AI Prompts for Insurance Underwriters: Risk, Pricing, Portfolio

Published: July 2026 | 8 min read

Insurance underwriting requires balancing risk assessment, pricing accuracy, regulatory compliance, and portfolio management. AI prompts can accelerate the analytical work — risk classification, loss projection, pricing models, treaty documentation — while keeping underwriting judgment in the hands of licensed professionals.

Below are production-ready AI prompts for insurance underwriters covering risk assessment, pricing, portfolio management, and compliance. These are adapted from Skillent's Finance AI Prompt Library.

Risk Assessment & Classification

1. Property Risk Assessment Worksheet

Role: You are a property insurance underwriter evaluating a commercial property risk.
Property: [type — office/warehouse/retail/manufacturing], [construction type], [year built]
Location: [address placeholder, distance to fire hydrant, fire protection class]
Input: [Square footage, occupancy details, hazards, sprinkler system, alarm system, construction details]
Task: Perform property risk assessment.
Evaluate:
1. Construction class (frame/joisted masonry/non-combustible/fire resistive)
2. Occupancy hazard (low/moderate/high)
3. Protection class (PPC rating based on fire protection)
4. Exposure (external — adjacent properties, internal — fire loads)
5. Construction quality and maintenance
6. Sprinkler coverage (% of building, type, certification)
7. Alarm system (burglary, fire, central station)
8. Roof condition and age
9. Electrical system (age, capacity, updates)
10. Plumbing (age, material, leak history)
11. HVAC (age, condition, maintenance records)
Recommend: Accept/Decline, coverage limits, deductible, key exclusions
Format: Property risk survey with rating and pricing recommendation.
Include: Loss control recommendations to reduce premium.

2. Commercial General Liability Risk Assessment

Role: GL underwriter evaluating a commercial liability risk.
Business: [type], [operations summary], [annual revenue], [employee count]
Input: [Prior claims history (5 years), safety program details, subcontractor usage, product information]
Task: Perform GL risk assessment.
Evaluate:
1. Operations hazard class (LOW/MOD/HIGH based on NAICS and operations)
2. Premises liability exposure (public access, parking, elevators)
3. Products/completed operations exposure
4. Contractual liability transfer (indemnification agreements)
5. Subcontractor exposure (certificates of insurance on file?)
6. Employee injury exposure overlap with workers comp
7. Prior claims analysis (frequency, severity, trend)
8. Safety program quality (formal/informal, training, enforcement)
9. Liquor liability exposure (if applicable)
10. Cyber liability exposure (if operations involve data)
Recommend: Accept/Decline, limits, deductible, premium class, exclusions
Format: GL risk assessment with classification and pricing basis.
Include: Required endorsements and manuscript coverage recommendations.

3. Workers' Compensation Risk Classification

Role: Workers' comp underwriter performing risk classification.
Employer: [business type], [operations summary], [state], [employee count by role]
Input: [Job descriptions, payroll by class, prior loss history (5 years), safety program]
Task: Classify WC risk and recommend structure.
For each employee class:
1. Class code (NCCI or state-specific)
2. Payroll (remuneration basis — include overtime rules per state)
3. Class rate (base rate for jurisdiction)
4. Experience modification factor (if applicable — explain calculation)
5. Hazard group (if applicable)
Determine:
6. Total estimated premium (sum of class code payrolls × rates × mod)
7. Minimum premium check
8. Optional coverages (disease, stop-gap, WA/monopolistic state considerations)
9. Deductible options and credit
10. Schedule modification (debit/credit based on safety program)
Format: WC classification worksheet with premium calculation.
Include: Loss run analysis summary and mod factor trend.
Flag: Any class codes that are commonly misclassified (e.g., clerical vs. outside sales).

Pricing & Actuarial

4. Rate Adequacy Analysis

Role: Underwriter performing rate adequacy analysis for a book of business.
Line of business: [type], state: [jurisdiction], period: [analysis year]
Input: [Earned premium, incurred losses, LAE, earned exposures, current rates]
Task: Perform rate adequacy analysis.
Calculate:
1. Loss ratio = incurred losses / earned premium
2. Expense ratio = expenses / written premium
3. Combined ratio = loss ratio + expense ratio
4. Target loss ratio (based on expense load + profit provision)
5. Indicated rate change = (actual loss ratio / target loss ratio) - 1
6. Loss cost = (incurred losses + LAE) / earned exposures
7. Pure premium = loss cost × frequency × severity
8. Credibility (Z) = based on exposure volume (limited fluctuation credibility)
9. Credibility-weighted indicated change
Format: Rate adequacy analysis with step-by-step calculations.
Include: Recommendations — file for rate increase, decrease, or maintain.
Reference: Actuarial Standards of Practice (ASOP) for rate review methodology.

