Claude Code for Lawyers, Part 1: From a Court Text Message to a Litigation Response Workflow

AI-assisted English translation of the original Chinese article, retaining its historical context. Use the language switch to read the Chinese source.

The "Claude Code for Lawyers" series begins with this article.

From a lawyer's perspective, I will record what Claude Code (hereinafter referred to as CC) can do, to what extent, and what its boundaries are in real-life scenarios such as contract review, due diligence, legal opinions, litigation response, and customer communication.

The first article starts with a court text message.

This time, the lawyer did not open Word first, create a folder, or flip through the scanned documents first.

Only one thing was done: forward the court text message to CC.

[Court] The (2026) Anhui XXXX Minchu Case No.

In less than 10 minutes, CC automatically completed the follow-up actions: downloading and archiving the court scanned materials, identifying the indictment and evidence with OCR, organizing the case timeline, generating case analysis and litigation strategy, and drafting a complete set of first drafts of response documents such as a statement of defense, cross-examination opinions, and a list of supplementary materials.

This is not the "ask and answer" experience in web AI. It is more like inviting a hands-on AI assistant into the computer, allowing it to continuously run through an entire case-handling assembly line around local files, case catalogs, scanned materials and document templates.

1. Browser-based AI can answer questions, but it is difficult to complete the work.

Many lawyers have already used web AI. Paste the question in and it can summarize, rewrite, and draft a paragraph. This ability has value, but its boundaries are also clear:

1. It usually doesn’t know where your case folder is;

2. It cannot reliably download, rename, and archive a batch of scanned materials for you;

3. It requires you to constantly copy and paste, upload one by one, and ask questions repeatedly;

4. It is difficult to string "text messages, PDFs, pictures, evidence catalogs, document templates" into a continuous process;

5. It often gives you an answer rather than a set of implementable case work areas.

Therefore, web AI is more like a smart consulting object. You ask and it answers, and it stops when you stop.

The work in this article is difficult to complete with web AI. It is not a single question and answer, but a set of file-level workflows.

2. The difference between CC: it is not a chat box, but a workflow

The key difference of CC is that it can work continuously around local files and task targets.

After the text message is sent, it does not simply explain the meaning of the text message, but treats the text message as the entry point to the case and automatically breaks it down into a series of actions.

Task	Web AI	CC workflow
Read court text messages	Can understand text but usually remains stuck in conversation	Use text messages as case process triggers
Download scanned materials	Unable to automatically enter the local file system and case directory	Can be archived, named and organized around local folders
OCR and batch processing	It usually requires manual uploading one by one and questioning round by round.	Continuous processing of entire batches of PDFs, images and text
Generate multiple documents	Easily generate only a single answer	Can simultaneously output timeline, analysis, strategy, defense statement, and cross-examination opinions
Leaving traces and reusing	After the conversation is over, it is difficult to settle it into a workflow	Catalogs, rules, and templates can be precipitated into reusable processes

This is the watershed between the two ways of using AI: web AI solves "how to answer this question", and CC solves "how to get this done".

3. What does CC automatically do after the text message is sent?

The first step is to identify the case information from the text message: case number, court, electronic service matters, our identity, and submission deadline. Then create a case workspace and arrange the defense and evidence nodes.

The second step is to process the materials served by the court. The complaint, evidence catalog, subpoena, notice of evidence, and scanned copies of the plaintiff’s evidence are automatically filed in the case catalog:

01－Court serves materials

02: Statement of Claim and Relief Sought

03－Scanned copy of plaintiff’s evidence

04－We have materials to be replenished

05－Reply and Cross-Examination Opinions

06－Case Analysis and Litigation Strategy

07－Case Timeline

The third step is to perform OCR on the scanned PDF, turning the visible but unretrievable image text into analyzable case file text.

The fourth step is to generate the first batch of results based on the indictment and evidence. In less than 10 minutes, the lawyer already had a package of response materials on his desk.

4. In less than 10 minutes, more than just a statement of defense will be generated

What is really valuable is that it generates a set of interconnected materials at once:

1. Case timeline: arrange the contract signing, performance, dispute occurrence, reminder, prosecution, service and other nodes according to time;

2. Case analysis report: Break down the plaintiff's requested relief, factual allegations, supporting evidence and disputed issues;

3. Litigation strategy: List the matters we can admit, key defense directions, risk points and reinforcement paths;

4. First draft of the defense: Respond to the claims item by item to form a draft document that can be modified;

5. First draft of cross-examination opinions: Provide review opinions on the authenticity, relevance and legality of the plaintiff’s evidence piece by piece;

6. Supplementary materials list: Tell the customer what contracts, payment vouchers, chat records or performance materials still need to be provided;

7. Pre-trial communication outline: Next time you have a meeting with your client, you can check the facts directly by question.

This is not "AI writes an article for the lawyer", but AI first organizes the case materials into a state where the lawyer can take over the judgment.

5. Only when 80% of the repetitive work is taken away will the 20% capability gap be revealed.

It’s no secret that there’s a lot of work in legal services, but it’s time-consuming. Downloading materials, changing file names, organizing directories, OCR, excerpting claims, arranging timelines, making evidence tables, and framing defense statements are important, but they are highly repetitive in nature.

In the past, a lot of lawyers' energy was consumed by this 80%. Whoever is more careful and who can endure better will be able to do a solid job in basic work.

But after the AI workflow matures, these 80% will be standardized, streamlined, and automated faster and faster. By then, what really widens the gap between lawyers is no longer whether they can organize materials, but the remaining 20%:

1. Can you see the breakpoints in the other party’s evidence chain?

2. Can you judge which facts should be acknowledged and which facts must be defended;

3. Can you design a more successful litigation strategy?

4. Can you explain legal risks into options that customers can understand;

5. Can you assume judgment and responsibility at key nodes?

The more powerful AI becomes, the less likely lawyers will be able to just be material movers. When repetitive work is taken away, true expertise is magnified and compared more clearly.

6. Lawyers are not replaced, but forced to upgrade

CC did not appear in court for the lawyer, nor did he sign for the lawyer. What it does is push the work that would have taken a long time to the lawyers in advance.

When the lawyer takes over, he is no longer faced with a blank Word, but a case timeline, case analysis, litigation strategy, defense statement, cross-examination opinions and a list of supplementary materials.

At this time, the lawyer's work becomes more pure: reviewing facts, verifying evidence, judging risks, adjusting strategies, and being responsible to clients.

AI does not make lawyers easier, but allows lawyers to enter the part that truly tests their abilities faster.

7. Boundaries must be clearly stated

1. The case number, deadline, and court information must be manually verified;

2. The OCR recognition results must be returned to the scanned original for review;

3. The case analysis, litigation strategy, defense statement and cross-examination opinions generated by AI are only the first draft;

4. Customer identity information, contact information, amount, and private materials should be desensitized first;

5. Every word ultimately submitted to the court must be the responsibility of the lawyer.

AI can advance the work to the lawyers, but it cannot bear the final judgment for the lawyers. The bottom line of legal services is still accurate facts, solid evidence, and clear responsibilities.

Conclusion: From being able to use AI to being able to command AI to work

A court text message used to be just a reminder; now, it can become the entrance to a case workflow.

Browser-based AI allows lawyers to feel for the first time that "a large model will answer." Workflow tools like CC allow lawyers to further see that "big models can work."

This difference will be the source of the efficiency gap in legal services in the future.

Leave 80% of the repetitive work to the process and leave 20% of the key judgments to the lawyers. Whoever can complete this division of labor earlier will be able to spend his time on areas of real value.