The habits I had to unlearn when I left law to build things

When lawyers ask me about moving from practice into building software, they usually want to know what transfers. The honest answer is that plenty does - and that the harder, more useful conversation is about what doesn’t. Legal training isn’t neutral. It installs instincts, deep ones, drilled in through years of training contracts and supervision and fee-earner incentives. Some of those instincts are genuine advantages in a product company. Others will quietly kill one. The problem is that from the inside they all feel the same: they all feel like professionalism.

These are the habits I had to consciously unlearn, and the ones I’m glad I kept.

Unlearn: shipping only when it’s finished

A lawyer’s work product goes out the door once, finished, with your name on it. If the advice is 90% right, it is wrong. Years of that regime build an instinct that anything leaving your hands must be complete, polished and defensible - and that instinct is close to fatal in a product company.

Software is not advice. It ships incomplete, on purpose, because the fastest way to find out what it should be is to put the unfinished version in front of real users and watch. The first release is not the final word on anything - it’s the opening question of a conversation. I knew this intellectually long before I could act on it. For months, “it’s not ready” was my most productive-sounding form of procrastination, and it took real effort to see it for what it was: perfectionism wearing a work ethic as a disguise. The unlearning isn’t lowering your standards. It’s relocating them - from “is this finished?” to “will this teach us something?”

Unlearn: treating risk as the whole answer

Legal training makes you superb at seeing what could go wrong. It does not train the other half of the judgement: weighing what going wrong actually costs against what moving slowly costs. In practice you rarely need that second half, because your role is to identify the risks and the client decides what to bear. As a founder, you’re the client now. Every decision is yours, and a founder who treats every identified risk as a reason to pause has outsourced their strategy to their anxiety.

The reframe that helped me: in an early company, the default outcome is death. Caution is not neutral. Every week spent de-risking a decision that could simply be tried is a week of runway spent buying certainty you didn’t need. Most decisions in a young company are reversible, and for reversible decisions, speed beats analysis almost every time. The skill of spotting risks stays valuable - the habit of stopping for all of them has to go.

Unlearn: winning the argument

Legal work is adversarial at its core. There’s a counterparty, there are positions, and being right - demonstrably, on the record - is the job. That trains a reflex to treat disagreement as an argument to be won, and it’s a reflex that misfires badly when pointed at users.

When a user says your product is confusing, they are not a counterparty making a claim you can rebut. There’s no version of “actually, if you read the tooltip” that ends with your product getting better. The user’s experience is data, and data can’t lose an argument - it can only be ignored. The same misfire happens inside a team: cross-examining a colleague’s half-formed idea is a great way to make sure people stop bringing you half-formed ideas, and half-formed ideas are the only kind that new products are made of. I had to learn to sit on the rebuttal reflex and ask another question instead. It still takes effort.

Unlearn: selling by the hour

The deepest habit of all, because it’s not a behaviour - it’s an identity. Practice teaches you that your value is your time and your effort: hours recorded, all-nighters survived, responsiveness as a virtue. Products invert this completely. A product’s value has nothing to do with how hard you worked on it, and a feature that took three months can matter less than a fix that took an hour. Effort is a cost, not the product.

This one matters commercially, not just psychologically. Lawyers who build companies often price like lawyers - anchored to effort and inputs - and underprice the thing whose entire point is that it removes effort. Value pricing feels almost indecent to someone trained on timesheets. Do it anyway.

Keep: everything about precision

Now for the other side of the ledger, because the transfer is real. The habits of precision - reading things properly, defining terms, noticing what a sentence actually says rather than what it gestures at - are rarer in the software world than lawyers would ever guess, and they compound. Contracts, obviously. But also product specs, investor terms, partnership agreements, the exact wording of a claim in your own marketing. I’ve watched founders sign things no lawyer would let a client sign, and I’ve watched ambiguity in a spec burn a month of engineering. The lawyerly allergy to vagueness is an edge. Keep it.

Keep: knowing the client

And the biggest advantage of all, the one that justifies the whole difficult migration: you know the user. If you’re building for lawyers, you have sat where they sit. You know what 6pm on a completion day feels like, what supervision actually involves, which corners get cut and which never do, and why the beautifully rational workflow in the pitch deck doesn’t survive contact with a Thursday. Outsiders building legal tech spend years and fortunes acquiring a fraction of that context, and most never do. It’s the moat you walk in with.

The trick - the whole trick, really - is holding both truths at once. Your legal training is simultaneously your best asset and your most reliable source of bad instincts, and nothing about how the instincts feel will tell you which is which. Rightness feels like rigour. Perfectionism feels like standards. Caution feels like wisdom. The founders who make the jump well aren’t the ones who abandon their legal formation, or the ones who trust it completely. They’re the ones who learned to audit it.