The Hidden Architecture of Trust: Navigating the Legal Tech Gap
In the early stages of building WorkRightsNow.org, I’ve been constantly reminded of the gap between "code-complete" and "system-ready."
I finally have the module "Find a Solicitor" feature ready. The form is live, the logic is sound. But the reality is that I haven’t onboarded any solicitors yet. This gap: between the technology being built and the ecosystem being populated—has turned into a deep exploration of the invisible architecture required for high-stakes legal tech.
Engineering via Subtraction
During a period over Christmas, I did some architecture refactory through its first major iteration.
Initially, the stack was "standard" for this era: a Next.js frontend, a FastAPI backend, and a RAG (Retrieval Augmented Generation) pipeline fueled by embeddings from 3 legal documents, Vercel cloud. But as I worked, I started questioning the complexity.
The models used are already trained on vast amounts of publicly accessible legal data; for an MVP, a heavy RAG setup felt duplicated. Similarly, since Next.js handles server-side logic elegantly, a separate Python-based FastAPI layer added more surface area than value.
From Data Broker to Data Controller
In software, there is a temptation to take the path of least resistance: collect data, pass it on, and call it a "lead". But to be a reliable partner in the legal space, I had to choose a different path before even making the first pitch to a firm.
A data broker is not enough, and it must be a Data Controller. This meant moving toward ICO registration and treating compliance as a core feature. When dealing with someone’s employment rights, potentially a defining moment in their life, integrity is the only currency that matters. If I want a solicitor to trust the data coming out of this engine, they need to know the foundation is legally and ethically grounded.
The Friction of Genuine Intent
In modern UX, we are taught to remove friction. But in legal tech, friction is a filter.
To ensure that once moving toward the first onboarding, the workflow isn't creating noise, I’ve introduced intentional "nudges" into the enquiry form. Beyond standard technical defenses like reCAPTCHA, asking users to upload a document: a contract, a letter, a piece of evidence.
In a "coding bubble," this looks like a conversion risk. In the real world, it’s a verification of intent. It ensures the requirement is genuine and allows the AI to offer value effectively before a human professional even enters the loop.
A Tiny Trial in a Transitional Period
This MVP is a milestone, but I view it as a "tiny trial." It’s a first step into exploring what I call an AI-blended reality.
I even thought about seeking an angel investor with legal reach, someone who believes in AI’s potential but also wants to hedge against the social risks if the technology isn’t managed responsibly, but at this moment, I just want to see how this lives organically for a year.
There might be a version 2.0 somewhere in the future, but I’m not in a rush. I’m doing this work because I feel excited to build. We are all living through a "strange alternative transitional period" with AI, and it’s easy to get lost in the pursuit of scale.
The Quiet Conclusion
I will try to find some solicitors who are happy to offer help through the platform, and learn how to do SEO, but ultimately, I’m at peace with the outcome. If it turns out that nobody uses it, or if it turns out that nobody actually needs it, I’m happy with that too. It would mean we’ve all found other ways to navigate these challenges.
We seek purpose in the work we love, but the goal is to solve the problem, not to stay in a state of "transition" forever. If WorkRightsNow.org offers even basic, effective information to one person in a defining moment, the experiment was worth it. If not, it was still a necessary step in understanding how the world runs outside my own bubble.
Comments
Post a Comment