Dumping PM Excel Dumps into TRAE Work Spits Out a Full Frontend Spec in 30 Minutes
Foreword
Daily life of a frontend developer at a small company:
The PM sent over 3 Excel files, claiming they are the complete requirements for a new project. Upon opening them — 12 functional modules, with modules further divided into "Scenario 1~4". Fields have different sources under different scenarios ("Data Management Platform Push" / "BIMS Auto-fill" / "Settlement New" / "Manual Input"). Some cells are marked "None, needs to be added", some notes are hidden in corners, and some are completely invisible because the column width is too narrow.
No interaction drafts, no field specification documents. All the information is in these 3 Excel files. You have to figure out relationships like "what is the source of this field under Scenario 1, and what does it become under Scenario 3" by yourself.
Previously, sorting out this kind of requirement would take a rookie at least a few days. This time, using TRAE Work, it took 10 minutes to output a three-piece set: Page Fields + Interaction Logic Description + Development Task Breakdown.
1. Real Pain Points: Small Companies Have No Swagger, No Postman, Only the PM's Excel
I am responsible for the frontend of the company's BIMS system. This time, I needed to develop two major modules: "Settlement & Delivery". The PM provided 3 Excel files:
| File | Module | Number of Functions |
|---|---|---|
| File 1 | Unsettled | 1 function (Data Management Platform syncs information to BIMS) |
| File 2 | Settlement | 3 functions (Create Settlement + Settled List + Send Settlement Email) |
| File 3 | Delivery | 8 functions (Create Delivery Task + Undelivered + Supplement Delivery Info + BIMS Interop + Delivering + Send Delivery Email + Delivered + Re-deliver) |
Total: 12 functions, 4 business scenarios, hundreds of fields.
For this kind of Excel requirement, the standard process is like this:
- Read through the 3 Excel files, understand what each module does (clarify with PM promptly if unclear, taking anywhere from half a day to two or three days).
- Annotate the source of each field — "This field comes from Data Management Platform Push under Scenario 1, and from BIMS Auto-fill under Scenario 3" (This is for the backend to consider, but as a lead, I have to think about it too).
- Sort out cross-module dependencies — the Settlement module depends on fields from the Unsettled module, the Delivery module depends on fields from the Settlement module.
- Organize interaction logic — which fields are editable, which are read-only, which are shown/hidden based on conditions.
- Break down development tasks — by page, by function, by priority.
The Most Torturous Parts
Unknown field sources: The same "Contract No." field has a source of "Data Management Platform Push" in the Unsettled list, changes to "Unsettled" (i.e., inherited from the Unsettled list) in the Settled list, and becomes "Settled" in the Delivery list.
Intersection of 4 scenarios: The logic for different scenarios needs to be sorted out separately, and data sources are also different.
Overlooked notes: For example, there's a note in the Undelivered list: "When handling a Scenario 1 delivery, if this project has already delivered Scenario 2, then the content of the Scenario 2 delivery needs to be displayed here" — business rules hidden in cells like this are easy to miss if you're not careful.
Cell width causing oversight: Some cell content is very long (e.g., "Need to solve the problem of duplicate folder names in non-POD5 data paths: do not overwrite duplicate files during hard drive transfer OR modify the naming rules for non-POD5 data path folders to avoid duplicates"). If the column width is too narrow, the full content isn't displayed, and you won't know there's content inside unless you click on it.
Conditional logic in email templates: The subject of the settlement email has different formats when the customer batch number has content versus when it's empty. The body of the delivery email needs manual editing (filling in courier number, uploading hard drive photos). These conditional branches are scattered throughout the Excel.
Cross-file dependencies: Fields in the Delivery module heavily depend on the Settlement module, which in turn depends on the Unsettled module. But the information is scattered across 3 Excel files, so you have to switch back and forth to view them.
Proof in the picture
File content:
Enterprise WeChat has watermarks, it's too troublesome to deal with. Providing one just to show I'm not making this up!
