AI PDF to Excel extraction with a source page behind every cell
Name the columns you need, point ChatDocument at a set of documents and get a table with one row per document. Party, term, renewal date, liability cap or any field you define: each value links to the page it was read from, and the whole table exports to XLSX or CSV for the spreadsheet your team already uses.
A register built from the documents themselves
Copying terms out of fifty contracts by hand takes a day and leaves no trail. Extract to table reads each document, fills your columns and keeps a link from every cell to the passage that supports it. A cell the document does not answer stays empty and is marked as not found, rather than being filled with a plausible value. You can open any cell to see the highlighted passage, correct it if your reading differs, and keep the table in the space for the next person.
- Custom columns in plain English, with an optional format such as date, amount or yes or no
- Works on PDF, Word, scanned PDFs read with OCR and other supported files
- Empty cells marked as not found in the document
Columns a legal team sets up in a minute
| Document | Party | Term | Renewal date | Liability cap |
|---|---|---|---|---|
| Master services agreement | counterparty name, p. 1 | 36 months, p. 3 | automatic, notice 60 days, p. 9 | 12 months of fees, p. 14 |
| Data processing addendum | counterparty name, p. 1 | runs with the main agreement, p. 2 | not found | uncapped for data breach, p. 6 |
Every value in the table links to its page, so a reviewer checks the reading instead of redoing it.
Five steps to an exported spreadsheet
- 1Open a space and select the documents, or a whole folder synced from a connected drive.
- 2Add columns: a name and a short instruction, for instance Liability cap, the maximum amount either party pays, as written.
- 3Run the extraction. ChatDocument reads each document, fills each column and checks every cited passage against the stored text.
- 4Review the table, open any cell to see its source, and save the column set to reuse on the next batch.
- 5Export to XLSX or CSV. The export keeps a page reference next to each value.
| Party | Term | Renewal date | Liability cap |
|---|---|---|---|
| Northgate Analyticsp. 2 | 24 monthsp. 3 | 1 Mar 2028p. 4 | 125% of feesp. 5 |
| Harbourline Logisticsp. 5 | 12 monthsp. 6 | 30 Jun 2027p. 7 | 100% of feesp. 8 |
| Brightford Instrumentsp. 8 | 36 monthsp. 9 | 15 Jan 2029p. 10 | 1.0 M USDp. 11 |
Extraction table with a cell opened
A table of twelve agreements with columns for party, term, renewal date and liability cap; one cell is open and the contract shows page 14 with the cap clause highlighted.
Documents per run and how rows are counted
Each document row in an extraction counts as 1 question, whatever the number of columns. The number of documents in one run depends on the plan: 10 on Starter, 50 on Team, 250 on Business and 1,000 on Enterprise. Rows draw from the monthly questions of your plan (500, 2,000 pooled, 7,000 pooled and 20,000 pooled) and from the daily limit. Pages count once per billing period. See chat with PDF pricing.
Who uses extraction most
- AI contract review: a renewal and notice register for every active agreement.
- AI financial document analysis: revenue, margin and headcount pulled from a stack of quarterly reports.
- Due diligence AI: change of control, assignment and exclusivity clauses across a data room.
- AI legal document analysis: governing law and dispute terms across a client's contract portfolio.
FAQ
How many columns can one extraction have
As many as the table needs. The row of each document counts as 1 question regardless of the column count.
What happens when a document does not contain a value
The cell stays empty and is marked as not found. ChatDocument does not guess a value from other documents or general knowledge.
Can I export with the page references
Yes. XLSX and CSV exports keep a page reference for each value, so the spreadsheet stays traceable after it leaves ChatDocument.
Can I reuse the same columns next month
Yes. Save the column set in the space and run it on new documents whenever they arrive.
Turn a folder of contracts into a register
Create your account, add your documents and define the first columns.