Q

2.0

📚 DocIQ — Document Intelligence Suite 6 Functions · 1 Platform

Documents in. Knowledge out.

Chat with them. Generate from them. Extract fields out of them. Organize them. Six functions, one platform.

DocIQ is xAQUA's document intelligence suite. The same indexed corpus powers six ways of working with your unstructured content — Chat for cited Q&A, Create for AI-generated documents from your templates, Extract for structured field pulls, Library for governed storage, Review for contract review against playbooks, and Comply for continuous regulatory compliance. One ingestion pipeline. One semantic layer. One audit log.

Production-proven at a $300B pension fund. Page-level citations. Confidence scoring. Human-in-the-loop everywhere it matters.

DocIQ in production
6
Functions
1
Indexed Corpus
100%
Cited Outputs
VPC
Private Deploy
$300B
Reference AUM
Why DocIQ Exists

Your knowledge work is six jobs in disguise.

Every document operation in your enterprise — answering a policy question, drafting an SOP, pulling a field from a death certificate, finding last quarter's BPA — is the same underlying problem: turn unstructured content into trustworthy output. But teams cobble it together with eight to thirteen different tools — chatbots, doc generators, OCR vendors, contract-review platforms, GRC suites, SharePoint, and the integrations between them. None of them share context. None of them share the audit trail.

DocIQ collapses them into one platform. The same indexed document answers a chat question, sources a generated draft, supplies fields for an extraction job, lands in the Library, gets reviewed against a playbook, and gets checked against the regulations that govern your business — with one set of permissions, one citation model, one audit log.

Six ways of working with documents. One platform that knows them all.

The old way
A whole document Frankenstack. Zero shared context.
CHATBOT
A standalone Q&A bot indexed against one folder. Hallucinates clause numbers. Has no idea what's in the doc generator.
DOC GENERATOR
A separate template app. Re-uploads of the same source files. Outputs vanish into someone's drive.
OCR VENDOR
A third tool for pulling fields. CSV export. No provenance. No way to ask "show me the source."
SHAREPOINT
The graveyard. Every output ends up here, unfindable, mixed with everything else, no version chain.
Federated architecture · No migration

Stop bringing documents to the AI. Bring AI to the documents.

Box AI runs on Box. Microsoft Copilot runs on SharePoint. Adobe AI runs on PDFs in Acrobat. Each vendor's "AI" deepens their own moat — and leaves your enterprise documents stranded across eight different stores. DocIQ runs on all of them. At the same time. With one search, one permissions model, one audit log.

b WHEREVER YOUR DOCUMENTS LIVE DOCIQ FEDERATION LAYER WHATEVER YOUR TEAM NEEDS Amazon S3object storage Azure Blobstorage Google Cloudstorage SharePoint+ Teams OneDriveM365 Google DriveWorkspace Boxcontent cloud Confluence+ wikis Uploaddrag & drop + Network shares, FTP… Q DocIQ FEDERATION LAYER 1 Connect & sync Set cadence · honor source ACLs · incremental 2Parse + OCRPDF · DOCX · images · scans · transcripts 3Embed + indexpgvector + BM25 · per-tenant · permission-aware 4Serve to six modesChat·Create·Extract·Library·Review·Comply YOUR BYTES NEVER LEAVE THE SOURCE OR YOUR TENANT 💬DocIQ ChatQ&A on docs · cited · per-domain agents ✍️DocIQ CreateGenerate any document from your templates 🧲DocIQ ExtractStructured fields · confidence-banded 📚DocIQ LibraryIndex, version, govern · source of truth ⚖️DocIQ ReviewContract review · playbook coverage · risks 🛡️DocIQ ComplyContinuous regulatory compliance · audits One semantic layer · one auth model
Compare Vendor-locked AI: works only on its own store vs DocIQ: works across every store, all at once
/ 01
Connect, don't migrate
Point at an S3 prefix, a SharePoint site, a Box folder, a Drive collection. Set a sync cadence — hourly, daily, on-change. DocIQ ingests, embeds, and indexes incrementally — including deletes, ACL changes, and new versions. No big-bang migration. No rip-and-replace.
/ 02
Honor source permissions
DocIQ doesn't re-implement your access controls — it inherits them. If a user can't read the source PDF in Box, they can't see it in Chat, Library, or Extract results. Permissions stay where they belong: on the source of truth.
/ 03
We index. You store.
Your bytes stay in S3, SharePoint, Box, Drive — wherever they already are. DocIQ stores the index, the embeddings, the metadata, and the workflow state. Migration is not a precondition for AI. Your data warehouse already taught us this lesson.
The Six Functions

One ingestion. Six ways to work.

