Multi-line editor for the cleanup logic
IFERROR-wrapped VALUE-wrapped DATEVALUE-wrapped INDEX gets long fast. The multi-line editor renders it as readable code instead of a 250-character ribbon you have to parse character by character.
Use case
The export ran fine.Now make it useful.
Numbers stored as text, dates the wrong way around, blanks where there should be zeros, totals that don't tie. The last mile between an ERP export and a report nobody questions is where Formula Foundry lives — the multi-line editor, the visual builder, and an AI assistant that explains why the SUMIFS keeps returning zero.
Trial days only count when you actually open the add-on. No credit card.
How it works
From raw export to clean dataset.
Spot the columns that are wrong
Color-coded editor + AI explain make it obvious when a column you thought was numeric is actually text — including the silent kind that looks right until SUMIFS returns zero.
Build the cleanup formula
Visual builder for the conditional logic (IFERROR / VALUE / DATEVALUE / TRIM cascades). Real-time preview so you know the formula works on row 1 before you fill 50,000 cells.
Save the pattern as a snippet
Same export structure next month. Same cleanup formula. Save the working version once; insert it on next month's export instead of rebuilding it.
What makes the last mile faster
The toolkit for the messy middle.
Visual builder for nested conditionals
Build the cleanup formula as a tree — pick the function, fill the arguments, see real-time output on row 1 before you fill the column. Stops the 'tweak and pray' loop.
AI to debug the silent failures
When SUMIFS returns zero on what looks like a numeric column, ask the AI to explain why. It catches the 'numbers are stored as text', 'date is in the wrong locale', and 'whitespace in the lookup key' bugs that every clean-up job hits.
Snippets for next month's export
Same export, same cleanup. Save the working formula once, insert it next month with one click. The team library means you're not the only person who can run the cleanup.
See it in action
Building the cleanup formula in the visual builder.
Got questions?
ERP cleanup FAQs
- The cleanup formulas run inside Sheets / Excel — your data never leaves your session. The AI assistant only sees what you paste into it (a single formula, a sample row), and only when you click 'explain' or 'suggest'. We don't read or upload the rest of your dataset.
- Formula Foundry works on whatever spreadsheet you're looking at — the cleanup is on the data after export, not on a connector. If you can drop the CSV into Sheets or Excel, you can use the editor, the builder, and the AI to clean it. The patterns generalize across every ERP we've seen.
- Phase 1 of the AI is one cell at a time, by design — auto-cleaning a 50k-row export without review is how teams ship reports with quietly wrong totals. The AI suggests the cleanup formula; you review it on row 1; then you fill the column. The review step is where the bugs get caught.
- If your team already has a pandas script and the discipline to maintain it, keep doing that. Formula Foundry is for the much larger group who don't — the analyst doing the cleanup in their browser, the ops lead who needs the report by 9am. We meet you in the spreadsheet you already have.
Stop fighting the export
Clean the next export in an afternoon.
Install Formula Foundry, drop in your messy CSV, and turn it into a dataset the rest of the org can actually use.