2. Practical Process: Running the Full Workflow in 30 Minutes
Step 1: Directly Upload the 3 Excel Files
Open TRAE Work, drag and drop the 3 Excel files provided by the PM directly into the dialog box. No need to pre-organize or convert formats; upload them as they are.
Step 2: Enter the Organization Prompt
After the files are uploaded, directly enter the following prompt in the dialog box:
I have uploaded 3 Excel files, which are the requirements documents for the Settlement & Delivery modules of the BIMS system (Bioinformatics Management Platform) provided by the PM. There are 12 functional modules and 4 business scenarios in total.
Please organize them into a frontend development document, including the following content:
1. Page Fields:
- Group by functional module (Unsettled / Settlement / Delivery)
- List all fields for each function
- Annotate the source of each field under different scenarios (Data Management Platform Push / BIMS Auto-fill / BIMS Auto-generate / Settlement New / Manual Input / Inherited from Unsettled / Inherited from Settled)
- Annotate the input method for each field (Read-only / Manual Input / Manual Check Single / Manual Check Multiple / Manual Select Multiple / System Calculated / BIMS Edit Tag)
- Annotate whether the field has a corresponding field in the Data Management Platform
- Annotate missing fields ("None, needs to be added")
2. Interaction Logic Description:
- Operation buttons and trigger actions for each functional page
- Differences in page behavior under different scenarios
- Cross-module flow logic (Unsettled → Create Settlement → Settled → Create Delivery → Undelivered → Delivering → Delivered → Re-deliver)
- Conditional logic for sending emails (Customer batch number has content vs. is empty)
- Hidden business rules in the notes (Extract all, do not omit any)
3. Development Task Breakdown:
- Break down development tasks by page
- Annotate dependencies for each task (which tasks must be completed before others can start)
- Annotate priority (P0 Core Process / P1 Auxiliary Function / P2 Edge Function)
- Annotate technical difficulties and points needing confirmation
Step 3: TRAE Work Outputs the Full Development Document Set
After about 10 minutes, TRAE Work outputs:
The generation is very detailed:
It even thoughtfully provided points that need confirmation:
3. Deliverables and Results
Final Deliverables
Page Fields: Complete fields for 12 functions, annotated with the source, input method, Data Management Platform correspondence, and missing items for each field under different scenarios.
Interaction Logic Description: Cross-module flow logic + differences in 4 scenarios + 4 delivery method branches for supplementing delivery info + email conditional logic + full extraction of hidden business rules.
Development Task Breakdown: 18 development tasks, sorted by priority (P0/P1/P2), annotated with dependencies and technical difficulties.
Efficiency Comparison
| Dimension | Before (Manually chewing through Excel) | Now (TRAE Work) |
|---|---|---|
| Read 3 Excels, understand modules | Half a day minimum | Paste once |
| Annotate field sources (Scenario Matrix) | 60 minutes | Auto-generate field matrix |
| Sort out cross-module dependencies | Half a day minimum | Auto-annotate flow logic |
| Organize interaction logic | Several hours | Auto-generate description |
| Break down development tasks | 1-2 hours | Auto-breakdown + annotate deps |
| Extract hidden notes | Often overlooked | Extract all, no omissions |
| Total | About several days | About 30 minutes |
| Confusing field sources | Frequent rework | Matrix is clear at a glance |
| Omitting notes | Bugs after going live | All extracted and annotated |
4. Reusable Experience
1. Universal Prompt Template
Save it in a memo, send it directly after uploading the PM's Excel files each time:
I have uploaded the Excel requirements documents provided by the PM (functional modules, business scenarios, field sources).
Please organize them into a frontend development document:
1. Page Fields: Group by function, annotate the source and input method of fields under different scenarios, annotate missing fields.
2. Interaction Logic Description: Operation buttons, scenario differences, cross-module flow, email conditional logic, full extraction of hidden business rules.
