TL;DR
This personal reflection explores why an MBA student chose to spend a summer at General Legal, an AI-native law firm, to test what human judgment, taste, and emotional intelligence are worth in a profession increasingly shaped by AI. The experience suggests that the future of legal practice may not be about replacing lawyers, but about building AI-native workflows that codify judgment while keeping humans responsible for the final decision.
Key Takeaways
- AI is changing how professional-services firms structure workflows, not just how individual tasks are completed.
- Human judgment, context, and relationship management remain important in legal work.
- AI-native legal workflows can combine agentic systems with human attorneys making the final call.
- Legal AI is increasingly moving beyond experimentation, with firms and AI companies developing more integrated tools and workflows.
- Harvard Business School has integrated AI directly into its MBA curriculum, including a required Data Science and AI for Leaders course for first-year students.
A personal quest
Last year I started at Harvard Business School during a strange time. More than once, we MBA students felt the ground shifting beneath us. The pace of AI was too fast to ignore; even the school rebuilt its curriculum around it. And a quiet question hung over everything: “Is business school still relevant? Shouldn't we all just drop out and build?”
I won't make the case for the value of the MBA degree today, but somewhere in that first year, I developed an itch I couldn't shake: to get my hands dirty at an AI-native firm. Not an AI startup, but a neofirm: a traditional professional-services business that has built an agentic workflow from the ground up, while keeping humans as the ultimate decision-makers. After hundreds of case studies (we couldn't get through a single class without hearing the word "AI"), I wanted to test a few personal hypotheses. What is the real value of human judgment, taste, and emotional intelligence in the age of AI? And what does that value look like today in high-touch, relationship-driven industries?
So I picked law - a profession full of, well, humanity (A rewatch of “12 Angry Men” in a leadership class had partially influenced that decision.) After a lot of digging into the companies I'd want to spend the summer with, I landed a Product internship at General Legal: a new entrant in the AI-native space, but no stranger to the elite tier of the profession.
At the time, the choice felt a little contrarian. Isn't the point of AI to automate repetitive work so humans can focus on what matters? And wouldn't that mean joining a legal-tech startup promising lawyers exactly that?
Why General Legal
I don't have formal legal training. But years as a VC investor, working alongside lawyers to get deals across the finish line, taught me something: a good lawyer is less a genius mathematician than a masterful poker player. There are rules, boundaries, and precedents. But alongside them sits a vast world of human context – the push and pull, the emotion, the unsaid. Navigating that is what separates good from great.
Here's the problem I see with much of legal tech today: it's AI bolted onto legacy systems. That co-pilot-as-afterthought approach compromises the premise of a truly agentic workflow, and it leaves the AI trained on generic and average consensus rather than hard-won institutional wisdom. Good lawyers know when to push and when to concede. They read business priorities and a counterparty's tolerance. They juggle tradeoffs under incomplete information.
An AI-native workflow, built from the ground up, codifies those judgment-dense decisions into playbooks from day one, so the agent takes a stronger first pass and lightens the load on the attorney making the final call. The result can be a turnaround measured in hours rather than days, at a fraction of the traditional cost.
But our engineering prowess isn't the whole story (our Casetext exit certainly doesn't hurt). The real edge comes down to an age-old adage: it's all about the people. Our carefully vetted attorneys are the secret sauce. They cut their teeth at Big Law and chose to build the future of legal practice instead. They've lived the judgment-dense moments and now they're teaching that instinct to our agent, one that compounds with every matter and across every industry. Most importantly, they’re willing to roll up their sleeves to build.
The future of legal practice
There's never been a more exciting time to ask these questions. In recent weeks alone, headlines have tracked Big Law firms pouring serious money into developing their own legal agents, or renting the infrastructure off the shelf, to widen client access to their institutional knowledge.
Even so, I don't expect Big Law to touch the billable hour anytime soon. The incentives run the wrong way, so the structural roadblocks are too entrenched to move quickly. Meanwhile, AI has accelerated product development and go-to-market in ways we've never seen, and in-house legal teams everywhere are scrambling to absorb an order of magnitude more legal work. Alternative legal service providers can certainly help, but only with routine matters. And frontier LLMs, for all their power, were never designed to carry liability or catch the edge cases only a good lawyer spots.
That leaves a void in the market too big to ignore. A month at General Legal has shown me a glimpse of what comes next - a world where elite legal services are within reach of everyone. And the real work has only just begun.
Frequently Asked Questions
How does an AI-native legal workflow differ from traditional legal technology?
The article describes an AI-native workflow as one designed around AI from the beginning, with judgment-dense decisions codified into playbooks so the AI agent can take a stronger first pass while the attorney makes the final call.
Is AI already changing legal practice?
Yes. Current industry reporting shows law firms and AI companies are developing increasingly integrated AI tools for legal research, drafting, contract review, and internal workflows. For example, Reuters reported in April 2026 that Freshfields and Anthropic were jointly developing AI tools for legal services.
Is Harvard Business School incorporating AI into its MBA curriculum?
Yes. HBS currently lists Data Science and AI for Leaders as a required first-year course, and its MBA programme describes AI as being integrated throughout the curriculum.
What is the broader question about the future of legal practice?
The article's central question is not simply whether AI will replace lawyers. It is what parts of legal work depend on human judgment and how those capabilities can be combined with AI-native workflows.