Strategic Vision

My Daily Driver Tech Stack in 2026

Drafting is not a document. It is a system. Here is what I use every day.
Pranav RamakrishnanAdvocate & Legal Engineer
January 9, 2026
12 min read
TwLiLink

Headnote

The last meaningful improvement to legal drafting tools was Microsoft Word. That was decades ago. I stopped using Word when drafting moved upstream into a system that understands intent and sequencing. Word became what it should have always been: the export format.

Is Drafting a Document or a System?

The last meaningful improvement to legal drafting tools for lawyers was Microsoft Word.

And that was decades ago.

Since then, law has grown more complex while the tools have stayed flat. Pleadings and contracts are still treated as pages instead of systems.

But law does not operate on pages. It operates through sequences of procedural bets, mapped obligations, and fallback positions under pressure.

I stopped using Word when I got early access to jhana Drafting. Drafting moved upstream into a system that understands intent and sequencing.

Key Takeaway

Word became what it should have always been. The export format.

Hardware That Preserves Legal Context

Once drafting moved upstream, hardware stopped being about power and started being about continuity.

My daily stack is a MacBook Pro, an iPad, and an iPhone. Not because of brand loyalty, but because Apple treats state as a first-class object. Legal work lives in half-finished thoughts. Marked PDFs. Open issue lists. Drafts that are not ready to be drafts yet.

The Apple ecosystem preserves that state. A document is never closed, only paused. I can draft on a Mac, review the same reasoning path on an iPad with a Pencil, and capture a procedural risk on my phone without rebuilding context.

Context Switching
This is the part most tools ignore. Lawyers do not switch devices. They switch moments, and Apple makes those transitions invisible.

LaaS: Law as a System

Law is not text. It is a system. Let’s call it LaaS: Law as a System.

A pleading is a sequence of procedural moves. A contract is an operating manual for future failure. Tools that treat law as pages force lawyers to carry structure and risk in their heads.

jhana is built for this model.

Paralegal

Handles drafting and research in one surface. Reads instructions and files together.

+ Briefs & Arguments+ Chronologies+ Draft filings+ Citation-linked outputs+ Work-till-done sessions
Suit

Ingests messy reality. Scanned annexures. Bulky filings. Handwritten records.

+ Structure extraction+ Timelines & indices+ Limitation tables+ Cited notes+ Full-record view

But lawyering is not just producing documents.

It is judgment. Sequencing. Knowing what to lead with, what to hold back, and where risk actually sits. That is what Playbook captures.

As I use Paralegal, Playbook starts to reflect how I actually practise. The order I raise issues. Where I place procedure. What I treat as non-negotiable. It does this by watching patterns, not by forcing rules. Nothing is enforced. It surfaces suggestions and I decide what to do with them. I can accept them, ignore them, turn them into rules, or add my own preferences. Over time, this lets judgment carry forward across matters without hardening into templates.

Pro is what lets that judgment scale. It gives you the headroom to work through large matters without constantly hitting limits or breaking context. Long sessions stay intact. The full record stays in view. When Suit is attached, Paralegal is not working off snippets or summaries. It is working with the entire matter as a whole.

Key Takeaway

That is LaaS in practice.

A Lawyer’s Soundboard

This is not my legal stack.

This is where ideas get pressure-tested.

Alongside jhana, I run a local AI setup using Chatbox and Ollama. The model I use is GPT-oss-20b. It is open-weights, which means the model itself is available to run and inspect instead of being locked behind an API. It runs entirely on my own machine, with no remote calls or third-party servers, and that boundary is deliberate.

Ollama runs the model, and Chatbox is the interface I work in. I use this setup to think out loud, test framing, sanity-check timelines, and pressure-test how an argument might be attacked before it ever becomes a draft.

This setup is not for drafting or producing work product. It is a place to introduce friction early, before anything becomes formal.

