AI Is Changing Legal Services, But Are Founders Asking the Right Question?
There’s a lot of excitement right now about what AI can do for legal services. Some of it is warranted. AI is genuinely good at drafting, reviewing, summarizing, and organizing — the kinds of tasks that used to eat hours of associate time and get billed to clients at premium rates. That’s a real efficiency gain, and firms that embrace it are delivering more value at lower cost. I’m one of them.
But here’s what I’ve noticed in conversations with startup founders lately: the efficiency conversation is crowding out a more important question. People are starting to ask “can AI do this?” when the better question is “what do I actually need from my lawyer?”
Those are very different questions. And how you answer them determines whether you end up with a faster, cheaper version of the wrong kind of legal help — or the right kind.
What AI Does Well in Legal Work (And Why It Matters for Founders)
Let’s give credit where it’s due.
AI tools can read a term sheet and surface troubling or non-standard provisions in minutes. They can cross-reference a cap table structure against what it finds as market norms, flag the clauses that tend to blow up a Series A, and generate a first draft of your SAFE or your NDA. They can do this at 11 p.m. on a Sunday, without a billing clock running.
This is genuinely useful. The first draft of a document, the initial pass on investor diligence, the research memo on a regulatory question — AI handles all of this faster than any associate, and increasingly at a quality level that’s hard to distinguish from competent human work if given the right prompts.
So yes: if what you need is document production, clause identification, or information retrieval, AI changes the value equation significantly. Smart lawyers use it for this as well.
Where AI Falls Short: The Limits of AI in Legal Advice
AI is very good at pattern recognition. It’s also somewhat dangerous for the same reason.
It finds the answer to your prompt that looks most like the right answer based on everything it’s been trained on. That’s powerful — and it’s exactly the wrong tool when your situation doesn’t fit the pattern. Even worse when your prompt doesn’t address the nuances of your situation.
Most significant founder decisions don’t fit clean patterns. The term sheet you’re about to sign isn’t just a document — it’s a bet on a relationship, a business model, a future. The question isn’t only whether the pro-rata rights are market-standard. It’s whether this particular investor will be a partner or a problem when things get hard. Whether the valuation sets you up for the next round or boxes you into a corner. Whether the deal structure actually fits where your company is going.
Your prompt will also significantly affect the response. “Draft an NDA” is a lot different than asking your lawyer to draft an NDA with a potential buyer who has significantly larger market power, may have history as a difficult partner and may be a competitor in some markets. A good lawyer will ask you about all of this and more. They may turn around and use AI to draft the first draft, but their prompt will incorporate all of this and more. It’s not AI failing, it’s the lawyer applying their experience and judgment.
AI doesn’t have judgment. Large Language Model reasoning is not the same. A good lawyer has years of earned legal judgment but also business judgment based on working with you and your company and pattern recognition across real relationships with multiple clients.
What Actually Protects Founders: 3 Things AI Can’t Replace
When I think about what I actually do for the startups and founders I work with — the founder and investor relationships I’ve built over more than 25 years — it comes down to three things. None of them show up in a chat with AI.
1. Asking the Strategic Questions Founders Don’t Think to Ask
The best legal conversations I have with founders are the ones where I say, “Before we get into the terms — where do you want this company to be in three years? Who’s your most likely acquirer? Are you building to exit or building to hold?” Half the time, that question changes the analysis entirely. The structure that looked smart on Monday looks different when you understand what the founder actually wants.
AI doesn’t ask questions on its own. It responds to questions. That’s a significant difference. The value of a good lawyer isn’t just answering correctly — it’s knowing which questions matter before you’ve thought to ask them.
An example: a founder came to us having already formed a company using AI generated docs. They soon discovered that they needed to be a Delaware c-corporation for a Series A investment and asked AI how to do it. AI gave them a technically correct answer, but it was an inefficient method. After talking to us about alternatives, he generated a whole new set of documents to do the conversion and asked us to review. That’s a lot of hustle, and I admire the commitment. But at a certain point, it makes sense to rely on experienced counsel — we know the forms, the process, and what needs to be done. Taking a DIY approach can sometimes end up being more costly in the long run.
2. Reading the Room: Why Human Judgment Still Matters in Deals
Fundraising and dealmaking involve people. People have unspoken priorities, deal-breakers they haven’t named yet, and signals they’re sending whether they mean to or not. I’ve sat across from investors and counterparties where the energy in the room told me something the term sheet didn’t — that this person was going to be a difficult board member, or that the real leverage in this negotiation was something we hadn’t touched yet.
You can’t train an algorithm on that. Empathy, intuition, and interpersonal intelligence are not soft skills — they’re strategic skills. And in a founder’s world, where so much depends on the quality of your relationships with investors, co-founders, and key partners, they’re often the difference between a deal that builds something and one that breaks it.
