From GUI to LUI: How AI‑Powered Copilot Transforms Report Development

Why Report Development Needs AI‑Driven Change

In the AI era, software development is being reshaped by intent understanding, code generation, debugging assistance, and automated test creation. Report development, however, remains a neglected island because it combines data extraction, business logic transformation, visualization, and interaction design, making it hard to describe with linear code.

From GUI to LUI: The Evolution

The launch of Ruankian Report Copilot marks a shift from a graphical user interface (GUI), where users click icons and menus, to a language user interface (LUI), where users issue natural‑language commands. This reverses the traditional model of "people adapting to tools" to "tools understanding people".

Examples of LUI commands include:

Generate a report: 生成报表 左表头 地区、城市 上表头 产品类别

Rank within a group: C3 组内降序排名

Apply conditional formatting: 选中所有金额小于 1000 的单元格 红色加粗

These commands replace menu navigation and complex formula memorization.

Ruankian Copilot in Practice

Data Loading

SQL source:

SELECT 货主地区, 货主城市, 雇员 ID, 订单金额 FROM 订单

File source (one‑line command):

加载文件 "F:/ 订单信息.xlsx"

Report Creation

One command can produce a multi‑level grouped cross‑tab report:

生成报表 左表头 货主地区，货主城市 上表头 雇员 ID 交叉格 订单金额求和

Style Adjustment

Complex formatting can be expressed in natural language, e.g.:

白色 背景 浅蓝色 居中 加粗

背景色 淡灰色 居中

左对齐 格式 "￥0.00"

Conditional formatting example:

小于 1000 红色 加粗 大于 10000 暗绿 下划线 否则 黑色

B3 小于等于 5 或者 C3 小于 5000 背景 红色 斜体 D3 小于等于 3 背景 碧绿 下划线 否则 背景 柠檬黄

Advanced Calculations

Previously difficult tasks such as irregular grouping, cross‑row ranking, and year‑over‑year calculations are now one‑line commands, illustrated by the following screenshots:

Examples:

分段扩展 横向 订购日期 列表 2022-12-25,2023-10-1,2024-5-1

C3 降序排名 列内

Batch Selection

After a report is completed, bulk modifications can be performed with simple Chinese commands, e.g.:

选中 字体 黑体

仿宋 加粗

选中 数据区 高度 大于 10

背景色 橙色

These two sentences replace minutes of manual work.

Real‑Time Help

The Copilot parses Chinese commands with an internal rule engine rather than a large language model, making it lightweight, efficient, and low‑cost. The interface provides instant help: typing a keyword shows the full command template and required parameters, guiding users to supply precise arguments.

Why a Rule Engine Instead of a Generic LLM?

Report generation demands 100% precision, not probabilistic "maybe correct" outputs. A rule‑engine‑based Copilot offers four key advantages:

No hallucinations : Every output is deterministic; unrecognizable commands trigger a clear error instead of fabricated results.

Low risk, private deployment : Runs entirely on‑premises, keeping data inside firewalls—crucial for finance, government, and defense.

Low latency : Local parsing and rendering happen in milliseconds, delivering an "instant answer" experience.

Low cost : The lightweight engine runs on ordinary servers without GPUs or token‑based pricing, enabling near‑zero marginal cost per user.

LUI Meets LLM: Combining Flexibility and Precision

Because the rule engine expects relatively structured language, casual phrasing can fail (e.g., "B2 C2 相差小于 100 时为黄色，大于 120 时为粉色，其他为红色"). An LLM can act as a "smart front‑desk" that normalizes free‑form input into the rule‑engine‑compatible syntax, after which the rule engine executes the precise command.

Supported LLMs include Deepseek via its API.

Example of LLM‑assisted translation:

B2 C2 相差小于 100 时为黄色，大于 120 时为粉色，其他为红色

Commercial Strategy: Scaling Without Extra Cost

The lightweight architecture means the Copilot runs without GPUs, token budgets, or even internet connectivity, and is included in the lowest‑price edition of Ruankian Report. This "volume‑without‑price‑increase" model lets teams of any size adopt the LUI paradigm at minimal expense.

Conclusion

Transitioning from GUI to LUI fundamentally changes how developers interact with reporting tools: from clicking menus to stating intent. Ruankian Report Copilot demonstrates that a rule‑engine‑driven LUI can deliver precise, low‑cost, private AI assistance, giving early adopters a decisive advantage in the AI‑driven data‑analysis race.