Q

2.0

Products Governance Qualix · Data Quality
Qualix — Data Quality & Metrics 7 Dimensions Governance · Vigil

Trust your data. Automatically.

Seven dimensions. One health score. Continuous.

Qualix scores every dataset across seven quality dimensions — continuously. Write rules in plain English. Watch trends. Get alerted before your dashboard does. Quality isn't a project you run once — it's a state you live in.

Built for stewards, CDOs, and data engineers who've had enough of chasing bad rows through tickets. Qualix scores. Qualix alerts. Your team owns the fix.

Qualix
Data Quality · Live scoring engine
Scoring
CUSTOMER DOMAIN · HEALTH Live
91 ▲ +2 this week
Across 23 tables · 847 columns · 1.5K+ records/scan
Completeness 94
Uniqueness 97
Validity 92
Consistency 78
Consistency dropped -4 on customer_id — upstream billing schema changed. Fix assigned to @steward.
Qualix in production
7
Quality Dimensions
83ms
Avg Execution
100%
System Success
1.5K+
Records / Scan
24/7
Continuous
Why Qualix Exists

You can't fix what you can't measure.

Most data quality "solutions" are weekly scripts, monthly audits, or quarterly tickets. By the time a problem is found, a dozen reports are already wrong and three people have been in meetings about it.

Qualix treats quality the way SREs treat uptime — continuous, measured, trended, alerted. Every dataset, every column, every day. Scores move. Thresholds fire. People act on problems instead of looking for them.

Data quality isn't a project. It's a continuous state.

The old way
You find out about bad data after the executive does.
01
Scripts run weekly. Drift happens daily. Gaps are found in dashboards, not in scans.
02
Rules live in notebooks and tickets. Nothing is reusable. Nothing is tracked.
03
Quality status is a spreadsheet update every quarter. No trends. No accountability.
Seven Dimensions · One Score

Every dataset, continuously scored.

Trends tracked. Thresholds enforced. Alerts before drift hits your dashboard.

94
Completeness
▲ +1 · 7d
Are required fields populated? Nulls, empties, placeholder values.
97
Uniqueness
— stable
Are primary keys unique? Duplicates surfaced by scan, not by production.
92
Validity
▲ +2 · 7d
Does the value match format, range, and domain rules? Email shapes, date bounds, enum lists.
78
Consistency
▼ -4 · 7d
Does the same fact agree across systems? Cross-source reconciliation, referential integrity.
89
Accuracy
▲ +1 · 7d
Does the value match ground truth? Reference data, lookup tables, external validators.
95
Timeliness
— stable
Did the data land on time? Freshness SLAs, staleness detection, late-arriving rows.
88
Governance
▲ +3 · 7d
Ownership, classification, and documentation coverage across the catalog.
One health score
A single weighted number per dataset, per domain, per project. Roll up. Drill down. Alert on drift.
What Qualix Does

Quality as a living layer.

📊
Continuous 7-Dim Scoring
Every dataset, every domain, every day. Scores stored, trended, and surfaced at every level of the catalog.
  • Per-column and per-table scores
  • Project, domain, and enterprise roll-ups
  • Historical trend snapshots
📝
Plain-English Rules (DQOs)
Data Quality Objects let stewards write rules the way they'd describe them — "customer_id must be unique and non-null" — and Qualix compiles and runs them.
  • Rule-based assertions on any column
  • Composite rules across joined tables
  • Reusable rule templates
🚨
Threshold Alerts
Configurable severity and routing. When consistency slips below 80, someone hears about it — in Slack, Teams, or Vigil's queue — before the dashboard breaks.
  • Severity tiers · info, warn, critical
  • Route to owners, stewards, or on-call
  • Suppress-and-learn for noisy rules
⏱️
Real-Time Execution
Qualix runs on DuckDB — millisecond profiling on massive datasets. No Spark cluster to warm up. No hourly batch to wait for.
  • 83ms average execution
  • In-memory scans at warehouse scale
  • No extract · scan in place
🔍
Root-Cause Analysis
When a score drops, Cezu and Vigil walk the lineage — which upstream changed, which rule tripped, which rows are affected.
  • Lineage-aware drift analysis
  • Affected-rows drill-down
  • AI-assisted fix suggestions
📈
Dashboards for Every Role
Project-level views for CDOs. Column-level drill-down for stewards. Rule-level detail for data engineers. Same scores, different cockpit.
  • Executive roll-up dashboards
  • Steward work queues
  • Engineer rule-debug views
How It Works

Scan. Score. Alert. Act.

A loop that runs on its own — without a notebook, a cron, or a Spark job.

