From Will Drafting to Family Matters: How a Prestigious Law Firm Turns Legal Expertise into Repeatable Workflows

When lawyers use AI, are they "asking AI" or "delegating tasks to AI"?

This distinction precisely defines the boundary between AI tools and AI workflows: the former merely improves efficiency at a single point, while the latter represents genuine process reengineering.

Starting August 2026, multiple lawyers at DeHeng (Hangzhou) Law Firm began integrating years of professional experience into Qwen Office, transforming it into reusable workflows. Even users without legal backgrounds can complete relevant consultations through guided interactions, while fellow legal professionals can significantly enhance case-handling efficiency using these workflows.

Lawyers Cai Huijuan and Xu Yan encapsulated over a decade of will service experience into a Skill titled "Will Drafting and Practical Guidance." Without needing any legal foundation, users can generate a comprehensive first draft of a will—complete with risk alerts—through approximately five minutes of conversational Q&A, sparking broader social discussions on topics such as digital asset inheritance and younger generations drafting wills earlier in life.

The Lian Yindi legal team launched the "Family Law AI Expert Suite," embedding years of case-handling expertise into modules for fact organization, asset investigation, and agreement generation, comprehensively covering divorce agreements, property division, child custody, and other family law scenarios.

Below is a first-person account based on Xu Yan's presentation:

Hello everyone in the legal tech community. I'm Xu Yan from DeHeng (Hangzhou) Law Firm.

Handling legal cases is like crafting handmade art. When identical case materials are given to different lawyers, the resulting litigation strategies and quality of legal documents may vary drastically. Precision stems from a lawyer’s experience and judgment, while the sense of burden arises from the large volume of work still requiring manual execution.

Documents, WeChat records, audio recordings, and video files sent by clients must first be read, timelines clarified, and dispute points extracted before proceeding to legal research, evidence analysis, and document drafting. Dozens of interdependent steps mean that one incomplete step blocks the next. Lawyers often spend their time on repetitive tasks such as searching, excerpting, organizing, and formatting.

With the arrival of AI, we started “asking AI”—for instance, inputting materials into a large model to generate complaints, answer legal questions, or conduct legal research.

This indeed improved efficiency. However, upon closer examination, we found that sometimes up to 80% of the AI-generated content required editing. Time originally spent on “writing” shifted to “revising,” while the workflow sequence, review burden, and accountability boundaries remained fundamentally unchanged.

Xu Yan, Lawyer at DeHeng (Hangzhou) Law Firm and Deputy Director of the Artificial Intelligence Committee

This does not constitute true workflow reengineering.

True reengineering transforms the original linear process into two parallel tracks: one handled by an intelligent agent responsible for material parsing, fact decomposition, evidence analysis, legal research, and initial draft generation; the other where the lawyer focuses on client communication, goal setting, critical judgment, strategy formulation, material selection, and final accountability.

01 From "Conversation" to "Task Delegation": A Fundamental Shift in Using AI

In practice, we discovered that the core shift in reengineering workflows lies in two words: moving from "conversation" to "task delegation." Previously, "asking AI" was essentially a single-point task—a form of dialogue. Now, it resembles delegating work to an assistant: clearly specifying the case background, objectives, materials, steps, and delivery standards, allowing the intelligent agent to break down complex problems, activate relevant skills, and deliver results.

At this stage, what lawyers need to master is no longer writing complex prompts, but how to clearly assign tasks, set standards, and evaluate outcomes. For example, for fact organization, instead of lengthy prompts, we now simply say: "Please organize the facts of this case." Qwen Office automatically invokes the litigation suite to produce structured output. Similarly, a single instruction can trigger pre-trial evidence analysis, reviewing materials from both opposing counsel and the judge’s perspective.

This capability is powered by Skills and expert suites.

A Skill is not a prompt—it defines triggering conditions, required inputs, execution steps, verification methods, and reference scripts, forming a reusable methodology. When multiple Skills are connected under a unified logic, they become an expert suite. It doesn’t just complete isolated tasks; it accompanies the lawyer throughout the entire business process.

02 Real-World Cases: The Will Skill and DeHeng Family Law Suite

Recently, our team launched the "Will Skill" on Qwen Office. Users input basic information and intentions via conversation, and AI generates a structurally sound, fully compliant first draft of a will based on a professional legal framework. While the media attention was notable, what mattered more to me was solidifying our team’s professional procedures, review habits, and risk awareness into a repeatable methodology.

This also marks our first step toward reimagining legal services through office-based AI agents in the realm of public-interest legal work.

Together with Attorney Lian Yindi’s team, we developed the "Family Law Expert Suite." Family law issues involve complex regulations, emotional entanglements, and diverse documentation. We packaged years of experience into multiple Skills, creating a workflow that covers the entire consultation and agreement drafting process:

Fact Organization Skill: Automatically analyzes communication logs and materials to extract key timelines and dispute points.

Asset Investigation & Analysis Skill: Structures marital assets and debts, generating division recommendations.

Agreement Generation Skill: Automatically drafts various agreements at different stages according to needs.

The entire suite transforms team experience from "individual memory" into "systematic assets," enabling every new case to build upon accumulated historical expertise.

03 Scheduled Tasks and Web Publishing: Underestimated Yet Powerful Use Cases

Beyond case handling, two seemingly minor yet highly practical scenarios stand out.

The first is scheduled industry updates. Our team specializes in capital markets dispute resolution, requiring continuous tracking of industry developments. Manual searches were unsustainable. Now, with scheduled tasks set once, Qwen Office automatically retrieves, organizes, and formats weekly industry legal reports—including charts, source links, and table of contents—requiring only one minute of setup.

The second is client relationship management. We created a legal information website backed by a database, continuously updating laws and policies relevant to our clients and pushing updates weekly. Clients simply click a link to access the latest content.

One client commented: "Attorney Xu, you really care about us." For clients, being consistently followed up with is itself part of high-quality service. For us, Qwen Office handles information aggregation and formatting, freeing us to focus on curating truly relevant content.

04 Turning Experience into Knowledge Assets: Every Lawyer Can Build Their Own Skill

Many lawyers assume creating Skills requires high technical proficiency. In reality, they already use their own "invisible Skills" daily: knowing what to check first upon receiving materials, how to assess factual completeness, which angles to examine evidence from, what additional information to request, and what items to verify before delivery.

Traditionally, lawyer development depends on years of case accumulation—valuable yet vulnerable to loss through staff turnover. Skills offer a new path: turning experience from a consumable into a compounding asset, transforming individual expertise into a shared team foundation.

And implementing all this on Qwen Office is remarkably simple—sometimes all you need is a few keystrokes.

Conclusion

AI can act as a search engine, tool, assistant, or partner—but most importantly, it’s about our mindset toward AI. Through AI, we can finally reclaim time for deep thinking.

The first half of 2026 Shanghai Legal Industry Achievements Release and Special Seminar on AI Legal Services Application was hosted by the Shanghai Lawyers Association's Information Technology and Legal Tech Committees, organized by Alibaba's Qwen Office, and co-organized by Deli Technology.

We dedicated to serving clients with professional DingTalk solutions. If you'd like to learn more about DingTalk platform applications, feel free to contact our online customer service or email at This email address is being protected from spambots. You need JavaScript enabled to view it.. With a skilled development and operations team and extensive market experience, we’re ready to deliver expert DingTalk services and solutions tailored to your needs!