Upload a document once. DocIQ indexes it once. Then chat it, generate from it, extract from it, and shelve it — all from the same canonical source.

💬
DocIQ Chat
Q&A · cited answers
Ask your documents. Get the answer with the page it came from.
Grounded retrieval over your unstructured corpus — PDFs, Word, wikis, transcripts, scans. Hybrid BM25 + vector search with page-level citations on every claim. If retrieval is below confidence, DocIQ refuses to fabricate.
  • BM25 + pgvector + RRF fusion retrieval
  • Page-level citations with highlighted passages
  • Per-domain agents (Pension, HR, Legal, Finance, IT)
  • Effective-date awareness for amended documents
Explore DocIQ Chat → 0 hallucination tolerance
📄
DocIQ Create
Generation · DocSynthesizer
Drop your sources. Pick a template. Get a branded, version-controlled document.
Generate any document type from your own templates — SOPs, BPAs, RACI, test cases, requirements specs, briefing memos, audit responses, board papers — using your own source files, in your own brand. Four-stage wizard: Sources & Template → Generate → Review & Edit → Publish & Export. Full approval workflow built in.
  • 5 system templates · custom .docx upload supported
  • Multi-stage LLM pipeline with progress visibility
  • Section-by-section review and inline editing
  • Approval state machine with immutable audit trail
Explore DocIQ Create → DOC-YYYY-NNNNN stable IDs
🔧
DocIQ Extract
Field extraction · structured pulls
Turn 500 PDFs into a clean spreadsheet — with confidence scores you can route on.
Define a schema (or pick a system one). Drop documents. DocIQ pulls every field with a confidence score and a provenance pointer. Low-confidence rows route to human review automatically.
  • System schemas: Death Certificates, Contracts, Financial KPIs, Audit Evidence
  • Custom schema designer · any field, any document type
  • Per-field confidence band (high · medium · low)
  • Inline correction · accepted values save back as ground truth
Explore DocIQ Extract → Confidence per field
📚
DocIQ Library
Knowledge store · governed
Every output you've ever generated. Every template you've ever used. One organized home.
The Library is the governed knowledge store for everything DocIQ produces. Three system collections — Templates, Output Documents, Extraction Schemas — plus unlimited custom collections. Pinned items, smart filters, and cross-collection search.
  • 3 system collections, unlimited custom folders
  • Cross-collection search (full-text + tag)
  • Pin, version, supersede, archive — all tracked
  • Permissions inherited from source-system ACLs
Explore DocIQ Library → Versioned · audit-ready
⚖️
DocIQ Review
Contract review · playbook coverage
Contract review at the speed of decision. Coverage check, risk scan, redline draft — in 90 seconds.
Drop a contract in. Within 90 seconds you have a complete review report — playbook coverage, risk findings, executive summary — with citations to the exact contract clause and the playbook excerpt that drove every verdict. Works with or without a formal playbook. Output: a redline-ready Word document with tracked changes.
  • Coverage check vs playbook · risk scanning · clause classification
  • Bootstrap a playbook from your historical contracts
  • Counterparty-specific learning suppresses recurring false positives
  • Tracked-changes DOCX export — send straight to counterparty
Explore DocIQ Review → ~90s per contract
🛡️
DocIQ Comply
Continuous regulatory compliance
Continuous compliance against the regulations that govern your business — not a quarterly fire drill.
Comply maintains a living map between the regulations that apply to you (HIPAA, SOX, GDPR, FedRAMP, ISO 27001, PCI-DSS, ERISA, custom obligations) and your policies, procedures, controls, and contracts. When a document changes, Comply re-checks it. When a regulation changes, Comply tells you what to review. Audit packages export in seconds.
  • Major-framework library + custom regulation support
  • Real-time compliance posture dashboard · cross-framework gap consolidation
  • Change-impact analysis when regulations are amended
  • Audit-evidence package generation — weeks of work, one export
Explore DocIQ Comply → Audit-ready · always
How It Flows

One pipeline. Six destinations.

Every document enters DocIQ once and becomes available to all six functions instantly.

