Legal tech 7 min
The lawyer who codes: how AI is changing what it means to practise law
1 April 2026
There is a particular kind of lawyer emerging right now. You might have met one. They draft contracts in the morning, review case law in the afternoon, and at some point during the day they open a terminal — or a chat window — and ask an AI to do something the rest of the firm does not know is possible yet.
They are not software engineers. They do not have computer science degrees. But they have learned, mostly out of curiosity, that the tools available to them in 2026 are not what they were in 2022. And that the distance between knowing that and acting on it is now a professional advantage.
This is a piece about that change. About what is actually happening in the legal profession, and what it means for the lawyers who are paying attention.
The end of legal work as a craft without tools
Law has always been a knowledge profession. The value a lawyer provides has historically been inseparable from time — time spent reading, analysing, drafting, advising. Billable hours exist because that time was genuinely scarce and genuinely skilled.
That scarcity is changing, unevenly and fast.
Large language models — the technology behind Claude, GPT-4, and their successors — are not replacing lawyers. But they are collapsing the time required for a significant portion of legal work. First-draft contract review. Precedent research. Summarising case bundles. Translating between jurisdictions. Generating standard correspondence. Work that once took an associate a morning can now take an AI three minutes and a lawyer fifteen to review.
This is not a threat if you see it clearly. It is a reallocation of time — away from the mechanical and toward the judgmental. The lawyer’s irreplaceable value has always been in the judgement. AI does not have that. What it has is speed, breadth, and tirelessness. The combination of human judgement and machine speed is not a diminished lawyer. It is a more capable one.
What “legal tech literacy” actually means now
A few years ago, legal tech literacy meant knowing how to use your firm’s document management system. Then it meant knowing which e-discovery platform your jurisdiction used. Then it meant being comfortable with contract lifecycle management software.
Now it means something different and more fundamental: the ability to direct AI systems to do useful legal work, and to evaluate what they produce.
This is sometimes called prompt engineering, a phrase that has aged poorly. What it actually describes is closer to delegation — the skill of giving a capable but literal-minded colleague a precise enough brief that they come back with something you can use. Every experienced lawyer already does this with junior associates. Doing it with an AI requires the same instincts: specificity, context, scepticism about the first draft.
Beyond that, a new layer is emerging. Tools built on the Model Context Protocol — MCP — allow AI systems to connect directly to the software lawyers already use. Not just to generate text, but to take actions. To log a time entry. To pull a matter’s history. To draft from a template stored in the firm’s system. To create a task in a matter management tool.
The lawyer who understands how to configure these connections — who knows what their AI assistant can reach and what it cannot, and what permissions that implies — is operating at a different level to one who uses AI only as a better search engine.
The coder-lawyer is not a fantasy
In Silicon Valley, a decade ago, there was a wave of “design thinking” in law — the idea that lawyers needed to think like product designers to survive the coming disruption. Much of that turned out to be overstated, at least in its original form. Law is not a product. Clients are not users. The analogy had limits.
The coder-lawyer analogy is different, because it is not an analogy. It is increasingly literal.
Not every lawyer needs to write code. But the lawyers who understand the logic of code — who can read a JSON response from an API, who understand what a webhook does, who can configure an MCP server without calling IT — these lawyers can automate things that other lawyers cannot. They can build workflows that compound over time. They can connect their practice management software to their AI assistant to their document repository in ways that save hours every week, permanently.
This is not the domain of legal technology specialists sitting in a separate team. It is becoming a skill that lives in the same person as the legal knowledge — the same way commercial lawyers developed financial literacy not to become accountants, but to advise clients properly.
The Danish bar has always valued technical precision. The move into technical fluency is a natural extension of that.
What this looks like in practice
Consider a senior associate in a Copenhagen firm handling a portfolio of employment matters. On a Monday morning in 2026, their workflow might look like this:
They open Claude. They describe the three client calls they have that morning — the matters, the open questions, the likely outputs. Claude pulls the relevant matter histories from LexTime via MCP, surfaces the last three time entries for context, and flags an unanswered question from a previous session. Before the first call begins, they have a brief for each matter that would previously have taken thirty minutes to assemble from scattered notes and a billing system.
After each call, they dictate a two-sentence summary to Claude. Claude logs the time entry to the correct matter in LexTime, drafts the follow-up correspondence, and marks the action items. The associate reviews, adjusts one sentence, and approves. The entire post-call administration takes four minutes instead of twenty.
At the end of the week, they ask Claude to generate the timesheets for the three matters with active billing cycles. LexTime produces the PDFs. The associate reviews them, sends them, and goes home.
None of this required writing a line of code. It required understanding which tools were connected, what they could do, and how to ask for it clearly.
The firms that are not paying attention
For every lawyer experimenting with MCP connections and AI-assisted matter management, there are twenty who are not. This is normal. Technology adoption in law has always been slow — the profession has good reasons to be conservative about change, particularly change that touches how advice is given and how records are kept.
But the gap between the two groups is widening faster than it has before. The compounding nature of AI-assisted workflows means that firms investing in this now are not just faster — they are building institutional knowledge about what works, training better internal systems, and attracting the kind of junior lawyers who expect these tools to be available.
The firms that wait are not standing still. They are falling behind a moving benchmark.
Why time is where this starts
Of all the ways AI is entering legal practice, time registration is the most immediate and the most overlooked.
It is immediate because the problem is daily and universal. Every lawyer in every firm, regardless of practice area or seniority, has to record time. It is the connective tissue of the business. Get it right and cash flow follows. Get it wrong — late entries, vague descriptions, rounded numbers — and the invoice is the first thing a client questions.
It is overlooked because it feels administrative. Lawyers do not go into law to fill in timesheets. They tolerate it. And that tolerance is precisely what makes it a good place to start: the AI that takes this off a lawyer’s plate is not competing with their most valued work. It is clearing space for it.
An AI connected to your time tracking system via MCP is not a novelty. It is the baseline infrastructure for a practice that runs efficiently in 2026. From there, the same connection, the same interface, the same model — can be extended to document drafting, matter summaries, client correspondence, deadline tracking. The integration deepens as the trust builds.
A note on what LexTime is building toward
LexTime is time tracking software for small law firms. That is where it starts. But the MCP connection — the ability for Claude to log time, retrieve matter histories, and generate timesheets on a lawyer’s behalf — is not a feature added to a time tracker. It is the foundation of something larger.
The vision is a practice where the AI assistant knows the context of every matter, can act within the firm’s systems with appropriate permissions, and handles the administrative layer of legal work so completely that the lawyer barely notices it happening. Where time is captured because the work was described, not because someone remembered to click a button. Where the timesheet is a byproduct of the conversation, not a separate task at the end of the week.
That is not a distant future. The infrastructure for it exists today. The lawyers who are building these habits now — who are learning to describe their work to an AI, who are connecting their tools, who are comfortable with the idea of a machine acting on their behalf within defined limits — are not chasing a trend. They are practising, quietly, what legal work will look like in five years.
The AI-empowered lawyer is not a different kind of lawyer. They are a better-equipped version of the same one.
Try LexTime free for 30 days.
No credit card required.
Start free trial