Q

2.0

Products / SemantIQ
๐Ÿ“š
Data Catalog ยท SemantIQ The semantic backbone of xAQUA

Auto-discovered. AI-enriched. Steward-confirmed Data Catalog for the AI Era.

The common language for conversational data management and analytics.

SemantIQ connects to your data, discovers the schema, and drafts the catalog with AI. Your stewards review and confirm. The result becomes the shared language behind every xAQUA capability โ€” so your team can go from raw data to advanced analytics in plain English.

๐Ÿค–
Includes Zyra โ€” the AI Data Steward built into SemantIQ. Ask anything about your catalog in plain English. Try her below โ†“
50+
Enrichment fields per table & column
7
ISO quality dimensions, continuously scored
40+
Semantic types auto-classified โ€” PII, dates, currency
30+
Metrics across 7 categories โ€” quality, volume, KPI
Auto
Categorical Data classification on every sync
0
Rows of data ever sent to the LLM
XAQUA AGENTS & PRODUCTS CEZU Concierge CONVERSESQL SQL Generation QUALIX Quality SENSEMASK PII Masking NARRATIX Reports + SemantIQ Semantic Layer ยท Living Data Catalog 50+ fields 7 quality 40+ types 30+ metrics ๐Ÿ“š Catalog ๐Ÿ“– Glossary ๐Ÿ•ธ๏ธ Lineage ๐Ÿ“Š Quality ๐Ÿ›ก๏ธ Governance ๐Ÿ”„ Change Mgmt YOUR DATA SOURCES Snowflake PostgreSQL Salesforce CSV ยท Excel ยท Parquet Databricks + more Auto-discovered ยท AI-enriched Steward-confirmed ยท Governed Zero raw data ever leaves your environment Schema metadata only โ€” never rows
One Semantic Layer ยท Many Catalogs ยท Many Sources

Search every catalog. In plain English. Or via API.

Most catalogs are silos โ€” one project, one set of sources, one isolated search. SemantIQ federates them. Ask one question and get one answer that spans HR, Sales, Finance, Operations. Across catalogs. Across sources. Across files and virtual assets.

YOUR ENTERPRISE CATALOGS ๐Ÿ“š Customer 360 Snowflake ยท Salesforce 412 assets ๐Ÿ‘ฅ HR Master Workday ยท PostgreSQL 287 assets ๐Ÿ’ฐ Finance & GL Oracle ยท Databricks 523 assets ๐Ÿ“ฆ Operations SAP ยท BigQuery ยท Files 386 assets ๐Ÿ›ก๏ธ Compliance & Risk PostgreSQL ยท Files 239 assets SemantIQ Federated Search Layer BM25 keyword + semantic vector hybrid ยท One query, all catalogs ยท Zero data movement + BM25 + Semantic = HYBRID TWO ACCESS MODES Z ๐Ÿ’ฌ Ask Zyra Conversational ยท Built for users "Find tables with PII data" โ†’ spoken answer { } โšก Hybrid Search API Programmatic ยท Built for developers POST /v1/semantiq/search
๐Ÿค–
SemantIQ Federated Search LIVE DEMO
Conversational + API ยท One semantic layer ยท Powered by Zyra
Online ยท Ready to help
Searching across: ๐Ÿ“š 5 catalogs โ€ข ๐Ÿ”Œ 23 bound sources โ€ข Cataloged assets: 1,847 โ€ข Avg health: 94/100
Z
Hi! I'm Zyra, your AI Data Steward. I have full context across all 5 of your enterprise catalogs โ€” every table, column, relationship, quality score, and governance classification.

Try a federated question โ†’ or ask me anything. My answers can span multiple catalogs and sources.
Try a federated question
Why traditional catalogs die

Buy a traditional catalog. Hire a team to populate it. Watch it go stale.

Traditional catalogs are documentation-first. Every business name, every description, every PII tag, every glossary entry โ€” typed by hand, by your team, on top of their day jobs. Six months in, half is wrong. A year in, half doesn't exist. The data team writes definitions. The business team disagrees. The catalog you bought to standardize answers becomes the source of three different answers.

