Address Integrity Skill Logo

Address Integrity Skill

Surfaces inconsistencies and address-linked risk signals across borrower documents before funding.

Design Intent

Address integrity checks are simple enough to seem mechanical, but subtle enough to carry real risk.

The intent of this skill is to treat address validation as a judgment-support problem, not a deterministic rule check.

Why This Is a Standalone Skill

Address integrity checks show up across underwriting and operations. Address data is spread across borrower documents, third-party reports, and internal systems.

Teams end up relying on careful, repetitive cross-checking across documents. This work is cognitively draining and judgment-sensitive. Isolating this capability allows comparisons to be handled consistently, while interpretation and decisions remain with human underwriters.

Below, we explain where this skill fits in the underwriting workflow, what it can and cannot do, and where human judgment takes over. If you want to explore it hands-on, you can try it later on this page.

Understanding This Skill

Why This Skill Exists

In private lending, addresses sit at the center of many risk signals. Underwriters routinely compare addresses across loan applications, bank statements, IDs, entity documents, purchase contracts, and third-party reports.

This work is manual and repetitive, but it requires sustained attention. Small discrepancies — a unit number, a formatting difference, or a shared address across parties — are easy to miss when teams are working under time pressure.

Where This Skill Fits

This skill fits into moments where borrower documents are already being reviewed, but attention is spread across multiple checks.

Typically used during:
  • Initial underwriting
  • Reviewing conditions
  • Pre-funding checks
Typically used by:
  • Underwriters
  • Credit analysts
  • Risk & QA teams

What This Skill Can Do

Identifies address-level inconsistencies and risk signals.

  • Find and compare addresses across loan applications and supporting documents
  • Normalize superficial differences in how addresses are written
  • Flag mismatches relative to the anchor address on the application
  • Surface address patterns that may indicate related-party or occupancy risk

What This Skill Cannot Do

This skill does not:

  • Make credit, approval, or exception decisions
  • Determine borrower intent or assign fault
  • Replace underwriting, legal, or compliance judgment
  • Guarantee fraud detection or eliminate risk
This skill supports human judgment. Accountability always remains with the underwriting and credit teams.

What You Provide (Inputs)

This skill works with information that teams already review, including:

  • Loan application data (anchor address)
  • Supporting borrower, property and entity documents (PDFs, images, spreadsheets)

What This Skill Produces (Outputs)

This skill produces structured findings for human review, including:

  • A consolidated view of all addresses found across documents
  • Indications of matches, mismatches, and anomalies
  • Contextual notes highlighting risk signals
  • A concise summary to support underwriting review

When Humans Must Take Over

Human review is required when:
  • Address discrepancies materially affect credit, collateral or occupancy risk
  • Patterns suggest potential related party relationships
  • Address usage conflicts with the stated loan purpose
  • Exceptions, overrides or judgement calls are being considered
Final accountability always remains with the underwriting and credit teams.

Monitoring and Oversight

This skill can make mistakes. It can surface false positives when legitimate address variations exist (unit formatting, recent changes, mailing vs property addresses).

It can also miss risk when source documents are consistently wrong in the same way.

Findings that carry material risk should always be reviewed by a human before any decision or action is taken.

Borrow This Capability

Explore this skill before deciding whether, and how, it fits into your workflow.

ChatGPT
Explore

Try the Skill in GPT

Explore how the Address Integrity Skill reviews documents and surfaces risk signals using sample or real files.

Open in ChatGPT

This is a lightweight, exploratory way to understand the logic behind the skill. It's useful for learning and validation, not for production use.

Gemini
Explore

Try the Skill in Gemini

See the same skill logic applied using a different foundation model.

Open in Gemini

This helps illustrate what stays consistent across models — and what changes — when the same capability is implemented on different LLMs.

Experience

Try a Purpose-Built VAL Skill

Experience this skill as a governed agent or co-pilot, with stronger safeguards, task-specific logic, and supporting tools.

Try VAL Skill

While LLMs are still used under the hood, purpose-built skills combine them with document processing, rules, and oversight — making the capability more reliable than using a general-purpose model directly.

Discuss

See It in Your Workflow

Understand how the same logic integrates into lending systems as a governed agent — including handoffs, human review points, and auditability.

View Workflow Integration

If this looks relevant to your workflow, we can walk through it together.

Frequently Asked Questions