The connective tissue is MCP. Model Context Protocol. A MCP lets the model interact with tools like Google Drive and Calendar in a limited, structured way. That makes it easier to reason about dates, task lists, and non-sensitive notes without pasting context into prompts.

For a lawyer, this works like a soundboard for non-substantive work. I use it for planning, sequencing, workflow design, and stress-testing logic before anything becomes real work product.

I do not upload confidential pleadings or full matter files here. That stays inside jhana, which is built for legal data, conflict safety, and full-record work.

Boundary
This is where I experiment. If something breaks, nothing client-facing does.

Where Conversations End and Cases Begin

Conversations and documents are not the same thing. I treat them differently because they demand different kinds of attention.

Read AI is my assistant for conversations. Calls, meetings, internal discussions. I am not looking for transcripts. I want outcomes. What was decided. What shifted. What now needs follow-up. Read turns spoken discussion into something I can act on, so nothing important stays stuck in memory or chat logs.

LiquidText is where documents live. This is my desk.

It is an interactive workspace where I can spread the record out. I connect highlights, notes, and different documents on a single canvas. Judgments, pleadings, contracts, compilations. Instead of scrolling, I link ideas, group issues, and follow threads across sources.

LiquidText replaces paper bundles. I can move material around, build context visually, and find facts by following connections instead of flipping pages. Pen, touch, and multi-monitor support matter here. This is review work that feels closer to thinking than reading.

Over time, it becomes a second brain. Not conclusions. Raw material, connected and reusable.

Key Takeaway

This is where the case starts to take shape.

Taking Dictation, Again

Some thinking is too fast for typing. Senior Advocates understood this long before software did.

When I was working under Senior Advocates, the instruction was simple. Take dictation. Sometimes someone else would type. Sometimes I would. The point was never the typing. It was speed. Get the argument out while it is still alive.

That instinct is what pulled me toward voice again.

Voice dictation quietly crossed a threshold with models like Whisper. Accuracy became good enough that speech could finally be trusted as input, not just convenience. That foundation has existed for a while. What was missing was a usable layer on top.

That is where Wispr Flow feels like a new sensation. The Whisper model has always been around, but Wispr Flow makes dictation usable in real workflows. Across apps. Low friction. You speak naturally. It writes cleanly. You move on.

I use Flow for early capture. Rough outlines. Issue sequencing. Emails. Notes while walking between meetings. Places where stopping to type would break momentum. For lawyers, that matters more than polish. Lose momentum and you lose the argument’s shape.

Feature: Steno by jhana

Steno was built for legal dictation from day one. With more reliable speech recognition underneath, this category is finally ready to matter again.

Voice is not a replacement for writing. It is how ideas survive long enough to be written properly.

Closing Thoughts

Control over work product has always mattered in legal practice.

As tools get faster and louder, that instinct matters even more. The question is not whether AI is powerful. It is whether it respects judgment, preserves context, and keeps the lawyer in the loop.

As the year winds down, this is a good moment to pause and ask a few honest questions.

Where does your thinking slow down unnecessarily?

Where does context get lost between reading, discussion, and drafting?

Where do good ideas disappear because capture is too hard or too late?

When considering AI or productivity tools, look past the demo. Ask what actually improves control. What reduces friction without taking ownership away. What fits into how law is really practiced.

AI should not replace lawyering. It should make room for it. Room for judgment. Room for sequencing. Room for the quiet work of reading, thinking, and deciding what actually matters. The best tools do not take the pen out of your hand. They remove the noise so you can use it properly.

That is the standard worth holding onto as this space keeps moving.

Hope everyone a restful holiday break. Wishing a clear, focused start to the year ahead!

Index Keywords
Tech StackLegal TechProductivityApple EcosystemLaaSLaw as a Systemjhana Paralegaljhana SuitPlaybookStenoOllamaMCPVoice DictationWispr FlowLiquidTextRead AILegal WorkflowsAI for LawyersDocument IntelligenceLegal AI 2026
See Also

Discussion