Pillar One
Validity
Diagnoses and corrects invalid data — across six categories of common defects.
Data Type
Values that are not type correct, such as symbols, parentheses or commas.
Range Constraints
Values in non-permissible ranges.
Mandatory Constraints
Values or columns that should not be missing.
Unique Constraints
Values that are not unique across a dataset (such as customer number).
Set-Membership Constraints
Values do not come from a pre-defined set (such as marital status or gender).
Expression Patterns
Values do not fit expected formatting (such as phone numbers or email addresses).
Pillar Two
Accuracy
Matching, merging, or appending data to increase the accuracy of the data. For example, appending the postal code (or ZIP+4 in the United States).
- Append ZIP+4 codes to U.S. addresses
- Merge with verified reference datasets
- Cross-reference for standardization
Verified, matched,
and appended.
Inferred from
predictive analytics.
Pillar Three
Completeness
Using predictive analytics to infer values for missing fields. Where data isn't available, we apply 30+ years of analytics experience to predict it.
- AI-driven inference for missing fields
- Predictive modeling backed by decades of experience
- Confidence scoring on every inferred value
Industries we serve
Built for teams that depend on clean data.
From healthcare records to e-commerce customer files, our cleansing engine adapts to your domain and your compliance requirements.
HIPAA-aware processing
Patient address records, provider directories, and claims data — cleansed without exposing PHI.
Customer file hygiene
Standardize shipping addresses, deduplicate accounts, and reduce undeliverable orders.
KYC & compliance ready
Validate addresses against authoritative sources for AML, KYC, and risk workflows.
Constituent records
Voter rolls, mailing lists, and outreach databases standardized to USPS specifications.
Average turnaround
< 10 min
per file processed
How it works
Five steps from raw to refined.
- 01
Upload
Send us a CSV, Excel, or pipe-delimited file. We support any size — encrypted in transit.
- 02
Profile
We scan every column to detect address fields, names, phones, and identify defects.
- 03
Validate
Each record is checked against USPS, postal authorities, and predictive models.
- 04
Standardize
Casing, abbreviations, and component order normalized to USPS specifications.
- 05
Return
Download your cleansed file with a full report of every change we made.
FAQ
Common questions
Which file formats do you accept?
We can accept all known file formats. The self-service portal auto-detects the file and encoding, and can accept CSV, Excel (.xlsx/.xls), pipe-delimited and tab-delimited.
What is the maximum file size you can accept?
We can accept files of any size for any/all services. For the self-service portal, files are limited to 256 columns and 2 million rows.
How long does a typical job take?
Most files under 2 million records complete in under one hour. Larger files run in batches with progress updates.
Do you support international addresses?
Yes. We can cleanse data from any country. The self-service portal (with CASS, NCOA and ZIP+4 append) is only for US-based addresses.
Is my data secure during processing?
All uploads are TLS-encrypted. Files are stored only for the duration of processing and purged within 24 hours of download.
Can I integrate cleansing into my pipeline?
Yes — we offer a REST API for automated submissions. Talk to us about SLAs and bulk pricing.
What happens to records we can't standardize?
They're flagged with the reason (e.g. invalid ZIP, undeliverable per USPS) but still returned, so you can review them manually.