CIRAS | Iowa State University
AI Fluency / Cycle 02
Organize Information
Experiment: The Same Information in Different File Formats
Explore how files store information, compare their strengths and limitations, and choose a useful structure for the job.
Start Here
Download the example files once, extract the ZIP, and use that folder for both experiments.
Windows: right-click the downloaded ZIP and choose Extract All. On a Mac, double-click the ZIP.
Guided Experiment: Explore the Same Information
15 minutes · Follow along with the instructor
- Inspect: Open the files for order
00481. Compare the plain-text message, Word document, two PDFs, Excel workbook, CSV, and JSON example. - Test: Try selecting text in the two PDFs; examine cells and sheets in Excel; look at rows in CSV and nested values in JSON. Notice the difference between what a file looks like and how its information is stored.
- Decide: Discuss which format best serves a person reading the confirmation, a receiving clerk sorting delivery lines, and software exchanging records.
- Demonstrate: Organize the original supplier message as one record per delivery line. Decide on useful fields, choose an output format, and verify the facts against the original.
Optional instructor prompt
Use this only if helpful. The instructor can demonstrate the same decisions without AI.
Using the attached supplier message for PO 00481, recommend what one record should represent and five or six useful fields for a receiving log. Then show a short table. Preserve the leading zeros in the order number and both delivery dates. Do not invent facts. Finish by recommending Excel, CSV, or JSON for this task and explain why in one sentence.
Compare the File Types
Use this as a guide while opening the files. Same facts do not mean the same structure or capabilities.
| Format | What to notice | Typical tradeoff |
|---|---|---|
| TXT | Plain characters without built-in tables or formatting. | Simple and portable; relationships must be inferred. |
| Word (.docx) | Headings, paragraphs, and actual document tables. | Readable and editable; harder to sort many records. |
| Text-based PDF | Fixed page appearance; text can usually be selected. | Easy to share; extracting table structure may be imperfect. |
| Scanned PDF | Looks similar but may contain only an image. | Text extraction may need OCR or vision and checking. |
| Excel (.xlsx) | Cells, sheets, types, formatting, and formulas. | Good for editing and analysis; workbook features may not export. |
| CSV | Plain-text rows and columns. | Good for flat data; no formulas or sheets, and IDs such as 00481 can be misread. |
| JSON | Named fields and nested lists of records. | Useful for software exchange; less convenient for casual editing. |
A format isn’t automatically better because it is more structured. Choose it for the work you need to do.
Independent Experiment: Make the Decisions
20 minutes · Work on your own
Part 1. Compare the files first
Explore the versions of order 00481 in the extracted folder. Compare at least four formats, including one PDF and one table-based format.
- Identify one thing a format preserves or makes easier, and one limitation.
- Choose a preferred format for reading a confirmation and another for tracking delivery lines. Explain your choices.
Optional comparison prompt
I am comparing TXT, Word, PDF, Excel, CSV, and JSON versions of supplier order information. Make a compact comparison of what each format can preserve and what can be difficult to extract or use. Distinguish scanned from text-based PDFs. Use any files I attach as evidence; do not claim to have examined a file I have not provided. End with two questions that would help me choose a format for my task.
Part 2. Organize a different supplier update
Only after the comparison: Open C2_Independent_Supplier_Update_00482.txt or its Word version. This is a new order with one quantity not yet confirmed.
- Decide whether one record should represent an order or a delivery line.
- Choose five or six useful fields, organize the information, and select the format you would use for receiving.
- Compare your result with the original message. Preserve missing information as unknown, not zero, and retain the restriction on finalizing the schedule.
Optional prompts: choose one approach
Option A: Generate a draft to critique
Using the attached supplier update for PO 00482, recommend whether to use one row per order or per delivery line. Choose five or six fields and produce a compact table. Keep unknown values unknown and preserve any instruction that limits what receiving can finalize. Suggest a file format and one way I should verify your result.
Option B: Make the decisions yourself
I will decide how to structure the attached supplier update for PO 00482. Ask me three short questions, one at a time, to help me choose record granularity, fields, and output format. After I answer, build only the structure I approved. Then identify information that is missing or may have been lost.
Compare and Reflect
5 minutes · Group discussion
- Which file-type difference was most surprising when you opened the examples?
- How did the purpose of the work influence your choice of fields and format?
- What could be lost or silently changed when AI converts one representation into another?

