AI transformation studio · Est. 2024 · Lawyers to legal departments
Every legal organisation says it uses AI. Ask which step, and the room goes quiet.
Buying AI is easy. Installing it is the job.
We name the steps your organisation actually runs on, systematise them, then automate them - in that order, without exception. Four to eight weeks, fixed fee, built on the structure you already have.
adnah is an AI transformation studio for lawyers, law firms of every size, in-house legal departments, and government legal functions.
Read why the order matters →What the call actually is →Two minutes, no email required. A scored assessment, not a sales call.
Built by a practising lawyer across India, the UAE and Singapore.
Marks the founder has worked alongside
6
Primitives that describe every legal organisation
2
Tracks - AI-Native or AI Adoption
3
Jurisdictions of operating experience
4-8
Weeks per transformation, on average
01 / The MATTER Method
Six primitives. The complete anatomy of a legal practice.
Matter
The atomic unit of legal work. One matter, one place, one state - legible to any competent stranger in ten minutes.
Actions
The named, repeatable moves that carry a matter from opened to resolved. Externalised playbooks, not private expertise.
Time
Managed deadlines and measured durations. Fixed-fee pricing becomes arithmetic on data, not a guess.
Team
Explicit roles on every action - doer, reviewer, decider. The delegation structure that makes AI safe to deploy.
Evidence
Every assertion anchored to a source. Work product becomes self-verifying; hallucinations become impossible to ship.
Research
A compounding stock of firm knowledge, not a folder of dead memos. The practice remembers what it has learnt.
02 / The Transformation
Three phases. Four to eight weeks. Working artefacts by week two.
The order is fixed. Skipping a phase produces expensive chaos at higher speed, which is what most of the profession is currently doing.
- 01
Phase 01
Weeks 1–2
Name
We spend the first stretch inside your organisation - matters, calendars, drafts, invoices, sign-off chains, the messy middle. Then we hand back a written map: your six primitives, in your language, with the gaps circled in red, and which track you're on.
- ·Structural audit
- ·Primitive map
- ·Bottleneck brief
- 02
Phase 02
Weeks 3–5
Systematise
Templates, playbooks, role charts, deadline structure. The work product of this phase is an organisation that would keep running - cleanly - even if we removed every AI tool tomorrow. That is the point, whichever track you're on.
- ·Matter template
- ·Action library
- ·Role chart
- 03
Phase 03
Weeks 6–8, occasionally to 10
Automate
Only now do agents get deployed - on the actions we named, inside the roles we defined, against the evidence rules we set, on MatterOS, LexOS, custom-built tools, or the market software that fits best. Automation without the prior two phases is theatre. With them, it is leverage.
- ·Intake agent
- ·Drafting agents
- ·Research stock
Fragment 04 · The AI-Native Practice
03 / The Thesis
The billable hour taught the profession to record time, not manage it.
Big-firm economics run on leverage: the associate is not the labour, the associate is the margin. AI removes exactly the work the associate is paid to do, which makes the tools an attack on the model, not an upgrade to it.
Independent firms and lean legal teams have the opposite problem - too much work, too few hours, no pyramid to defend. Every automated hour is recovered life. Identical technology, opposite economics.
adnah exists for both. We do not sell tools. We build the operating model - AI-Native where the structure allows it, AI Adoption where it does not.
04 / Before / After
What actually changes, line by line.
Every column on the right is a directly falsifiable claim about your practice after the transformation. We track them.
05 / The Fit Test
We say no more often than we say yes. On purpose.
A yes usually looks like
- +Any size - solo practice to a large firm, in-house department, or government legal function
- +You already have real matters and real work moving
- +Someone in the room can actually decide something on a Tuesday
- +You want to stop selling hours or stop drowning in volume
- +You are willing to change process, not just add tools
- +You want us to say which track - AI-Native or AI Adoption - applies to you, honestly
A no usually looks like
- −You want a pilot with no owner and no decision at the end
- −You need a multi-year procurement cycle before anyone will talk
- −Your real bottleneck is that you need more clients, not a rebuilt operating model
- −You want a tool bought, not a process rebuilt or installed
- −You want us to run it forever, not hand it back
06 / AI Readiness Audit
Which stage is your practice actually at?
Twelve questions - two per MATTER primitive - scored with a full breakdown across Matter, Actions, Time, Team, Evidence and Research. Most practices assume they're further along than the scoring shows - that gap is the whole reason transformations work.
