Strawberry can read an approved public directory, collect the fields you specify, and prepare a structured table for Google Sheets. The sheet should retain the page URL and collection date so every row can be checked or refreshed later.
01
Define the row before you collect the first page.
“Get all the companies” is not a usable specification.
Decide whether a row represents a company, location, person, listing, or product, then name the fields and the exact format each field should use.
02
Preserve the page that supports every row.
A clean spreadsheet becomes unreliable when nobody can tell where a number, category, or URL came from. Keep the listing page or profile URL beside the extracted facts, even when several listings appear on one page.
Strawberry can collect the visible directory data and attach the source trail needed for a later verification pass.
03
Clean the dataset before it becomes a decision.
Directories repeat listings, use inconsistent place names, and leave key fields blank.
Deduplicate on the fields that actually identify the record, standardise only where the source allows it, and flag records that need a human look.
04
Refresh the sheet without losing its collection rules.
Save the approved directory source, row definition, deduplication logic, and validation checks as a data-collection skill. A routine can revisit the public pages later and prepare a change set instead of silently overwriting the original collection.
The useful output is a reviewed dataset with clear provenance, not the largest possible row count.
AI can collect specified fields from approved public directory pages and prepare them as structured rows for a Google Sheet.
Include the entity name, relevant category or location, website when present, source URL, captured date, and an exception note for uncertain records.
Choose stable identifiers such as website and organisation name, then review potential matches before merging locations or separate listings.
Yes. Keep the source list and row rules, then run a controlled refresh that separates additions and changes from the original dataset.
Requirements vary by source and use case; review the site terms, applicable law, access rules, and the sensitivity of the information before collection.