AI financial document analysis with the source for every figure
Finance teams spend hours finding one number in a 90-page report or a workbook with twelve tabs. ChatDocument answers from your own filings, board packs and spreadsheets, cites the page or the sheet and cell range behind each figure, and tells you plainly when the number is not in the files.
For FP&A, controllers, analysts and finance operations
- Annual and quarterly reports, management accounts and board packs
- Budgets, forecasts and models in Excel (XLSX) or CSV
- Audit letters, loan agreements and covenant certificates
- Investor presentations in PowerPoint (PPTX), cited by slide
- Expense, travel and procurement policies the team answers questions about
Questions that used to mean opening five files
- What was gross margin in Q2 and Q3, and which page reports it?
- How is the leverage covenant defined in the credit agreement, and what is the threshold?
- Which cost centers went over budget in the latest management accounts?
- What does the auditor's letter say about revenue recognition?
- How did the guidance in the latest investor deck change from the previous one?
Cited answers on reports, decks and workbooks
Create a space per entity, per reporting period or per deal, and upload files or connect a finance folder from Google Drive, SharePoint and OneDrive or Dropbox so new versions sync on a schedule.
- Citations that fit the format: a page for PDFs, a slide for decks, a sheet and cell range for spreadsheets. Click to open the source at the highlighted passage.
- Compare periods: put two quarterly reports or two versions of a forecast side by side and see what changed, cited on both sides. More on compare documents ai.
- Extract to table: pull figures from a set of reports into columns such as Period, Revenue, Gross margin, Operating expenses, Net debt, Covenant ratio and Auditor. Each cell links to its page, ready for XLSX or CSV. See ai pdf to excel.
- Summaries: a one-page summary of a long annual report, or a custom template that always covers results, outlook and risks, each line cited. See summarize documents with ai.
- Sharing: give the CFO asker access to a space, save threads per close, share one answer by link.
No guessing on numbers
Each quoted passage is checked in code to exist in the document and contain the quoted text before the answer is shown. When a figure is not in the selected files, ChatDocument says so and shows the closest passages it checked, rather than estimating from general knowledge.
Plans for one analyst up to a finance department
- Starter, 49 USD per month or 24 USD per month billed yearly (288 USD a year): 1 seat, 500 questions, 2,000 new pages, 50 MB per file. Fits one analyst.
- Team, 149 USD per month or 74 USD per month billed yearly (888 USD a year): 5 seats, 2,000 pooled questions, 8,000 pages, 100 MB per file, 15 spaces.
- Business, 499 USD per month or 249 USD per month billed yearly (2,988 USD a year): 15 seats, 7,000 pooled questions, 30,000 pages, 200 MB per file, audit log and API for pulling figures into your own tools.
- Enterprise, 1,490 USD per month or 745 USD per month billed yearly (8,940 USD a year): 50 seats, 100,000 pages, 500 MB per file, SSO (SAML), SLA 99.9 and invoice billing.
Compare all limits on chat with pdf pricing. Save 50% with yearly billing.
What ChatDocument is not
ChatDocument reports what your documents say. It is not investment advice, it does not audit your numbers, and it does not build forecasts. It does not pull market data from outside your files. Check each figure against the cited page before it goes into a board pack. For due diligence work on a target company, see due diligence ai.
FAQ
Does it work with Excel files?
Yes. XLSX and CSV files are indexed and answers cite the sheet and cell range they come from.
Can it calculate ratios from the figures?
It answers from the figures in your documents and cites them. Treat any arithmetic in an answer as something to check against the cited cells.
Are our financial documents used to train models?
No. They are encrypted in transit and at rest and never used to train models. Details are on the security page.
Find the figure and its source in one question
Create your account, upload this quarter's reports and ask the question you would otherwise answer with Ctrl+F. No card needed for the demo.