1
🔌
Scan
Qualix reaches into every source — warehouse, lake, file, catalog — and profiles what's there.
2
📐
Score
DQO rules and the 7 dimensions fire. Every column, every row, gets a number.
3
🚨
Alert
Thresholds trip. Owners get pinged. Vigil opens a review. Dashboards stay honest.
4
🛠️
Act
Root cause is in the bubble. Fix upstream, suppress the rule, or quarantine the rows.
DQOs · Data Quality Objects

Write a rule in plain English.

You describe the rule the way a steward would describe it. Qualix compiles, schedules, runs, scores, and surfaces the result across every dashboard. No notebooks. No cron jobs. No brittle scripts.

  • Rule-based assertions on any column or combination
  • Severity tiers, thresholds, and routing — per rule
  • Reusable templates for common domains (customers, transactions, claims, accounts)
  • Full audit trail — who wrote it, when it ran, what it scored
DQO · customer_id.uniqueness
# Written by steward · compiled by Qualix
rule customer_id_is_unique
  on    "sales.customer"
  assert "customer_id" is unique
  and    "customer_id" is not null
  severity critical
  threshold 99.5
  alert   "#data-stewards"

── last run ──
  scanned   1,547,203 rows
  unique    1,547,198  (99.9997%)   ✓ PASS
  nulls     0                          ✓ PASS
  duration  71ms

── trend · last 7 days ──
  mon  99.9998   thu  99.9998
  tue  99.9997   fri  99.9997
  wed  99.9997   sat  99.9997
                      sun  99.9997

STATUS: GREEN · next run in 23 min
Use Cases

Where Qualix earns its keep.

🏛️
Pension & Benefits Accuracy
Public pension funds · retirement systems
Score beneficiary records, contribution histories, and service credits across legacy systems. Catch duplicate SSNs, wrong birthdates, and orphaned accounts before actuarial reports lock.
✓ Audit-ready beneficiary data, continuously verified
🏦
Financial Reporting Integrity
Banks · insurers · asset managers
Enforce validity and consistency on trade, position, and ledger data as it lands. Catch broken joins, missing cost bases, and currency mismatches before close — not after.
✓ Close-of-period reports that survive audit
⚕️
Healthcare Data Integrity
Payers · providers · health systems
Score patient identity, claim completeness, and code validity continuously. Surface upstream drift from each EMR, each claims feed, each partner — before it corrupts analytics.
✓ Clinical and claims data you can defend
📋
Regulatory Reporting
State agencies · federal submissions
Encode regulatory rules as DQOs. Every reporting cycle runs the same scan. Every exception has an owner, a deadline, and an audit trail — without a fire drill.
✓ Submissions filed clean, on time, every cycle
Why Qualix

Not another DQ tool.

Qualix is a module of a unified platform — not a standalone product that needs its own integrations, its own catalog, and its own semantic model.

Same semantic layer · every module
Rules fire against the same definitions your analysts, BI tools, and stewards use. No two versions of "customer." No two versions of "revenue."
Millisecond scans · no cluster
Runs on DuckDB. Profile 1.5K+ records in 83ms — without a Spark job, a warehouse ticket, or an approval.
English rules · engineer execution
Stewards write rules the way they describe them. Qualix compiles them into real, tested, observable assertions — no Python required.
Lineage-aware root cause
When a score drops, Qualix follows the lineage backward with Cezu — tells you which upstream changed, not just that something broke.
Continuous by default
No "go run the quality job." Qualix is always on — trends are captured, drift is caught, before the dashboard shows it.
Deploys where your data lives
Private VPC, air-gapped, on-prem, or cloud. No data leaves your boundary. Scans happen next to the data.
Qualix + Vigil

Qualix scores. Vigil acts.

Qualix is the scoring engine. Vigil — xAQUA's AI Data Governance Agent — watches the scores, triages the alerts, and turns drift into work the right person can own.

Ask Vigil in plain English: "which catalogs failed quality this week?" — it pulls every DQO result, ranks the fires, and hands back a report ready for the CDO.

Meet Vigil · The AI Data Governance Agent →
Ask Vigil Governance Agent
👤
Which catalogs failed quality this week?
Scanned 47 catalogs · 4 below threshold:

· sales.orders — Consistency 78 (▼4) · upstream schema drift
· claims.raw — Completeness 81 (▼6) · late-arriving rows
· hr.employees — Validity 84 (▼3) · format breach on phone
· finance.ledger — Timeliness 79 (▼8) · feed 2h late

Report routed to stewards. Fix ETAs logged.

Ready to stop chasing bad data?

See Qualix score a live dataset across all seven dimensions, write a DQO in plain English, and watch it fire — in under fifteen minutes.