That's not a team failure. It's a catalog failure. Documentation-first catalogs scale with human effort โ€” and human effort doesn't scale. Schemas evolve faster than your stewards can type.

SemantIQ inverts the problem. The AI drafts every business name, every description, every classification โ€” read directly from the schema, never from your data. Your stewards review, edit, and confirm. The catalog stays current because curation scales. Transcription doesn't.

Your stewards stay in charge. The AI does the heavy lifting.
๐Ÿ“š Traditional Catalog ยท 11 months in Coverage: 27%
โš 
customer_master.email_addr
No description โ€” added 8 months ago
Stale
๐Ÿ“
orders_fact.line_total_amt
"This is the total" โ€” last edited 14 months ago
Vague
โš 
product_catalog.sku_status_cd
Conflicts with two glossary entries
Conflict
โณ
employee_master.dept_id
42 columns awaiting description โ€” backlog: 9 weeks
Backlog
โœ“
finance_gl.posting_date
Confirmed and certified
Certified
Documentation-first architecture. Human effort doesn't scale.
Most catalogs document. SemantIQ executes.

A catalog that does work, not just describes it.

This is the line that separates SemantIQ from every legacy catalog on the market.

Collibra, Atlan, OpenMetadata โ€” these are documentation tools. They tell you what exists. They put a glossary in front of it. They draw lineage diagrams. After that, the work happens somewhere else: a SQL editor, a BI tool, a transformation product, an ML platform. The catalog is a reference manual.

SemantIQ is built differently. The same enriched metadata you see in the catalog is the metadata that drives execution across xAQUA. Update a column's business name in SemantIQ โ€” ConverseSQL generates better SQL. Mark a column as PII โ€” SenseMask masks it automatically. Confirm a cross-source join โ€” ConverseDataIQ uses it in query plans. Certify a table โ€” Trust Score reflects it for every business user.

The catalog isn't a side product.
It's the layer the platform runs on.
Legacy Catalogs
  • Inventory of assets
  • Glossary of terms
  • Lineage diagrams
  • Workflow approvals
  • Audit logs
SemantIQ
  • Everything legacy catalogs do
  • + Drives SQL generation
  • + Drives PII masking
  • + Drives query planning
  • + Drives Trust Score
  • + Drives natural-language access
โ€” Updates flow to runtime โ€”
โ†’ ConverseSQL โ†’ ConverseDataIQ โ†’ Cezu โ†’ Qualix โ†’ SenseMask โ†’ Narratix โ†’ Reeve โ†’ Entity 360
One product. Five pillars. Eleven capabilities.

Catalog. Glossary. Lineage. Quality. Governance. All native.

A single shared semantic model โ€” organized into five pillars. Every module reads and writes through the same governed layer.

PILLAR 01 Connect & Discover Sources in. Schema out. Continuously.
๐Ÿ”Œ
Bound Data Sources

Connect once. Bind many. One catalog can bind to DEV, TEST, and PROD versions of the same source โ€” schema drift between environments is tracked automatically.

Snowflake Databricks Postgres Salesforce Files
๐Ÿ“
File Assets

Drag-and-drop a CSV, Excel, Parquet, or JSON file. SemantIQ infers the schema, detects multi-sheet workbooks, runs the same enrichment pipeline as a database source, and binds it to a catalog.

CSV ยท Excel ยท Parquet ยท JSON
๐Ÿ”„
Schema Sync & Drift

SemantIQ watches every bound source. Breaking changes raise alerts before they hit production. Non-breaking changes update the catalog. Schema change history is audit-grade and queryable.

Continuous Drift alerts
PILLAR 02 Enrich & Classify Schema becomes business meaning.
โœจ
AI Enrichment Studio

Business names. Plain-English descriptions. Domain. Sensitivity. PII detection. Column roles. Glossary links. Over 50 enrichment fields per container and per column โ€” drafted by the AI, confirmed by your stewards.

