DOCUMENT AI

Turn contracts, reports and PDFs into live, queryable data

Upload a PDF, ask a question, and get the answer with the page it came from. Extract the fields that matter into a dataset and join them to your SQL, NoSQL and API data, on the same governed layer.

PDF, Word, Excel, CSV and text · up to 100 MB per file · source-referenced answers

Document AI with 14 vendor contracts indexed: the question which of these auto-renew and what is the notice period is answered from Acme_MSA_2026.pdf, page 11, section 8.2, with the cited clause highlighted. Answers across documents, page-level citations, permissions inherited

Chat with your documents. Then query them like a database.

UPLOAD AND CONNECT

Upload once, or keep them in sync.

Drag a folder in to start. Connect Drive, DocuSign, SharePoint or OneDrive and the set stays current without anyone re-uploading.

  • Drag-and-drop upload or file picker, up to 100 MB per file
  • PDF, Word, Excel, text, CSV and more
  • Continuous sync from Google Drive, DocuSign, SharePoint and OneDrive
Connected document sources
DRIVELegal / Contractssyncing
DOCUSIGNExecuted agreementssyncing
SHAREPOINTPolicy librarysyncing
UPLOADQ3 board deck.pdfindexed
MANUALRe-upload on changenot needed
ASK, WITH THE RECEIPT

Every answer carries its source.

The page, paragraph or section comes back with the answer, so a reviewer checks it instead of trusting it. That is what makes it usable in legal, finance and compliance work.

  • Ask in plain English across long, dense documents
  • Jump straight to the cited page, paragraph or section
  • Works on 15 to 100+ page documents without manual scanning
Grounded answer
“What is the liability cap in the Northwind agreement?”
Aggregate liability is capped at fees paid in the twelve months preceding the claim, excluding breaches of confidentiality.
Northwind_MSA_v3.pdf · page 24, section 12.4
Carve-out for indemnified IP claims, uncapped.
Northwind_MSA_v3.pdf · page 25, section 12.6
SUMMARIES AND MULTI-DOC Q&A

Compare a folder, not a file.

Ask one question across a set of documents and see which one each answer came from. Reading an RFP pack becomes querying it.

  • Executive summaries of long documents in seconds
  • Key terms, risks, dates and figures pulled automatically
  • Query many documents in a single question
  • Built for RFPs, vendor comparisons, policy packs and contract folders
One question, 3 contracts
Acme_MSA60 days notice, auto-renew
Northwind_MSA90 days notice, auto-renew
Contoso_MSAno auto-renew clause
FLAGContoso deviates from standardsurfaced
DOCUMENTS AS DATASETS

Documents become datasets.

The extracted fields are written back as a structured dataset, so contract terms sit on a dashboard next to billing and usage data and can be alerted on like anything else.

  • Extract amounts, dates, entities and clauses as fields
  • Store the results as structured Knowi datasets
  • Join document data with SQL, NoSQL and API sources
  • Use the result in dashboards, alerts and embedded analytics
See the unified data layer
contracts_extracted · joined
renewal_datefrom PDF, page 3
annual_valuefrom PDF, exhibit A
notice_daysfrom PDF, section 8
JOINpostgres.invoiceson account_id
JOINmongodb.product_usageon account_id
ALERTRenewal in 90d & usage < 40%armed
GOVERNED DOCUMENT AI

Your permissions still apply.

Document AI runs inside Knowi and inherits the access model you already run. Cloud, on-premises or hybrid.

  • Runs inside Knowi, without sending content to third-party services
  • Cloud, on-premise or hybrid deployment
  • Respects your existing permissions, governance and security model
Security and deployment
Access, unchanged
LEGALContracts folderfull
FINANCECommercial terms onlyscoped
SALESOwn accounts onlyrow-level
EXTERNALThird-party model callnone
THE FLOW

From a folder of PDFs to a live dashboard.

Four steps. The last one is the reason for the first three.

01 · INGEST

Upload or sync

Drag files in, or connect Drive, DocuSign, SharePoint or OneDrive for continuous sync.