📥
Ingest
PDF · DOCX · XLSX · HTML · email
🔍
Index
Chunk · embed · BM25 + vector
💬
Chat
Cited Q&A
✍️
Create
Generated docs
🧲
Extract
Field pulls
📚
Library
Governed store
⚖️
Review
Contract review
🛡️
Comply
Continuous compliance

Upload once, ingested once, indexed once. The output of Create, Extract, Review, and Comply lands in the Library — where it's discoverable by Chat on the next question. The loop closes.

Shared infrastructure

One stack underneath. Six products on top.

DocIQ isn't six products bolted together. It's one ingestion pipeline, one semantic index, one auth model, one audit log — exposed through six purpose-built UIs.

That's why a document you indexed for Chat is instantly available as a source file for Create. It's why a corrected extraction in Extract becomes ground truth for the next Chat answer. Shared substrate, not stitched APIs.

  • Multi-tenant — every query scoped by tenant_id
  • JWT auth · SSO via Keycloak · role-based access
  • Air-gapped or VPC deployment — documents never leave the boundary
  • Local embeddings · local LLM inference · zero-retention configurations
⚙️ The DocIQ stack
🌐
Single-page app
HTML · 6 mode panels · JWT + SSO
UI
FastAPI lifecycle layer
Python · 51 REST endpoints · async
API
🧠
LLM Gateway
Open-weight LLMs · 128K context · pluggable
LLM
🔎
Hybrid retrieval
BM25 + vector · RRF fusion · per-tenant
SEARCH
🗄️
PostgreSQL + pgvector
Multi-tenant · async asyncpg · HNSW
DATA
📦
Object storage
S3 · GCS · Azure Blob · local
FILES
Use Cases

Where DocIQ earns its keep.

Most enterprise use cases need two or three of the six functions. DocIQ ships them as one workflow.

🏛️
Pension Benefits Operations
$300B pension fund · production today
Caseworkers chat the post-retirement earnings limit rules with citations. They create branded SOPs from legacy CSSR reviews. They extract death certificate fields for DAR processing. Every output is shelved in the Pension Benefits collection.
Chat Create Extract Library
⚖️
Contract & Compliance Review
Legal · procurement · GRC
Extract obligations, dates, parties, and indemnity caps from contract portfolios. Chat the corpus to find every contract with a non-standard liability clause. Library houses the redlined drafts. Audit teams never paste between tools again.
Chat Extract Library
🛡️
Audit & Evidence Collection
Internal audit · SOC 2 · regulatory
Extract evidence fields from policy PDFs, control runs, and exception logs. Create the audit narrative from those fields and the underlying policies. Chat answers ad-hoc auditor questions with citation trails.
Create Extract Chat
📋
SOP & Process Documentation
Operations · ITSM · enablement
Create SOPs and BPAs from legacy memos, transcripts, and code repos. Library publishes them with full version history and supersession chains. Chat lets every employee ask the SOP directly instead of reading it.
Create Library Chat
Why DocIQ

Not a chatbot. Not an OCR vendor.

DocIQ is the unified document intelligence platform that the Frankenstack of point tools has been pretending to be.

Federated — not vendor-locked
Box AI works on Box. Copilot works on SharePoint. DocIQ works on every store at once — S3, Azure, GCS, SharePoint, OneDrive, Drive, Box, Confluence. One search across all of them. No migration required.
One ingestion, six outputs
Index a document once. Chat it. Generate from it. Extract fields. Shelve it. Review against a playbook. Check against regulations. No re-uploads. No reconciliation between tools.
Citation-first, always
Every Chat answer cites a page. Every Create section cites its source files. Every Extract field cites a paragraph. Provenance is the default, not a bolt-on.
Confidence scoring everywhere
Refuse-to-generate on Chat. Confidence bands on Extract. Approval gates on Create. The model never pretends to know what it doesn't.
Approval workflow built in
Six-state machine: draft → in_review → approved → published. Reject and revision-request supported. Immutable audit trail per document.
One platform, one agent
DocIQ rides on the same xAQUA semantic layer as ConverseDataIQ. Ask about a number, then about the policy that defines it. One conversation, one context.

Six functions. One platform. Thirty minutes.

See DocIQ index your documents and answer them, generate from them, extract their fields, and organize the outputs — in a single live demo. Bring your own corpus.