3. Development Task Breakdown: Break down by page, annotate dependencies, priority, and technical difficulties.
2. Scenario-Specific Prompt Variants
Multi-Scenario Intersection Version (Same function has multiple scenarios, different field sources):
The same function has multiple business scenarios, please:
- Output a field × scenario matrix for each function
- Annotate the source of each field under each scenario (Push/Auto-fill/Auto-generate/Manual Input/Inherited)
- Annotate differences in fields between scenarios (present in one scenario but not another)
- Annotate missing fields ("None, needs to be added")
Hidden Note Extraction Version (Business rules hidden in corners of the Excel):
There may be notes and business rules hidden in the cells of the requirements document, please:
- Extract all, do not omit any
- Categorize by functional module
- Annotate the fields and interactions affected by each note
- Specifically annotate "If... then..." type conditional logic
Cross-Module Dependency Version (Multiple modules have field inheritance relationships):
There are field inheritance and data flow relationships between multiple functional modules, please:
- Output a cross-module flow diagram (A → B → C)
- Annotate which fields in each module are inherited from upstream modules
- Annotate which fields are newly added in this module
- Annotate which fields require cross-system synchronization (e.g., BIMS ↔ Data Management Platform)
3. Advanced Technique: Generate a Confirmation Question List
Append a prompt to have TRAE Work organize all uncertain points into a list:
Please organize all content marked "None, needs to be added", "Needs confirmation", "To be confirmed" into a list:
| Functional Module | Field/Issue | Current Mark | Content to Confirm | Confirmation Target |
Send it directly to the backend and PM, confirm all questions at once, no need to ask back and forth.
4. Advanced Technique: Generate Component Type Mapping
Please map the input method of each field to a frontend component type:
- Read-only → Text/Tag
- Manual Input → Input/InputNumber
- Manual Check (Single) → RadioGroup
- Manual Check (Multiple) → CheckboxGroup
- Manual Select (Multiple) → Transfer/MultiSelect
- System Calculated → Read-only display (with loading)
- BIMS Edit Tag → Tag + Popover Edit
Directly guides component selection, no need to think about which component to use for each field again.
5. Advanced Technique: Generate API Field Requirements
Please output the API field list required for each function based on the field matrix:
- Annotate which fields need to be returned by the backend
- Annotate which fields need to be passed as parameters by the frontend
- Annotate which fields need to be newly added by the backend ("None, needs to be added")
- Annotate linkage relationships between fields (B is only shown when A is selected)
Send it directly to the backend as input for API design, enabling parallel frontend and backend development.
5. Summary
| Item | Data |
|---|---|
| Scenario | Organizing PM's Excel requirements into a frontend dev doc |
| Number of Excel Files | 3 |
| Number of Functional Modules | 12 |
| Number of Business Scenarios | 4 |
| Total Fields | Hundreds |
| Original Process Time | About several days |
| New Process Time | About 30 minutes |
| Efficiency Gain | 200x |
| Deliverables | Page Fields + Interaction Logic Description + Dev Task Breakdown |
| Reuse Method | Apply universal template + replace Excel text |
For a small company frontend developer chewing through a PM's Excel requirements, the most torturous part isn't the complexity of the requirements themselves, but the information scattered across Excel cells — field sources differ under different scenarios, notes are hidden in corners, column widths are too narrow to see everything, and cross-file dependencies require constant switching. Previously, manually sorting this out meant that missing one note or confusing one field source would lead to another round of revisions.
The core idea of using TRAE Work: Hand over the entire process of "Excel parsing → Field matrix generation → Interaction logic sorting → Task breakdown → Missing item extraction" to AI, with the human only responsible for defining the output format and doing the final confirmation.
The most critical value isn't just saving time, but that field sources are traceable + notes are not omitted + missing items have a checklist. Previously, confusing field sources led to rework, overlooking notes led to bugs after going live, and missing fields were only discovered halfway through development when there was no API. Now, TRAE Work automatically generates a field matrix and a missing field checklist, so a quick glance before development tells you which points need confirmation from the backend.
Note
You still need to understand the requirements yourself. This Trae Work is just an aid. If you don't figure it out yourself, you won't know if it generates something wrong, and that's big trouble!