign their rights to the company.
This is often done through an invention assignment agreement.
But even if you later use formal docs, your founder agreement should still make the intent crystal clear.
Also talk about “before company” work.
Many founders come in with old code, old models, or research.
You must decide:
Is that coming into the company?
If yes, on what terms?
If no, how do we prevent mixing it with company work?
This matters because mixing is easy.
A founder copies code from a past project. Or they reuse a dataset. Later, you can’t prove clean ownership. That is a red flag in due diligence.
If you’re building AI, be extra strict about datasets and training sources. If you don’t own the right to use the data, you may not own the output in the way you think. At the very least, it can create risk that slows deals.
Put a simple rule in writing:
What can be reused.
What cannot be reused.
How you document it.
If Tran.vc is supporting you, this is one of the biggest levers. Patent strategy is not only “file later.” It’s about building clean invention records now, so your filings are strong and your ownership is clear. You can apply here: https://www.tran.vc/apply-now-form/
Confidentiality: not just “don’t share,” but “how we handle info”
Founders often assume trust covers this.
Trust is good. Systems are better.
You need clarity on:
What is confidential.
How you store it.
Who can access it.
What happens when someone leaves.
This is even more important when you work with contractors, labs, or manufacturers.
In your founder agreement, define:
Company materials stay in company tools (company email, shared drive, repo access controls).
No founder keeps the only copy of anything important.
All key passwords are stored in a shared secure vault.
When someone leaves, access is removed same day.
This is not about fear. It’s about being a real company.
Time, effort, and “side projects”
Here is a common founder breakup line:
“I didn’t know you were doing that.”
So put it in writing.
If the company is meant to be full-time, say it.
If side projects are allowed, define the boundary.
For example:
No side project that competes.
No side project that uses company code, data, tools, or time.
Any outside consulting must be disclosed.
Any outside income that uses company relationships must be approved.
Again: you are not policing each other. You are keeping trust clean.
Because when the startup gets hard, founders look for reasons the other person is the problem. Don’t give the future angry version of you extra fuel.
How founders leave: the most important section you hope to never use
If you only get one thing right, get this right.
People leave.
They get sick.
They burn out.
They get a better offer.
They move.
They disagree.
A founder agreement that doesn’t cover leaving is not complete.
You want to define:
What counts as leaving (resigning, being removed, not meeting time commitments)
What happens to unvested equity
What happens to vested equity
What happens to IP and confidential info
What happens to titles and public statements
You also want to define “good leaver” and “bad leaver” in plain language.
Good leaver might mean: left due to health, family, or a company decision not related to wrongdoing.
Bad leaver might mean: fraud, theft, serious misconduct, or refusing to do agreed work.
The reason this matters: you may want different outcomes.
For example, you might want the company to have the right to buy back shares in certain cases.
Don’t try to invent complex legal rules yourself. But do write down the intent and the business logic.
Also, decide what happens if one founder wants to fire another.
This is uncomfortable. It is also real.
Write:
Who can propose removal.
What process happens.
What vote is needed.
What happens after.
This will not “cause” conflict. Conflict exists whether you write this or not.
Writing it simply makes conflict less deadly.
The “what if we sell?” question
Most founders dream of building forever, until a real offer shows up.
Then things get weird.
One founder wants to sell.
The other wants to keep going.
So decide early:
What vote is required to sell the company.
Whether drag-along rights exist (so a majority can require the rest to sell).
How proceeds are distributed.
This is one of those topics that feels far away until it isn’t.
A basic agreement here can prevent a deal from dying later.
Conflict rules that keep the team calm
How to disagree without burning trust
Most founder conflict is not about the topic. It is about the feeling behind the topic. One person feels unheard. The other feels blamed. Then even small choices start to feel like fights about respect.
Put a simple “how we disagree” rule in writing while you still like each other. This is not therapy talk. This is a speed and focus tool.
A clean rule is: when a disagreement shows up, you name it within 48 hours. You do not let it sit for weeks. Waiting makes stories grow in your head. Those stories become “truth” even when they are not.
