LearnAI
Новинки AI · Код без программиста · Lesson 9/10
← To the catalog
Task

Auto-table: let it calculate the rule

Order table automation. Return JSON {"build":"what we automate","input_data":"description of columns and 2-3 example lines","rule":"processing rule","output":"what the result should be","done_when":"how you check"}. The examples contain only fictitious data.

Unlock access to submit solutions for instant AI review.
Unload routine in tables
💡 A piece of theory

Table automation is when the routine with rows and columns is done not by a person, but by a rule. The scheme is the same: input data - processing rule - finished result. AI needs to know what your data looks like (which columns, what’s in them), which rule to apply (filter, calculate, categorize, remove duplicates) and what the result should be (new column, separate sheet, summary). Describe the columns with words (“Column A is date, B is amount, C is category”), formulate the rule as an order (“Add the amounts for each category”) and specify the type of result. Attach 2-3 example lines - but fictitious ones, without personal and secret data: from them the AI ​​will understand the format more accurately than from any description, and will not guess the structure incorrectly. Ask not to touch the source data, but to write the result separately - then an error in the rule will not erase your table. A common mistake is to describe the desired outcome, but not show how the input is structured: the AI ​​will start thinking up the columns, and the rule will miss.

How is it assessed · passing 70

  • 1Valid JSON without text around it25%
  • 2There are fields build, input_data, rule, output, done_when30%
  • 3Shown is an input with examples and an order-rule45%
Your promptClaude ⌄
🔒

Unlock access to submit solutions for instant AI review.