Risk AssessmentHigh riskComparison recommended

AI Housing Assistance Subsidy Integrity Audit Playbook

A public housing authority administers 3,400 Housing Choice Vouchers. An HUD monitoring review found that 14% of units have rents above the payment standard, 8% have households with unreported household members, and the authority has not conducted required annual inspections on 22% of units.

When to use this playbook

  • Use this playbook when the decision looks like the situation above: A public housing authority administers 3,400 Housing Choice Vouchers.
  • It is a fit when you have source files in hand and need a structured, reviewable analysis — not a generic chat answer about "Housing Assistance Subsidy Integrity Audit".
  • Do not use it as a substitute for licensed, legal, clinical, or authorized official judgment in the domain.

What you'll need

  • HCV program data (all 3,400 vouchers: rent, household composition, inspection dates)
  • HUD payment standard by bedroom size and zip code
  • HUD monitoring letter and findings
  • HAP agreements for flagged units
  • HUD CFR 24 Part 982 requirements

Attachments: Documents (Documents)

The Prompt

You are a housing authority compliance officer responding to an HUD monitoring review with three findings. I am attaching:

Work only from the attached source files. If a conclusion is not supported, say so.

Produce:
1. For the 14% with rents above payment standard: calculate total overpayment to landlords and whether excess was improperly absorbed by tenants.
2. For the 8% with unreported household members: calculate the subsidy impact and financial exposure.
3. For the 22% with missed inspections: assess the HUD penalty exposure and tenant health and safety risk.
4. Develop the corrective action plan for each finding with HUD's expected timelines.
5. Tell me the overpayment recovery process from landlords and the authority's fair housing obligations when correcting household composition issues.

Call out where independent models are likely to disagree, and list follow-up documents a reviewer should request.

What to expect

  • Rent overpayment calculation
  • Unreported household member financial exposure
  • Inspection failure penalty and safety risk
  • Corrective action plan per finding
  • Overpayment recovery and fair housing obligations

Review before you act

  • Validate this output against source files before relying on it: For the 14% with rents above payment standard: calculate total overpayment to landlords and whether excess was improperly absorbed by tenants.
  • Validate this output against source files before relying on it: For the 8% with unreported household members: calculate the subsidy impact and financial exposure.
  • Validate this output against source files before relying on it: For the 22% with missed inspections: assess the HUD penalty exposure and tenant health and safety risk.
  • Validate this output against source files before relying on it: Develop the corrective action plan for each finding with HUD's expected timelines.
  • Confirm every cited figure, date, counterparty, or requirement against the attached originals — models compress and can drop a qualifier.
  • Treat disagreement between models as a review item, especially on classification, materiality, and recommended next action.
  • Do not authorize an operational, clinical, legal, credit, or enforcement action solely because the models agree.

Why compare models on this

For Housing Assistance Subsidy Integrity Audit, running the same attachments across independent models is useful because the hard part is classification and completeness, not fluency. The workflow is already designed to surface rent overpayment calculation; unreported household member financial exposure; inspection failure penalty and safety risk; corrective action plan per finding. Those are comparison artifacts — they only exist if more than one model runs. Models split on trafficking versus legitimate high-volume redemption, and on identity-fraud versus data error. Divergence is a reason to pull the case file, not to auto-disqualify.

Public BenefitsCompliance and Integrity OperationsRisk AssessmentHighDocuments

See governed multi-model AI on your own prompt

Compare GPT-5, Claude, and Gemini side by side, with human review and a decision record built in.