Then you agree that each founder gets uninterrupted time to explain their view. Not as a debate, but as a full explanation. The goal is not to win. The goal is to understand what each person is protecting.
A clear path when you can’t agree
Some topics will not resolve quickly, even with good talk. When that happens, you need a tie-break method that does not ruin the relationship.
Write a short escalation path. Start with founder-to-founder talk. If that fails, bring in one neutral advisor that both founders trust. If that still fails, move the decision to a pre-chosen tie-breaker like the CEO, or a board vote once a board exists.
The key is that the method is chosen before the conflict. In the middle of conflict, any “new” process feels like a trick.
Meeting rhythm that prevents silent resentment
It is easy to skip meetings when you are busy. Then you only talk when there is a fire. That pattern creates a team that lives in stress.
Put a standing founder meeting on the calendar. Keep it short, but consistent. Use it to share what is going well, what is blocked, and what is worrying you.
Also agree that big decisions are not made in passing. If a decision changes priorities, budget, timeline, or ownership, it goes into the founder meeting. That makes decisions feel fair and visible.
What to write into the agreement
You are not trying to script your lives. You are trying to avoid chaos.
In writing, include a short section that says disagreements will follow a set path, and that both founders commit to using that path before taking drastic actions.
This is one of the simplest ways to protect the company from the “slow breakup” that kills momentum.
Founder promises that become problems later
The hidden danger of “we’ll figure it out”
Early on, founders say things like “I’ll handle fundraising,” “I’ll build the model,” or “I’ll start selling.” These are often said with good intent.
The problem is that “intent” is not a plan. When the work becomes hard, each person remembers the promise in a different way. That difference becomes friction.
A founder agreement should turn vague promises into clear commitments. Not because you want control, but because clarity keeps trust clean.
Defining output instead of effort
Many teams track effort, not output. They talk about how many hours they worked. That is not useful for a startup.
Write down outputs. Outputs are things you can point to. Shipped features, signed pilot terms, delivered prototypes, filed patents, completed test runs, or completed customer interviews.
This keeps the team focused on progress, and it reduces the “I worked harder” arguments that often show up when stress rises.
A simple way to document commitments
You do not need heavy tools. You need a simple system.
Agree that each founder writes their main goals for the next two weeks. Keep them few, and make them clear. At the next founder meeting, you review what happened and what did not.
If something did not happen, you do not attack. You ask what blocked it, and whether the plan needs to change.
This habit makes your founder agreement real. Without habits, the agreement is just a file in a folder.
What to write into the agreement
Include a section that says founders will set and review commitments on a set rhythm, and that commitments will be measured by outputs.
This helps you stay aligned while the company changes fast.
Equity fixes when roles change
Why “fair at the start” can become unfair later
Even if you split equity well at the start, life changes. One founder may take on more risk. Another founder may become less available. A new executive may come in and change the work balance.
If you never revisit the equity conversation, resentment can grow quietly. Then it explodes at the worst time, like when you are raising or signing a major deal.
You do not want equity to be a constant argument. But you also don’t want it to be frozen forever if reality changes.
The power of vesting and earned ownership
Vesting is one of the best tools for handling change without drama. It already connects equity to staying and building.
If one founder slows down or leaves, vesting reduces the damage. That protects the active founders and keeps the cap table clean for investors.
Also consider whether vesting starts at the same time for all founders, especially if someone joins later. If one founder starts six months later, you can still be fair, but it needs to be written clearly.
Handling major role shifts without a fight
Sometimes a founder’s job changes in a big way. For example, a technical founder becomes a manager and stops building. Or a business founder stops doing sales and focuses on operations.
These shifts are normal, but they affect contribution.
A simple approach is to put in writing that founder roles will be reviewed at set milestones. Not every month, but at moments that matter, like after a fundraising round, after a product launch, or after major hiring.
The review is not about blame. It is about aligning roles with what the company needs now.
What to write into the agreement
Include a clause that roles can change, and that founders will review roles at key milestones.
If you want to go further, you can include a process for adjusting equity or adding performance-based grants later. Many teams keep this as an option, not a default, because changing equity can get complex.