Get my score- Stage 1 - Manual01
- Stage 2 - Assisted02
- Stage 3 - Augmented03
- Stage 4 - AI-Native04
Two minutes. No email required to see your score.
07 / Engagement Model
One way to work with us. Optionally, a lighter one after.
We hold ourselves to the model we run: fixed price, named deliverables, honest scope, no lock-in.
The full pass. Name, systematise, automate - the operating model delivered end to end, working artefacts inside the organisation from week two. AI-Native or AI Adoption, decided up front.
- ·Structural audit of what you use now, and how
- ·Written map of your six primitives
- ·Matter template, action library, role chart
- ·MatterOS, LexOS, custom-built, or market software - whichever fits
- ·Handover so the organisation runs it, not us
Optional. A light monthly relationship that extends the model as the practice grows - new actions, new agents, quarterly review, no lock-in.
- ·New actions added as needed
- ·Agent tuning against real matters
- ·Quarterly primitive review
- ·One month notice, either side, always
08 / Our Work
View our work09 / Studio Voice
“The firms that will define the next decade of legal work are already open. They are two lawyers, or five, in a room you have never heard of. Our job is to make sure the room runs like the next decade, not the last one.”
Raghav R Handa · Founder, adnah
10 / Client Voices
Anonymised, in their own words.
Firm names withheld at the practice's request. Each quote is from a live transformation between 2024 and 2026.
“By the fourth week we had stopped talking about AI and started running the practice. That was the point I understood what they'd actually built.”
Managing partner
IP litigation shop · India
“The transformation felt less like a consulting engagement and more like someone rebuilding the engine while I kept driving. The firm didn't stop for a day.”
Founder
Immigration boutique · Singapore
“Adnah refused half the things I asked for. That was the sign they were serious. Everyone else had said yes to everything.”
Solo practitioner
Cross-border arbitration · UAE
11 / The First Fortnight
Working artefacts by day fourteen. Not slides.
This is the honest calendar of the first two weeks of a transformation - from read-only access to a first agent shipping on real work.
Day 01
Read-only, everywhere
We take read access to your case system, drive, calendar and billing. Nothing changes for the organisation. We start building the map, and start seeing which track you're on.
Day 02
The structural audit
First working session with the people who run the work. We name what a matter is, what an action is, and where the definitions currently break.
Day 03
The bottleneck brief
A one-page written brief goes back to you: the two or three specific places where the practice is losing time and pricing accuracy right now.
Day 05
First matter template
A working matter container lands in the practice on the first live matter of the week. Not a mock. A real matter, running on the new container.
Day 09
First action playbook
One repeating action - usually intake or first draft - gets extracted into a named playbook and used by an actual person on an actual matter.
Day 14
First agent, in production
One drafting agent deploys against a named action, inside the role structure, with the evidence rule enforced. It ships on real work by end of fortnight two.
12 / Questions we get
Short answers, held honestly.
If your question isn't here, put it in the intake note - we read every one.
Are you a software company?
No. We do not sell software, licences, or dashboards. What we deliver is a set of primitives your organisation owns, running on MatterOS, LexOS, custom-built tools, or market software - whichever fits.
Do you work with big firms?
Yes, along with solo practitioners, in-house departments, and government legal functions. A large or heavily regulated organisation usually cannot go fully AI-native on any honest timeline, so we run AI Adoption instead - same primitives, same sequence, installed within the structure that already exists. We say which track applies to you before any work begins.
How many organisations do you take on?
A deliberately limited number, each year. The founder and the team behind him will not stretch thin enough to make the model dishonest. If the fit is right and the calendar is full, we say so and give a date.
Where are you based?
The studio is remote-first, async-first. The founder is a practising lawyer in India, UAE and Singapore. Transformations happen wherever the organisation is; most sessions are video, some are on-site by request.
What does a transformation cost?
Fixed fee, quoted after the intake call, scaled to the size and constraints of your organisation. Range is public on request. We hold ourselves to the pricing model we run.
What happens after the transformation?
Either you run it - that is the success case - or we keep a light monthly retainer if you want us present as the organisation grows. There is no lock-in, ever.
13 / Engagement
Start your AI transformation.
We take on a deliberately limited number of organisations. Send the shape of your practice or department, the bottleneck you can't get past, and the model you're trying to reach.
