Community Discussion · Company Watch

WorkBuddy turns dates into text when summarizing Excel; has anyone seen this?

KevinZhao_FinKevinZhao_FinAug 312026/08/31 72 views

I recently read a popular science article about AI data analysis tools claiming they can automatically clean, classify, and generate charts. That got my hands itching. I've been using WorkBuddy for about a month, mainly for research report summaries, which felt okay. This time, I tried merging client payment details and several PDF invoices into a single analysis table, but I got stuck right away.

I threw both the Excel and PDF files in together and asked it to summarize by supplier, month, and amount. It ran for about ten minutes. The table came out, but the date column turned entirely into text, so I couldn't filter by month. There were also several #VALUE! errors in the amount column. Even worse, one row split "2026-08-20" into two columns: "2026" and "08-20". I did a quick check, and the pivot table was completely broken.

I tried re-uploading and clicked the option to "strictly match output template," but the formatting was still messy. Did I fail to clarify field rules initially, or does this mixed PDF-and-Excel scenario inherently require breaking down the steps? Previously, when I used the knowledge base feature, it seemed stable with templates, but this time the data is in Excel, and I'm unsure where to place the template.

Has anyone else experienced WorkBuddy converting dates and amounts to text? Did I set something up wrong, or is it actually better suited for summaries rather than running these kinds of analysis tables directly?

1 replies

?
Ctrl + Enter to reply
Mo Mo
Mo MoAug 31

I wanted to mention this in your previous post about "scheduled tasks extracting dates from scanned documents returning empty results." Mixing PDFs and Excel files really messes with WorkBuddy. I suggest converting PDFs to plain text or CSV before merging them. Don't expect it to auto-clean formats; I tried splitting the steps once, and efficiency actually improved.