English
Contact us

Answers you can
verify.

Ask anything across your documents and databases.

DataFact turns enterprise data into clear, traceable answers — with the source behind every result.

1.5M+ users · 15,000+ organizations

Documents + Databases · Traceable answers · Enterprise-ready

( FIG. 01 — EVIDENCE )
An answer about a margin drop with its two sources: a highlighted database row and a boxed figure on page 12 of a PDF report.

The gap

01

AI can give you an answer.
The hard part is knowing if it's right.

Most AI tools stop at the answer. DataFact keeps the evidence attached, so your team can see where every result came from and verify it without repeating the analysis by hand.

  • 01

    You get an answer, but no source

    If the result cannot be traced, someone still has to check it manually.

  • 02

    Documents lose their structure

    Tables, merged cells, figures, and complex layouts often break when documents are treated as plain text.

  • 03

    Documents and databases stay disconnected

    Important business questions rarely live in only one place. DataFact lets you work across both.

Connect · Understand · Analyze · Verify

02

One question.
All your data.

Connect documents and databases to DataFact and ask questions in natural language. DataFact finds the relevant information, analyzes it, cross-checks the sources, and returns an answer you can verify.

  1. INV-2024-081MAR 15TOTAL ABCDEREGIONUNITSPRICEAMT12345678NORTH1282.4307EAST863.1267SOUTH2101.8378NW572.2125SE1492.6387CENT922.9267WEST843.5294TOTAL1,758Q3 INV-2024-081MAR 15TOTALDOC ABCDEREGIONUNITSPRICEAMT12345678NORTH1282.4307EAST863.1267SOUTH2101.8378NW572.2125SE1492.6387CENT922.9267TOTAL1,758Q3WEST843.5294TABLE DOC · P.1 TBL · R.4

    01Connect

    Bring in all your sources

    Documents and databases, together

    Bring in documents, spreadsheets, reports, and databases.

    • Merged & nested tables
    • Scanned mixed layouts
    • Source and time on every record
  2. VERSION SNAPSHOTSMETRIC LOGICNET OF TAXANALYSIS RULESEXCLUDE TRIALSHANDBOOKEMEA = EU + UKSHARED MODELNEWMETRIC LOGICNET OF TAXANALYSIS RULESEXCLUDE TRIALSINCLUDE TRIALSHANDBOOKEMEA = EU + UKREVENUEQ3 OUT1,6241,712Q11,455Q21,548Q31,624EXCLUDE TRIALSMETRIC LOGICNET OF TAXANALYSIS RULESEXCLUDE TRIALSHANDBOOKEMEA = EU + UK

    02Understand

    Keep structure and meaning

    Context stays intact

    Keep document structure, data definitions, and business context intact.

    • Versioned snapshots
    • Topology across models
    • Reviewable rule by rule
  3. REVENUE?PRICEREGIONCHURNANALYSISNEWPRICECHURNREGION

    03Analyze

    Ask in plain language

    One question can cross every source

    Ask questions and let DataFact retrieve, query, and reason across your data.

    • Planned retrieval paths
    • Multi-agent split
    • Cross-channel corroboration
  4. 04Verify

    See the evidence

    Every answer stays traceable

    See the exact evidence behind every answer.

    • Row & table level rules
    • Full question audit trail
    • Feedback into knowledge

03Document intelligence

Understand documents as they were meant to be read.

DataFact keeps tables, figures, layouts, and page locations intact instead of flattening everything into text.

So when an answer points to a source, you can go directly to the exact place it came from.

  • Keep complex tables intact — merged cells, nested headers, and multi-column layouts remain structured
  • Trace every result — every extracted block keeps its page and location
  • Review what AI reads — edit, split, disable, or re-index individual content blocks
  • See figures in context — charts and figures can appear directly with the answer
Explore Document Intelligence

04Data analysis

Ask in plain language. See the data behind the answer.

Ask a business question and DataFact turns it into a query against the data and definitions your team already uses.

The SQL stays visible, so analysts can inspect exactly how the result was produced.

  • Natural language to SQL — ask questions without writing queries manually
  • Use your business definitions — keep metrics and data meanings consistent
  • Respect data permissions — users only see the data they are allowed to access
  • Inspect every query — keep SQL, results, and definitions visible together
Explore Data Analysis

05Data agents

Ask bigger questions.

Some questions need more than one query. DataFact agents break complex questions into steps, search documents, query databases, compare sources, and bring the findings together.

The result is not just an answer — it includes the analysis and evidence behind it.

  • Break complex questions into steps
  • Work across documents and databases
  • Cross-check conflicting sources
  • Return charts, findings, and citations together
Explore Data Agents

Governance

06

The right answer
for the right person.

DataFact applies access rules before data reaches the answer. Permissions, data roles, and business rules determine what each user can see, while every question and result remains traceable.

01
Scope
Control access by organization, team, and user.
02
Data permissions
Apply table-level and row-level access rules.
03
Business rules
Keep shared definitions and analysis rules consistent.
04
Audit trail
Review the questions, sources, queries, and answers behind each analysis.

How it fits

07

Works with the data
you already have.

No need to replace your warehouse or move everything into a new system. Connect DataFact to the documents and databases your organization already uses.

Sources

what you already have

  • PDF · Word · Excel
  • PPT · CSV · scans
  • Postgres · MySQL
  • Warehouse tables

Memory

parsed, modelled, versioned

  • Chunks with coordinates
  • Table & column semantics
  • Analysis model
  • Version snapshots

Agents

planning and retrieval

  • Document Q&A
  • Query generation
  • Multi-agent runs
  • Scheduled reports

Surfaces

where people meet it

  • Console
  • Embeddable chat
  • Open API — Slack, Teams, your own apps
  • Native bots: DingTalk, Feishu

Two ways to run it

08

Use DataFact
your way.

Start with DataFact in the cloud, or run it inside your own environment.

Cloud

Start without managing infrastructure.

Connect your data, create an agent, and start asking questions.

  • Working the same day, not after a procurement cycle
  • Subscription billing — scale seats and volume up or down
  • New capabilities arrive without an upgrade project
  • Regional options where data residency applies
Try demo
Private deployment

Keep DataFact inside your environment.

Deploy DataFact in your own cloud or infrastructure with your existing identity and security systems.

  • Your data centre or your private cloud tenancy
  • SSO / LDAP integration with your identity provider
  • Model serving on your own hardware
  • Licence term with your own upgrade windows
Contact us

Same features, same governance model, same API either way — so a team can evaluate on a subscription and move to a private deployment later without rebuilding its workspace.

Built for questions that matter.

09

The change wasn't speed. It was that the finance team stopped asking us to prove the number, because the answer already showed them where it came from.
Head of data platformAutomotive manufacturer · 40,000 staff
  • Manufacturing & automotive
  • Telecommunications
  • Metals & mining
  • Financial services

DataFact is used where answers need to be more than fast — they need to be understandable, traceable, and trusted.Customer stories

Start here

Bring us a question your team actually cares about.

Connect a document and a database, ask the question, and see exactly where the answer comes from.

Your data stays connected to every answer.