How to Use AI to Generate Weekly Reports and Other Rich Text Content: A Practical Guide

Imagine having an intelligent assistant that can quickly produce document content tailored to a specific topic or format. This is the core idea behind AI-generated collaborative rich text in Fenxiang (纷享销客). By configuring a prompt template, users can guide the AI assistant to produce high-quality rich text. Users insert variables as prompt inputs and craft precise prompts to control the output. Rather than being a simple add-on, the feature works like a smart writing partner that helps teams turn business data into readable, structured documents with less manual effort.

The feature supports several types of variables that give users powerful generation capabilities. Primary object and its fields: reference field data from the current object. Other objects and their fields: reference field data from related objects. Scenario-related variables: automatically populate dynamic variables tied to the current context. Custom functions: write APL functions in Java or Groovy syntax to handle complex logic. Together, these variable types allow the AI to draw on both the record being edited and the broader business data around it, so the generated text reflects real, context-specific information.

The feature is currently applicable to a range of scenarios. The most prominent example is automatically generating sales activity summaries, which helps employees complete daily, weekly, and monthly reports with minimal effort. Because the template can pull in related records and business fields, the same approach can be adapted to other report-writing and content-drafting tasks where repetitive, data-driven text needs to be produced consistently.

Configuring a prompt template follows an overall workflow of three stages: basic settings, prompt content, and testing with preview. Each stage builds on the previous one, so completing them in order ensures the final output matches your requirements. The sections below walk through each stage in detail.

Start by selecting the generation language (for example, Simplified Chinese), the stylistic preference, the sender role, and the content length. Then choose an AI model that fits the task. Clear basic settings establish the tone and scope of the output, while the model choice affects quality and speed, so it is worth matching the model to the complexity of the report.

The prompt content is assembled from several sources. Object fields: select data fields from the object as inputs. Scenario variables: use special variables specific to the current user scenario; for example, when writing a marketing email, the email subject is a scenario variable. Custom functions: APL functions based on Java or Groovy, which require selecting a namespace and business object and then binding the function to the business scenario. Prompt: write explicit instructions so the AI understands your intent. Combining these sources lets you build a template that pulls in the right data and tells the AI exactly how to structure the result.

Validate the configuration with test data before rolling it out. In the preview test area, click the generate button, review the output, and adjust the template until the result meets your expectations. Iterating on the prompt at this stage is far cheaper than discovering problems after users begin relying on the template.

Consider a scenario where the AI retrieves the current user's sales records within the log's time period and automatically generates a summary of this period's work. First, enable the AI generation feature on the collaborative rich text field named 'Work Summary for This Period' in the log object. Next, under 'Select a prompt template,' choose 'Create new template,' give it a name and description, and click Next. Then select your preferred model, set the generation language to Simplified Chinese, and leave style and length unrestricted. In the prompt content section, configure the weekly report template in detail, making sure it covers all the necessary information. Finally, click 'Generate Preview' in the upper right corner and select test data. Clicking 'Generate Preview' again shows both the prompt preview and the generated text, completing the example of automatically summarizing sales records into a work summary.

After saving the prompt template, return to the field editing page and bind the newly created template to the field, then save. Next, go to the frontend, open the log object, and click the New button. An AI icon will appear under the 'Work Summary for This Period' field. Click the icon and select the bound prompt template, which displays a preview of the prompt for the current data. Click the Generate button, and the AI produces rich text based on the template and the log's time range. If the result meets your needs, click 'Insert Below' to automatically populate the 'Work Summary for This Period' field. If it does not, use the regenerate button in the lower left corner to try again.

The AI-generated collaborative rich text feature gives organizations an intelligent and efficient way to produce field content. By configuring prompt templates, users can readily guide the AI assistant to generate high-quality rich text tailored to their actual needs. As you continue to explore the feature, you will find more ways to apply it, bringing added convenience and value to everyday work.