3. Giving Hard Advice: Why Honest Legal Counsel Matters
This might be the most important one.
Another example: a business owners was ready to move onto the next thing and she wanted to sell. She found a buyer who seemed eager and compatible. At the start, the terms were technically fine. The acquirer seemed reputable.
But the whole thing was a mistake. Not a legal mistake — a business mistake. The buyer couldn’t answer basic business questions about why they wanted certain things structure the way they wanted. The purchase price was too tied to seller financing. The buyer was too agreeable when we pushed back to see where they’d resist. Red flags all over.
I told my client to walk away. I told her the structure put her at substantial risk of not getting paid. I told her the buyer seemed wrong for her business and her needs. She pushed back. We talked through it. After a lot of renegotiation, she got better terms and the deal closed.
An AI tool, prompted to “review this LOI,” would have delivered a competent legal analysis. It would not have looked the seller in the eye and said “you are risking everything you built on a bad deal because you are tired.”
Clients have a right to do what they want, but it’s my job to help them understand the issues and achieve their goals.
AI Makes the Wrong Lawyer Cheaper. It Doesn’t Make Them Right.
Here’s the thing that worries me a little about the AI efficiency conversation: it can create the illusion that the quality of legal advice is rising across the board. In some ways it is — the floor has gone up. But the ceiling is still determined by the human behind the work.
A lawyer who was always going to tell you what you want to hear, generate documents on request, and avoid the hard conversation — that lawyer is now faster and cheaper. That’s not a win for you. It’s a cheaper version of advice you shouldn’t be taking.
The AI tools I use make me faster. They make the routine work better and more consistent. But the founders I work with don’t keep me around for the routine work. They keep me around because when the round isn’t coming together, when the co-founder relationship is fracturing, when the term sheet has a clause that looks fine but isn’t — they want someone who will think alongside them, tell them the truth, and help them find a path forward.
That’s not a service AI delivers. It’s what a trusted advisor delivers, using AI as one of their tools.
How to Choose the Right Startup Lawyer in the Age of AI
I’ll leave you with something practical.
When you’re evaluating a startup lawyer — especially for an ongoing outside general counsel relationship — pay attention to the questions they ask, not just the answers they give. A lawyer who’s going to protect your interests is curious about your business. They want to understand your cap table, your investor relationships, your exit horizon, and where you’re going — not just what you need today. Good lawyers want to know about you, this helps us understand why you are doing what you are doing. We provide better advice when we understand you.
Watch how they handle disagreement. Can they push back on your position without it becoming adversarial? Can they deliver a hard message in a way that keeps the relationship intact and the momentum moving? That’s a skill — and it’s a preview of what they’ll do when the stakes are real. My father was an executive at a Fortune 100 company and when I was going to law school, he told me “be a lawyer that finds a way to accomplish the business need. Find a way to say “yes” don’t just say “no.””
And notice whether they treat legal work as a transaction or a relationship. Transactional lawyers get the document done. Relationship lawyers help you build something — and make sure you don’t accidentally break it along the way.
AI is changing what legal services look like. The things worth paying for — judgment, empathy, trust, and the willingness to tell you what you need to hear — haven’t changed at all.
Frequently Asked Questions: AI vs. Lawyers for Startups and Business Owners
Can AI replace a startup lawyer?
No. AI can assist with drafting documents, reviewing contracts, and summarizing information, but it cannot provide strategic legal advice, business judgment, or relationship-based guidance. Founders still need a lawyer to navigate complex decisions, negotiations, and long-term risk.
When should I use AI for legal work?
AI is useful for early-stage tasks like drafting basic agreements, reviewing documents for common issues, and organizing information. However, once decisions involve negotiations, risk allocation, or long-term business strategy, it’s important to involve experienced legal counsel.
What are the risks of relying only on AI for legal advice?
AI can miss important context, misinterpret nuanced situations, or provide answers that seem correct but don’t fully apply to your business. Relying solely on AI can lead to inefficient structures, overlooked risks, or costly mistakes that require correction later.
How does a lawyer add value beyond what AI can do?
A lawyer provides judgment, asks the right questions, and helps you think through business decisions—not just legal ones. They can identify risks that aren’t obvious, guide negotiations, and give candid advice based on experience working with similar situations.
Is using AI for legal documents more cost-effective?
It can be in the short term, especially for simple drafting. But if the documents don’t fit your specific situation or require later correction, the total cost can be higher than involving a lawyer earlier in the process.
What should I look for in a startup or outside general counsel lawyer?
Look for a lawyer who takes the time to understand your business, asks thoughtful questions, and is willing to challenge your assumptions when necessary. Strong counsel should act as a long-term advisor, not just complete transactions.
What does an outside general counsel actually do for a business?
An outside general counsel provides ongoing legal guidance across contracts, transactions, risk management, and strategic decisions. They act as a trusted advisor, helping business owners make informed choices as the company grows.
