announced a strategic investment and partnership with Arabic.ai. 
LLM-X
Legal
Sharper preparation. Professional judgment.
Arabic AI for research, document comparison, drafting and multilingual delivery. Professionals stay responsible for interpretation and advice.
The quality of the work starts with its preparation.
Professionals work across instructions, sources, successive drafts and language versions. The original material stays central, and every change is a matter for their judgment.
- Source material spread across a matter
- Document changes that need interpretation
- Multilingual delivery with exacting review
Support each stage before the professional signs off.
Choose a task to see what goes in, what is prepared, who reviews it and what comes out.
Document and matter teams
Bring the change into focus.
Compare selected versions and assemble questions about the changed passages, with the source wording beside each one.
- Input
- Two supplied versions and the scope of the comparison.
- Assistance
- Prepares a summary of the differences and questions to examine.
- Review
- The responsible professional checks the original passages and their significance.
- Output
- A review pack with the changes and open questions.
Research and advisory teams
Build the brief around the supplied sources.
Organize a defined research set against a clear question, with findings kept apart from interpretation.
- Input
- Selected research documents and professional instructions.
- Assistance
- Structures the relevant passages into a working research brief.
- Review
- The practitioner verifies the sources, their currency and the interpretation.
- Output
- A source-led brief for professional review.
Professional and engagement teams
Get to a structured working draft.
Prepare an outline or draft from approved instructions, with missing information made explicit.
- Input
- An agreed brief, supplied references and drafting instructions.
- Assistance
- Develops a structure and drafts the requested sections.
- Review
- The professional checks substance, wording and fit with the instructions.
- Output
- A working draft ready for substantive review.
Language and engagement teams
Keep language work connected to the matter.
Coordinate the source, terminology and review for a multilingual deliverable, with language approval kept apart from approval of substance.
- Input
- Approved source material, terminology and a delivery brief.
- Assistance
- Drafts each language version with controlled terminology.
- Review
- Language reviewers check the meaning. Professionals approve the substance.
- Output
- Reviewed multilingual material for the agreed purpose.
Compare two versions of a service agreement.
Both excerpts stay side by side with a review question. The practitioner decides what the change means against the instructions.
Version AService period: 12 months
Version BService period: 18 months
Review question
Confirm the changed period against the instructions.
Responsible professional
The responsible professional verifies every material interpretation and approves any advice or final deliverable.
Keep professional review in the workflow.
Scope the document set, protect information boundaries and check every material interpretation against the source.
What reviewers check
- Source fidelity
- Material interpretation
- Review readiness
Set the matter and information boundaries.
Agree the permitted documents, purpose and review responsibility before confidential material or repositories are involved.
Evaluate the material differences.
Use representative versions, ambiguous passages and incomplete instructions. Check source fidelity and interpretation.
Refine through professional review.
Start with a defined group, examine the corrections it needs, then extend task by task.
Sovereign legal AI with Qistas.
Arabic.ai and Qistas combine LLM-X, LLM-S and on-premises deployment with Qistas's legal technology for research, litigation, transactions, and policy and legislative studies.
- Research
- Litigation
- Transactions
- Policy and legislative studies
Same models, same UAE legal questions. Only what they can read changes.
0.99With the statute in the prompt
0.495Without it
Best score by any model. Stanford CRFM, HELM Arabic Enterprise, May 2026.
What teams use.
The AI workspace
Find, compare and translate passages.
The agentic platform
Review workflows with named sign-off.
Make your next ambition operational.
Bring intelligence into your knowledge, people and operations. Talk to our team about the work you want to transform.