50+ fields Confidence-scored Lockable
๐Ÿ“Š
Quality & Metrics Intelligence

Seven ISO quality dimensions scored continuously. Over 30 metrics across 7 categories. Trend snapshots. Threshold alerts. The same scores feed Qualix and the Trust Score visible to every business user.

30+ metrics 7 categories Trend alerts
๐Ÿ“–
Business Glossary

A living glossary of business terms linked directly to the physical columns that implement them. Cross-catalog. Cross-source. When the business says "Active Member," every system that uses the term points to the same definition.

Cross-catalog Column-linked
PILLAR 03 Relate & Reason From isolated tables to a governed graph.
๐Ÿ•ธ๏ธ
Cross-Source Relationship Intelligence

Affinity scan finds joins you didn't know existed. Four signals weighed: name (40%), type (20%), value overlap (30%), cardinality (10%). One click to confirm. Confirmed joins flow straight into the query planner.

4-signal scoring One-click confirm
๐Ÿงฌ
Analytical Semantics

Drill-down hierarchies. Metric decomposition trees. Asset inheritance. The structures business analysts navigate by โ€” codified once, reused everywhere. Narratix uses them. ConverseDataIQ uses them. Reeve uses them.

Drill paths Metric trees Inheritance
PILLAR 04 Govern & Control Strict governance. Audit-grade. By design.
๐Ÿ›ก๏ธ
Governance & Certification

Sensitivity levels from PUBLIC to TOP_SECRET. Certification from UNCERTIFIED through CERTIFIED. Owner and steward assignment. Compliance reporting. Every classification captured as part of the semantic layer โ€” not bolted on.

5 sensitivity levels Steward-owned
๐Ÿ›‚
Governed Change Management

The catalog is under strict governance. No one edits definitions, classifications, or relationships directly. Every change goes through a change request โ€” submitted, reviewed by the catalog owner or steward, accepted or rejected with full reasoning captured.

CR workflow Audit trail Owner approval
PILLAR 05 Conversational Intelligence Ask anything. In plain English.
๐Ÿง 
AI Insights ยท Powered by Zyra

Zyra is the AI Data Steward built into SemantIQ. She writes insights as the data changes โ€” root-cause analyses when quality drops, volume anomalies, distribution drift, schema impact, governance recommendations. Every insight rated by your team. Ask her anything about your catalog in plain English. Try Zyra below โ†“

Root cause Conversational Self-improving Catalog-grounded
Six steps. Mostly automatic. Always under steward control.

From cold connection to enriched catalog โ€” in minutes.

1
๐Ÿ”Œ
Connect
Point at a database, file, or SaaS source. Data is never copied.
2
๐Ÿ”„
Sync
Schema introspection runs. Every change logged.
3
๐Ÿ†”
Register
Permanent UUIDs. Renames tracked. Deletes are soft.
4
โœจ
Enrich
AI drafts. Stewards review and confirm via change request.
5
๐Ÿ•ธ๏ธ
Relate
Cross-source affinity scan. One-click confirm joins.
6
โšก
Power the platform
Every xAQUA agent reads from the catalog.
โšก
One catalog. One language. Every product.
Three magic moments

This is what changes for your team.

MOMENT ยท 01
Connect a Snowflake database. Walk away. Come back.
You connect one source. Twenty minutes later, you have a fully drafted catalog: every table named in business language, every PII column flagged, every quality dimension scored, every probable join surfaced for your stewards to review.
โ†’ The work that used to take a quarter takes a coffee.
MOMENT ยท 02
Ask Zyra: "Why did customer_profile quality drop?"
Zyra finds it in seconds. The completeness score dropped because null rates on customer_email and customer_phone spiked overnight. The trigger: a column was renamed in the CRM source at 11:47 PM. The ETL pipeline still references the old name.
โ†’ Root cause. Suggested fix. Change request ready to submit.
MOMENT ยท 03
Add a new source. The whole platform gets smarter.
A new file lands. SemantIQ enriches it. The cross-source affinity scan finds three joins to existing tables. ConverseDataIQ can now answer questions that span the new data on day one. Nothing else changed.
โ†’ Everything works.
Not a catalog. Not a table format. Something else.

