The AI drafts, the lawyer decides.
Prompts built around Turkish law produce a draft petition, a client reply or a file analysis. The model runs on a local system on MTY Cloud, inside Turkey, and you are the one who reads, corrects and approves the text.
AI draft generation screen
- Choose the draft typeOn the assistant page you open Generate Draft and pick the kind of document to prepare.
- Enter the case and contextOptionally pick the related case, then write the party details and the content you want in the context field.
- The draft is generatedOnce the request is sent the button switches to Generating; the form stays open until the reply arrives.
- In the result cardThe form closes and the generated text appears in the Result card.
- Lawyer reviewYou read and correct the text, and save it to the file only after you approve it.
Choose the draft type
01/05
Human reviewThe prompt on the left, the draft on the right.
You write what you need; the assistant answers with a draft built on a Turkish-law system prompt. The draft opens in the editor and does not enter the file until you have corrected it.
Prompt
Prepare a statement of defence for case 2026/42 E. The claimant's allegation: a receivable arising from a contract. Our document: payment receipts.
Draft
To the Honourable Court, we submit our response to the claimant's allegation of a receivable.
The payment receipts show that the debt was paid before the action was filed, so the claim should be dismissed.
We request that the costs of the proceedings and the attorney's fee be borne by the claimant.
Sample record; real output depends on your file.
The AI is an assistant, not a replacement for the lawyer.
Every draft, message and analysis is informational. Checking sources, legislation and facts stays with the lawyer, which is why outputs are always editable and presented for review before you save.
02/05
Four jobs, all inside the file.
- 01/04
Petition
Pick the case file and write what you need: a statement of defence, a formal notice, an appeal. The file data becomes the context of the prompt; you edit the draft in your own style and attach it to the case record.
- Prompt
- Prepare a formal notice for the other party in 2026/42 E., with a 7-day deadline.
- Sample output
Dear Sir or Madam, the debt arising from the contract must be paid within seven days of service of this notice.
Otherwise legal action will be taken.
- 02/04
Client message
Paste the client's message and choose the tone of the reply. The assistant prepares a clear, courteous answer; you read and correct it before sending. The reply stays attached to the client record.
- Prompt
- The client asks: when is the hearing and what should I prepare? Reply calmly and clearly.
- Sample output
Hello, the hearing is on the date shown in the file, and it is already in your agenda.
I will send you the list of documents to prepare separately.
- 03/04
File analysis
Ask for an analysis of your file and the assistant returns a summary, risks, opportunities and a score. The result is stored as an analysis record you can open, update or delete. You choose what goes into the analysis; no document is sent on its own.
- Prompt
- Analyse file 2026/42 E.: give a summary, risks and opportunities.
- Sample output
Summary: a contract receivable case; the defence rests on payment records.
Risk: the claimant may not accept the payment receipts.
Opportunity: verifying the receipts through an expert report.
- 04/04
Outcome estimate
From the information in the file the assistant offers a possible outcome estimate. It is informational, not legal advice or a guarantee, and every output carries a disclaimer. The decision belongs to the lawyer who knows the file.
- Prompt
- Give a possible outcome estimate for 2026/42 E. and summarise the reasoning.
- Sample output
Based on the evidence, a possible outcome: the claim is partly upheld.
This output is informational and does not replace legal advice.
03/05
The quota is counted in a separate package.
HukukBis AI is not part of the Starter, Professional or Corporate plans; it is a separate package you add to any plan. Every request is deducted from the package's monthly quota and the counter shows what is left. When the limit is reached new requests are declined, and the counter resets at the start of the next period. The price is a quote based on the monthly request quota; see the HukukBis AI package on the pricing page.
Sample counter; the quota follows your package.
04/05
Data stays in Turkey.
The AI runs on a local model hosted on MTY Cloud. Your prompts and the case context you select are never sent to a provider abroad. The general data protection approach is described on the security and KVKK page.
- Local model
Runs on a local model hosted on MTY Cloud.
- No transfer
No data leaves Turkey for the AI.
- Not for training
Client data is not used to train the model.
- Firm decision
You switch the AI features on or off per firm.
- Saved on approval
An output is not part of the file until you save it; analysis records are stored privately for your company.
05/05
What do lawyers ask about the AI assistant?
- 01/05
Are the outputs legal advice?
No. Drafts, message replies, analyses and outcome estimates are informational. Checking legislation, case law and facts, and the final decision, stay with the lawyer; lawyer review is essential.
- 02/05
Is my data used to train the model?
No. Client data is not used for model training. The AI runs on a local model on MTY Cloud, entirely inside Turkey, and no data is transferred abroad for it. Details are on the security and KVKK page.
- 03/05
Is the AI included in my plan?
No. HukukBis AI is a separate package that can be added to any plan. The price is a quote based on the monthly request quota; each request is deducted from the quota and the counter shows what is left. See the pricing page for a quote.
- 04/05
Which language does it work in?
The assistant produces Turkish drafts and replies from prompts built around Turkish law. The HukukBis interface is available in Turkish and English.
- 05/05
Does document content go into the analysis?
You choose which case information and content goes into the analysis. No document is sent to the model unless you ask for it. For the documents themselves see document management.