02 · ASK

Question in plain English

Across one document or the whole set, with the source location returned alongside every answer.

03 · EXTRACT

Fields into a dataset

Amounts, dates, entities and clauses become structured columns you can query.

04 · JOIN

Blend and monitor

Join to SQL, NoSQL and API data, put it on a dashboard, and alert when something changes.

WHO USES IT

One question across every contract.

LEGAL & CONTRACTS

Read the clause, not the contract

Pull renewal dates, termination clauses, obligations and key terms without manual review. Summarise changes for stakeholders and compare agreements across vendors, regions or versions to spot deviations at scale.

“Which contracts deviate from our standard termination language?”
FINANCE & REPORTING

Summarise 100 pages in seconds

Condense annual reports, board decks and financial statements, surface key figures and trends, and turn note disclosures and narrative sections into structured fields you can analyse.

“Summarise the risk factors that changed since last year's 10-K.”
COMPLIANCE & POLICY

Evidence, with a citation attached

Ask across policy documents, SLAs and regulatory guidance and get source-backed answers. Q&A interactions flow into structured logs that support audits and evidence collection.

“What is our stated breach notification window, and where is it documented?”
SALES & CUSTOMER SUCCESS

Know what you promised

Pull commitments, SLAs, pricing terms and obligations from customer contracts, then use that contract data to power renewal planning, risk dashboards and proactive account management.

“What did we commit to this customer on uptime and support response?”
WALKTHROUGH

Document AI in the product.

Uploading, asking, and checking the source.

FAQ

Questions about Document AI.

Can I chat with a PDF in Knowi?

Yes. Upload or sync a PDF and ask questions of it in plain English. Every answer comes back with the page, paragraph or section it was drawn from, so it can be checked rather than trusted.

What is chat with documents?

Chat with documents means asking questions of a file, or a folder of files, in conversation instead of reading or keyword-searching them. In Knowi it also writes the extracted fields back as a structured dataset you can join to your databases and put on a dashboard.

What is Document AI in Knowi?

Document AI lets you upload documents such as PDFs, Word files and spreadsheets, ask questions in plain English, and get answers with source references, inside Knowi and alongside your database data.

Which file types are supported?

Common business formats including PDF, DOC and DOCX, XLS and XLSX, CSV and TXT, up to 100 MB per file.

Can I query multiple documents at once?

Yes. Upload or sync multiple documents and ask a question across all of them. Document AI returns the most relevant answer and shows which document it came from.

Is my document data secure?

Document AI runs within Knowi's environment and can be deployed in the cloud or on premise. Your existing permissions and governance controls stay in force.

Can I combine document insights with my other data?

Yes. Extract key fields from documents, save them as datasets in Knowi, then join them with SQL, NoSQL and API data for dashboards, alerts and AI analytics.

How accurate are the answers?

Every answer carries the exact source location inside the document, so a reviewer can jump to the page or paragraph and confirm it rather than taking it on trust.

Can documents stay in sync automatically?

Yes. Connect sources such as Google Drive, DocuSign, SharePoint and OneDrive for continuous sync, so new and updated documents flow in without a manual upload.

How do I enable Document AI?

If Document AI is not yet in your Knowi sidebar, contact your Knowi account manager or support to activate it.

Can Knowi be used for contract analysis?

Yes. Upload a set of contracts, ask questions across all of them at once, and extract the fields you care about, such as renewal date, term, notice period, liability cap and payment terms. Those extracted fields become a dataset you can chart, filter and alert on like any other.

How does AI contract review work in Knowi?

Documents are ingested and indexed, then queried in natural language with the answer linked back to the passage it came from. Because the extracted fields land in a dataset, review stops being one document at a time and becomes a view across the whole portfolio.

Can Knowi extract data from invoices and PDFs?

Yes. Document data extraction pulls structured fields out of PDFs, invoices and scanned files, and those fields can then be joined to the SQL, NoSQL and API data already in Knowi, so unstructured documents end up on the same dashboard as everything else.

Put your contracts on a dashboard.

Upload a folder, ask a question, and join what comes back to the rest of your data.

Source-referenced answers · documents as datasets · your permissions