Platform

Document intelligence with the evidence still attached

Pramanis helps Nepal CA firms bring financial statements, scanned PDFs, audit supporting documents, client-provided schedules, invoices and other supporting records closer to the workbook question they need to review. The aim is not to turn a folder into an automatic conclusion. It is to make the original source context easier to retrieve, inspect and discuss.

Explore evidence-linked analysis

The buyer problem: evidence is usually scattered

A material balance can be explained across a stock schedule, valuation working, supplier invoice, delivery record and notes sent by the client. Some items arrive as structured tables; others are unstructured PDFs, photographs or scans. When those records remain in separate folders, staff can spend too long locating the right page and reviewers may receive a conclusion without the document context that supports it. A stronger workflow keeps the question, the workbook movement and the relevant source record connected.

Extract for retrieval, then return to the original

Document extraction can make text and tables in uploaded files easier to search and organize. OCR can help where a PDF is a scan rather than selectable text. Pramanis can retain source-page references so a user can move from an extracted passage or finding back to the page it came from. This matters because extracted text is a working aid, not the document itself. A professional should read the original page, note its date and context, and decide what weight it carries.

Synthetic workflow — inventory support for a Nepal CA firm

In a synthetic audit engagement, a workbook shows a material increase in inventory. The team traces the movement to the stock schedule and brings together the client’s valuation document, selected purchase invoices and other supporting records. Pramanis helps organize those materials beside the workbook finding and point the reviewer to relevant pages. The reviewer checks whether the documents genuinely support the balance, whether the quantities and dates align, and whether further procedures or a client query are needed. The platform organizes evidence; it does not decide that evidence is sufficient or appropriate.

Keep the link between a finding and its support

A useful evidence trail does more than attach a PDF. It lets the team relate a workbook observation to a particular source document and page, while keeping the surrounding record available for review. That can help with financial statements, audit files, invoices, reconciliations and client schedules where a short extracted phrase could otherwise be misleading. Draft working-paper notes and reviewer questions should continue to point to their source, so a manager or partner can understand what was inspected and what remains unresolved.

Extraction has limits that matter in practice

Poor-quality scans, skewed pages, faint text, handwritten documents and complex tables can lead to incomplete or uncertain extraction. A file may have missing pages, an unreadable stamp, columns that do not preserve their meaning, or a total that is not clearly associated with the right row. These are signals to inspect the original, not reasons to assume the system has recovered the intended meaning. Teams should confirm completeness, compare important values with the source and record or resolve uncertainty before using extracted material in professional work.

Secure processing and clear source boundaries

Client files should be processed only through the firm’s approved upload and access process. Uploaded document text is treated as untrusted evidence, not as trusted AI instructions: language inside a file must not change the task, permissions or trusted source rules. That distinction is especially important when documents contain copied correspondence, prompts or unexpected text. Access, scope and the professional purpose of the review remain matters for the engagement team. Pramanis can help retrieve material, but the team remains responsible for confidentiality, authorization and review.

Professional responsibility stays with the firm

Document intelligence can reduce time spent finding records and make a review path easier to follow. It cannot independently determine whether a document is authentic, complete, relevant, sufficient or appropriate for an audit, tax or accounting conclusion. The responsible professional evaluates the facts, applies skepticism, performs required procedures and approves the final work. Where source material is missing or unclear, the correct next step may be a further request, alternative evidence or an explicit unresolved matter—not a more confident summary.

Continue exploring

Workbook intelligence Report drafting Audit & Assurance

Next step

Bring evidence closer to the review

Explore how an evidence-linked workflow can be discussed with synthetic material before confidential client files enter a new process.