When do you need a lawyer, a legal tech tool, or nothing at all?
When is chatting with Claude enough and when do you need a human lawyer to put their name on an answer?
Most founders end up asking the wrong question. The question isn't whether AI can be used for legal work, or whether a contract template is "safe." The question is what the specific situation actually calls for.
What do we need help with right now, and how much is at stake if we get it wrong?
Sometimes you don't need legal help at all. Other times a good template is more than enough. AI can be useful for understanding an issue or getting a first draft down. But some situations call for an advisor who can weigh the whole picture, give a clear recommendation, and stand behind it.
The cost usually shows up when the solution doesn't match the situation. A free template gets used on an agreement that should have been looked at properly. Or the business pays for lawyer hours on something a standard document could have handled just fine.
The goal isn't to use as much legal help as possible. It's to use the right help at the right time.
Start with the real question: what happens if this goes wrong?
Two questions usually give you a good starting point:
What are the consequences if this goes wrong? Is the situation standard, or is there something specific to your business that makes it different?
If the case is simple, familiar, and low risk, you often need very little help. If there's real money, ownership, employees, personal data, or a possible dispute on the line, that call should go to someone with legal expertise.
A simplified breakdown looks something like this:
No help needed: the situation is fully routine, with no meaningful legal or financial risk.
AI: you want to understand a topic, prepare questions, or put together a first draft.
A template or digital contract tool: the situation is standard, the document comes from a credible source, and you understand what's in it.
A permanent legal partner: legal questions come up regularly and carry real consequences for the business.
A lawyer or specialist: the matter is especially important, complex, contested, or could turn into litigation.
Can AI be used for legal work?
Yes. But be aware of what AI is genuinely good at, and what it shouldn't be used for.
Tools like ChatGPT and Claude can help you:
Get complicated contract terms explained in plain language
Summarize long documents
Put together a first draft
Work out which questions to raise before a meeting
Get up to speed on a topic you don't know well
That makes AI a useful sparring partner. But it shouldn't be the only basis for an important legal or commercial decision.
An answer can sound well put together and still be wrong. It might be missing important context, rest on the wrong legal rule, or miss something that matters specifically in your case. The business still carries the responsibility for that decision.
You also need to be careful about what information you put in. Customer names, employee details, salaries, health information, and other personal data shouldn't be shared in open AI tools unless the business has control over how that information is handled. Privacy rules still apply when AI is involved.
So the rule of thumb is simple: use AI to understand, explore, and prepare. Don't let AI alone decide what the business commits to.
We go into more depth on this in a practical guide to using AI in legal work and the privacy basics for growing companies.
When is a template good enough?
A template can be a genuinely good choice when the situation is standard, the consequences are easy to map out, and you understand what you're signing.
A simple mutual NDA, basic standard terms, or an ordinary privacy policy don't always need to be built from scratch. If your situation fits what the template was designed for, a well-built template can be both efficient and defensible.
The hard part isn't filling out the template. It's knowing whether it actually fits.
A template always rests on certain assumptions. It's built around what a typical agreement or situation looks like. If your business runs on a different model, has an unusual risk split, or has specific commercial needs, the standard language can miss the mark.
Be especially careful when:
The agreement involves significant value
The relationship is meant to last a long time
The other party has much more negotiating power
It's unclear who carries the risk if something goes wrong
The agreement touches on ownership, employees, or intellectual property
You don't fully understand some of the terms
If it's the first time you're using this type of agreement, a focused legal review can be a good investment. That doesn't mean redoing the whole document. Often it's enough to have someone confirm the agreement fits the situation and flag what needs to change.
When do digital contract tools make things worse?
Digital agreement tools usually work well for standard cases. The problem shows up when a polished, professional-looking document creates more confidence than the content actually deserves.
The agreement looks finished. Every field is filled in. The document gets signed.
It's only once the parties disagree that it turns out a key clause is missing, the risk sits in the wrong place, or the agreement never actually covered what the business needed it to.
