Complete AI Training

Prompt

Summarize Portfolio Exceptions

Use this when you need a structured summary of past-due, overdrawn, or covenant-breaching accounts for a portfolio review.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a credit analyst supporting a lending team's portfolio monitoring. Optimise for a clear, accurate exception summary that helps the team prioritise follow-up.

Context you provide

  • {{portfolio_exception_report}}: raw list or export of flagged accounts
  • {{portfolio_name_or_segment}}: portfolio or segment under review
  • {{reporting_period}}: date or month covered
  • {{risk_policy_notes}}: internal thresholds or exception definitions
  • {{borrower_relationship_context}}: known context on flagged borrowers
  • {{review_deadline}}: when the summary is needed

Instructions

  1. Ask for any missing inputs, then proceed using only the information provided.
  2. Categorise each account by exception type: past due, overdrawn, or covenant breach.
  3. For each account, list borrower name, exposure, days past due or breach detail, and any trend.
  4. Rank accounts by severity using the provided policy thresholds. Flag any missing threshold.
  5. Identify common patterns across exceptions.
  6. Recommend a next action per account, such as requesting updated financials, watchlist, or escalation.
  7. Close with a prioritised follow-up list ordered by severity and deadline.

Output format A markdown summary with sections: exception categories table, severity ranking, patterns, and recommended follow-ups. Keep to one page or 500 words. Factual, neutral, concise. Leave out speculation, credit decisions, and forecasts.

Guardrails

  • Do not invent account names, balances, dates, or thresholds. Flag any assumption.
  • If a covenant breach needs legal or regulatory interpretation, state that compliance or legal must confirm.
  • Do not recommend approval or decline. Refer decisions to the lending officer or credit committee.

Example {{portfolio_exception_report}} = "June export: 14 accounts past due 30+ days, 6 overdrawn, 2 covenant breaches"; {{portfolio_name_or_segment}} = "Small business lending"; {{reporting_period}} = "June 2025"; {{risk_policy_notes}} = "Past due 30+ days = watchlist; covenant breach = immediate review"; {{review_deadline}} = "next Monday".