SemantIQ vs. legacy catalogs and modern table formats.

The catalog category is crowded and confused. Buyers compare products that don't actually do the same thing. Here's the clean separation.

Collibra ยท Alation
Legacy Catalogs
Documentation systems. Built when the job was to inventory enterprise data and put a glossary in front of it. They do that well. But the work they enable still happens elsewhere โ€” in SQL editors, BI tools, ETL platforms, ML systems. Updating a Collibra entry doesn't change how anything else behaves.
Atlan ยท OpenMetadata
Modern Catalogs
Better at automation. Active metadata, lineage from logs, AI-assisted descriptions. The architecture is more current, but the role is the same: documentation and governance. The catalog is still a reference manual the rest of the stack ignores at runtime.
Apache Iceberg
Table Format
Not a catalog at all. It's an open table format that engines like Spark, Trino, and Flink query against. It often sits under a catalog, not instead of one. Treating Iceberg as a catalog comparison is a category error.
One is a reference manual. The other is the operating layer.
Seven things every catalog should have. Most don't.

Why SemantIQ is built right for the AI era.

๐Ÿงฌ
Semantic-first, not schema-first

Most catalogs index your schema. SemantIQ enriches it. Every container and column carries 50+ semantic fields. That's why xAQUA agents can answer questions in plain English.

๐Ÿ›‚
Strict governance, not optional governance

Most catalogs allow direct edits and call the audit log "governance." SemantIQ doesn't. Every change flows through a formal change request reviewed by the catalog owner or steward. Accepted or rejected with reasoning logged.

๐Ÿ”’
Zero raw data ever leaves your environment

The LLM never sees customer data. Schema metadata only. Enforced at the service layer. The same model can run inside your private deployment if you need it to.

โœ“
Auto-generated, human-confirmed

Every AI-generated field carries a confidence score. Every confirmed field is treated as absolute truth. Stewards can lock fields against re-enrichment. The catalog gets smarter as your team works in it.

๐Ÿ†”
Immutable IDs, additive enrichment

A column rename never breaks your catalog. Permanent UUIDs survive structural change. Enrichment is overlay, not overwrite. Curation work compounds over time instead of getting wiped on every refresh.

๐Ÿ”€
Physical and virtual assets, one model

Most catalogs only know about source tables. SemantIQ catalogs both โ€” discovered sources and xAQUA-created assets: queries, blends, pipelines, reports, ADLs, ML applications.

๐ŸŒ
Federated by design โ€” across every catalog

Most catalogs are silos. SemantIQ is a unified semantic layer across all your catalogs โ€” HR, Sales, Finance, Operations, Compliance. Hybrid BM25 + semantic search returns one answer that spans every catalog, source, and asset you have access to. Through Zyra or via API.

Built for the whole data team

One catalog. Four roles. Four wins.

๐Ÿ‘ค
Data Stewards

Stop writing descriptions by hand. Curate AI-drafted ones. Confirm what's right, edit what's wrong, certify what's ready. Every change flows through the change-request workflow you control.

๐Ÿ“‹
Chief Data Officers

A defensible, board-ready foundation. Sensitivity classification, certification status, governed change management, and full audit trails โ€” all native, all current, all auditable.

โš™๏ธ
Data Engineers

The catalog tells your pipelines what changed. Breaking schema changes raise alerts before they hit production. Cross-source joins are reusable. Migration work is catalog-driven.

๐Ÿ’ผ
Business Users

You don't have to know what SemantIQ is. You just notice that when you ask Cezu a question, it returns the right answer โ€” because the platform finally speaks your language.

Ready to see your data understood?

Stop writing your catalog.
Start curating it.

See SemantIQ enrich a real source from your stack โ€” in a 30-minute walkthrough.