A digital tool lowers risk when the alternative is no written agreement at all. It can raise risk when it's used in a complicated or unusual situation that neither the tool nor the user is really equipped to judge. At that point, the business is left holding a binding agreement and real obligations, without anyone having checked whether the document actually fits.
Which questions need a person who can be held accountable?
The single most useful word in this whole framework is accountable. A template can't be accountable. A tool can't be accountable. An AI tool can't be accountable. When the answer carries real consequences, you want someone who can own it.
Bring in a person when the question touches on any of the following:
Anything you're about to sign or commit to that involves real money, ownership, or obligations.
Employment decisions (terminations, layoffs, disputes). Norwegian rules here are strict and specific.
Ownership and cap table matters (shareholder agreements, vesting, investment terms).
Disputes, or anything that could turn into one.
Regulatory exposure, including serious privacy issues.
Anything new or specific to your company, where no template captures your actual situation.
If it's on that list, the question isn't which template or which tool. It's who.
Not everything that needs accountability needs a big law firm
It's a common assumption that "an accountable person" means a traditional law firm, with the hourly rates and formality that comes with it. So people avoid it, and reach for a template they shouldn't, or act on an AI answer they shouldn't, because the alternative feels like overkill.
But most of the real legal needs of a scaling company aren't high-stakes, one-off cases that call for a specialist firm. They're recurring questions an in-house lawyer would normally handle. "Look over this contract. Sanity-check this hire. Get our privacy basis in place. Tell us if this is actually a problem." These issues are too important for a template, but nowhere near big enough to justify a full law firm.
For a scaling company, the answer is often a built-in legal partner like Frank. Accountable like a lawyer, available and proactive like a tool, and priced for a growing company instead of running on an hourly model. We exist precisely because the old choice was "too little or too much," and neither one actually fit most growing companies.
How do you know which situation you're in?
Run whatever's in front of you through these questions, in order:
Does it involve real money, ownership, employment, personal data, or a possible dispute? If yes, you need an accountable person, skip to question 4. If no, keep going.
Is it fully standard, and do you understand every part of it? If yes, a template or DIY tool is probably fine. If you're not sure, treat that uncertainty as a no.
Do you just need to understand or prepare, rather than commit to anything? Then AI is a good sparring partner. Just don't treat the answer as final, and keep personal data out of it.
If you need a person: is it a one-off, high-risk, specialized, or contested situation? That's a traditional lawyer. Is it recurring, with real risk, but more in-house in nature? Then it's the middle option, a built-in partner, not an ad hoc firm.
The goal isn't "always get a lawyer" or "never get a lawyer." It's to stop paying, in money or in risk, for a mismatch between the help you chose and the situation you're actually in.
FAQ
When do I need a lawyer?
When the matter carries real consequences and needs someone who's accountable for the answer. Anything involving money, ownership, employment, disputes, regulatory exposure, or a situation specific to your company. For recurring, in-house-style needs, a built-in legal partner can fill that role without the hourly rates of a big firm.
Can I write a contract myself?
For genuinely standard, low-risk documents that you fully understand, yes, a good template can be enough. The risk is using a self-drafted or template-based contract on a situation that isn't standard, where a missing or poorly fitted clause can get expensive.
Are legal templates safe to use?
They're safe when your situation matches the template's assumptions and the risk is moderate. They become risky when used on non-standard or high-risk matters, because they create a false sense that the document has been "handled."
Is it legal to use AI tools like ChatGPT for legal work?
Yes, there's no law against it. But you're fully responsible for anything you act on, and you shouldn't put personal data into consumer AI tools, since GDPR applies as soon as personal data is involved.
Are digital contract tools enough on their own?
For standardized documents, often yes. For anything high-risk, new, or not fully understood, they can add risk rather than remove it, because the result looks finished even when no accountable person has actually reviewed it.
Written by
Meagan Leber
meagan@frank.legal
