The Scenario
You inherited a Google Sheet from the analyst who left three weeks ago. Column A has 200 news article snippets she'd been collecting for a market intelligence project — tech company names, executive quotes, product announcements, all of it unstructured. The brief you received says the next step is to build a structured entity database: which companies appear, which executives, which locations. The sheet is supposed to feed a dashboard by end of week.
The bad version:
- Copy the first snippet from column A, paste it into the TextRazor API explorer, scroll through the JSON response looking for the entity array, manually note the names and types, paste them into columns B and C, move to row 2.
- Repeat for 200 rows, making judgment calls each time about which entities matter and whether a relevance score of 0.4 is worth including.
- Realize at row 40 that you haven't been consistent about how you're formatting the type field — "ORG" vs "organization" vs "Company" — and now you have to decide whether to go back and fix rows 1–39 or live with the inconsistency.
This kind of cleanup compounds. By the time you've sorted the inconsistencies, documented what you extracted, and caught the rows where you accidentally overwrote the wrong cell, you've spent most of the day on a task that was supposed to take an hour. The dashboard still isn't done.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent that lives inside your Google Sheet. It reads the sheet, understands the column structure, and through its built-in TextRazor integration can run entity extraction across every row and write structured results back into the columns you specify. Open the sidebar and ask.
For each text in column A, use TextRazor entity extraction and write the extracted entity names and types (person/organization/place) into columns B and C, one cell per row.
What You Get
- Column B receives a comma-separated list of the top entities extracted from each row's text, ranked by relevance score.
- Column C receives the corresponding type for each entity — person, organization, or place — in the same order.
- Rows where TextRazor returns no high-confidence entities get a note in column B rather than a silent empty cell.
- The output format is consistent across all 200 rows — same delimiter, same type vocabulary, no judgment calls per row.
What If the Data Is Not Quite Ready
The snippets include HTML tags from a scrape
For each text in column A, strip any HTML tags from the text before sending it to TextRazor entity extraction, then write the extracted entity names and types into columns B and C.
Some rows are blank or contain only URLs
For each text in column A, skip blank rows and rows that contain only a URL, run TextRazor entity extraction on the remaining rows, and write entity names and types into columns B and C. Leave B and C empty for skipped rows.
You need entities from two tabs combined
Combine the texts from column A of the 'Q1 News' sheet and column A of the 'Q2 News' sheet, run TextRazor entity extraction on each, and write entity names and types into columns B and C of each respective sheet.
Full kill chain: clean, extract, frequency count
For each text in column A, remove HTML tags and trim whitespace, run TextRazor entity extraction, write entity names and types into columns B and C, and then add a 'Top Entities' summary tab that shows each unique entity name, its type, and how many rows it appeared in — sorted by frequency descending.
The more reliable pattern is to handle cleanup and extraction in the same ask — you get consistent output without a separate pass.
Try It
Get the 7-day free trial of SheetXAI and open any Google Sheet where column A holds a batch of text snippets waiting to be structured. Ask SheetXAI to run TextRazor entity extraction and write the results directly into the adjacent columns. Then check out the IAB classification spoke if your next step is content categorization.