From startup exits to AI-powered legal innovation, Nick Holzherr has built his career around spotting everyday pain points and turning them into practical solutions. GitLaw is an AI-powered legal platform designed for startups and small businesses, offering free, lawyer-vetted contract templates, an AI agent for instant drafting and review, and Git-style version control for tracking changes and collaboration.
In this Eqvista interview, the founder shares how his frustrations with costly, time-consuming legal work inspired him to create a platform that makes contract drafting and review faster, simpler, and more accessible for founders. It’s a conversation about product-building, legal inefficiency, and how AI is reshaping the way early-stage businesses handle one of their most essential tasks.
Nick, you’ve gone from The BBC Apprentice finalist to building and exiting Whisk, then leading Samsung Food, and now founding GitLaw. How do you describe the through-line in your career so far?
I did The BBC Apprentice because I thought it would be fun – and it was! But generally, I’ve always had the urge to fix problems that I personally experience. With Whisk it was cooking for myself. With Gitlaw it was the time, cost and confusion with getting contracts written or negotiated when building businesses.
Cooking and contracts look totally different on the surface, but technically there are some important similarities.
Recipes and contracts are both long pieces of unstructured or semi-structured text and our goal is to help people take action faster. With Whisk, we indexed recipes and turned those messy paragraphs into structured ingredients, steps, timings to then offer simple workflows on top (shopping lists, substitutions, meal plans, etc). With contracts our workflows are free and super fast creation, analysis or editing of those contracts – and we’ll soon add more options like free eSign.
The team is also an important through-line. A lot of the same people are working with me – this is now our third business together.
What key event or metric made you decide that 2024 was the perfect year to launch an AI- native legal tech venture like GitLaw?
The models crossed the threshold where they could handle long context, nuanced drafting, and reasoning well enough to be productised – not as a demo, but as something founders would find it truly helpful.
I always wanted to serve founders – and pass on the benefits of AI to those founders. I noticed that most people in the space are helping lawyers be more effective – and I don’t believe that the full value of AI will be passed onto builders that way.
You’ve spoken about being shocked by early-stage legal bills in your first startup. Can you walk through one concrete example of those costs and how it directly shaped GitLaw’s product thesis?
I paid about $150k for the legal bill to sell Whisk.com to Samsung. Part of what I paid for was a law-firm partner who had negotiated hundreds of other deals like this before, negotiating the details with me. That was super valuable. But that might have been 20% of the bill.
The other 80% was spent on more junior colleagues building due diligence packs, writing agreements to clean up past mistakes and generating a bunch of very simple and standard agreements like shareholder consents. We had many meetings where we would go through large physical folders of contracts – each folder was something like 300-400 sheets large. It felt inefficient. Inefficiency can be mildly annoying in general, but when you’re paying $500-$1000 for that junior lawyer to wade through a process like that – it’s super frustrating.
A more regular frustration example is generating a legal agreement for a partnership with a customer. Sometimes you need something with a bespoke agreement. There was an example where we had agreed terms with a really important partner at Samsung Food but it then took about 3 months to go through the back and forth of negotiating the minor details. Each time we needed our lawyer to take another turn it would take one to two weeks. It cost about $8k for the contract – but it was the time that was more costly. During that time priorities can shift and cost the entire deal. In a previous world where things moved more slowly that could be just about acceptable. Today things move so far much faster and I think businesses would benefit from instant or at least much faster turnarounds.
GitLaw offers hundreds of free, lawyer-contributed templates and an AI agent to draft and review contracts. What was the original wedge: templates, the agent, or workflow — and how has that evolved?
We started with legal document templates. We had originally planned features that auto-build smart forms from those legal document templates to fill them out. But as we started to refine our form builder we saw how LLMs were becoming so much more capable – so we pivoted into replacing our smart forms with an AI Agent instead. We wasted some time building smart form builders that we then threw away but it was entirely worth it – the new experience is so much better.
GitLaw positions itself as “an AI-native legal platform built for founders, not law firms.” Practically, what are two or three UX or workflow decisions that reflect that positioning?
Ultimately our goal is to become the platform for everyone – including individual lawyers, in- house lawyers, micro law firms and very large ones too. Over the next year or two I believe most law firms are disincentivized to implement AI to its fullest potential. Ultimately using AI reduces the amount of work lawyers need to spend on each case, which when you’re charging by the hour, means less money. It’s a typical “innovator’s dilemma”.
Some product decisions that we’ve taken are:
1. We offer trusted and lawyer-vetted legal document templates that businesses/founders need. Many lawyers already have access to their own libraries (that’s how they do the work) so building this out is specifically targeted at users.
2. We’ve priced the platform to be entirely free and every user gets $5 free AI credit each month. If users run out of credits they can buy more for $20. That’s a stark contrast to many lawyer-focused platforms who charge $250-$1000+ per month per seat for access.
What’s been the hardest technical or product challenge in building an explainable, auditable AI system for legal documents, rather than a generic text generator?
While there is quite a lot of refinement and orchestration needed to get AI outputs to be high quality – the biggest challenge is to make it simple and easy to use. So many solutions out there are confusing to use. Building something beautifully simple is actually very difficult! It took us over 100 people in my previous company in Whisk.com to achieve that – and once we had done it we started winning “App of the day” by Apple – hundreds of times over. The difference between our first version and the version that won all the awards and high reviews is subtle – but hugely important.
What was the most controversial product decision you’ve made at GitLaw so far, and what did you learn from the pushback?
One of the most controversial truths is that access to law is so unfair. The stats show that 90% small businesses don’t go to a lawyer when they have a problem because it’s too expensive. It feels very unjust that big, rich enterprises can use law to bully small companies. Obviously some in the legal profession don’t see it that way. They say that it’s highly controversial to provide AI that does part of what a lawyer previously did for free. Many countries around the world have very specific laws to stop people doing it (laws that were put in place by lawyers). Because of those laws you have to be very careful about how you provide the service.
Personally I don’t think it’s controversial – I think it’s reasonable to let users decide themselves how they handle legal documents. Obviously it’s important to make the impact of their choices clear to them.
You can already generate legal documents in ChatGPT, Claude or Gemini – it just won’t be as good as the ones we generate (and sometimes it will be entirely wrong!). I also think that using AI is just a natural evolution of the downloadable documents and form builders that existed on so many websites like SimplyDocs in the past.
You recently raised a $3 million pre-seed round to scale GitLaw. What metrics or signals were investors most focused on at this stage: user growth, document volume, agent accuracy, or something else?
For the $3m pre-seed the focus was the team and vision – not any specific metrics. The team has a track record – we have build two successful startups together.
For founders who are currently burning cash on bespoke legal work, what is the pragmatic, low-risk way to start moving parts of their legal process into AI-assisted workflows today?
Personally I think most routine legal document generation or analysis/review can be done by AI first and reviewed by a human afterwards. Whether that human is a founder, HR/sales/marketing executive, in-house lawyer or outside counsel really depends on how important and sensitive the contract is.
Lastly, for founders bootstrapping their first SaaS startup in 2025, what’s your top piece of tactical advice on structuring equity grants and early contracts to avoid costly legal rework down the line?
My advice is to minimise the amount of money and effort you spend on anything that is not directly tied to finding product/market fit. For most businesses that includes minimising money and effort spent on legal work.
In the past that meant doing it yourself based on free online templates. I’ve done that many many times. The quality of what a platform like GitLaw can do is significantly better than the free online templates I’ve come across!