5. Loss Development Analysis

Role: Underwriter performing loss development analysis.
Line: [type], accident years: [list], maturity: [current evaluation]
Input: [Loss triangles — paid and incurred by accident year and development period]
Task: Analyze loss development patterns.
Calculate:
1. Age-to-age loss development factors (LDFs) for each interval
2. Selected LDFs (based on volume-weighted average, simple average, or geometric mean)
3. Tail factor (for ultimate development)
4. Cumulative development factors (CDF) from each age to ultimate
5. Projected ultimate losses by accident year
6. IBNR (incurred but not reported) = projected ultimate - reported to date
7. Bornhuetter-Ferguson estimate (if prior expectation available)
Format: Loss development triangle with LDF analysis and IBNR projection.
Include: Commentary on unusual development patterns (upward development, changes in claim handling).

6. Exposure Rating Worksheet

Role: Underwriter using exposure rating for a large commercial risk.
Insured: [company], line: [property/GL/auto/WC], exposure base: [payroll/sales/area/units]
Input: [Exposure amount, ILF factor, rate manual reference, hazard group]
Task: Perform exposure rating calculation.
Steps:
1. Determine exposure base (payroll for WC, sales for GL, building value for property)
2. Apply rate per $100 of exposure (manual rate)
3. Apply ILF (increasing limits factor) based on limits requested
4. Apply experience modification (if applicable)
5. Apply schedule rating credits/debits (±25% typically)
6. Apply deductible credit (if applicable — use deductible discount factor)
7. Apply optional coverage loadings
8. Calculate minimum premium
9. Verify against rate adequacy rules
Format: Exposure rating worksheet with each step and final premium.
Include: Comparison to loss-rated approach (if data available).

Portfolio Management

7. Underwriting Portfolio Dashboard

Role: Underwriting manager building a portfolio dashboard.
Book of business: [line], [geographic scope], [period]
Input: [Premium by segment, loss ratios by segment, policy counts, exposures]
Task: Create a portfolio dashboard specification.
Metrics:
1. Written premium (YTD, trend vs prior year)
2. Earned premium (YTD)
3. Loss ratio (accident year and calendar year)
4. Combined ratio
5. Policy count and retention rate
6. New business vs renewal mix
7. Average premium per policy
8. Coverage limit distribution
9. Geographic concentration
10. Industry concentration (for commercial lines)
For each metric: Current value, trend (3-month/12-month), target, status (green/yellow/red)
Format: Dashboard specification with alert thresholds.
Include: Drill-down dimensions (by agent, territory, class, size).

8. Reinsurance Treaty Summary

Role: Underwriter documenting reinsurance treaty structure.
Treaty type: [quota share / surplus share / excess of loss / facultative]
Line of business: [type], effective period: [dates]
Task: Create a treaty summary for underwriting reference.
Include:
1. Treaty type and structure
2. Retention and cession limits (min/max)
3. Commission structure (original, profit commission, sliding scale)
4. Reinsurer participation shares
5. Coverage trigger (loss occurrence vs. risks attaching)
6. Definition of subject premium
7. Reporting requirements (borard statements, bordereaux)
8. Commutation/cancelation provisions
9. Event limits and hours clauses (for XOL treaties)
10. Aggregate limit (if applicable)
Format: One-page treaty reference document for underwriters.
Include: How this treaty affects individual risk underwriting authority and capacity.

Compliance & Documentation

9. Underwriting File Documentation Checklist

Role: Underwriting manager creating a file documentation standard.
Line: [type], authority level: [individual underwriter authority]
Task: Create a comprehensive underwriting file documentation checklist.
Required documentation:
1. Application/declaration page
2. Loss history (5 years minimum, from insured or CLUE/ISO)
3. Inspection report (if required by file plan)
4. Credit score (if used and permitted by state)
5. Vehicle/property valuation
6. Underwriting rationale memo (why this risk was accepted at this price)
7. Any deviations from guidelines (with supervisory approval)
8. Reinsurance cession records (if applicable)
9. Premium finance documentation (if applicable)
10. Policy issuance checklist
11. Compliance review (state-specific forms, endorsements, rate filing compliance)
For each item: Required, conditional, or optional, with source and retention period
Format: Documentation checklist with sign-off.
Include: Audit triggers — what gets flagged in QA review if missing.

10. Rate Filing Compliance Memo

Role: Underwriter preparing a rate filing compliance memo.
State: [jurisdiction], line: [type], filing type: [new rate / rate change / rule change]
Task: Create a rate filing compliance memo.
Include:
1. Filing type (rate/rule/form)
2. Effective date and implementation timeline
3. Rate change magnitude (±% and $ impact)
4. Supporting actuarial documentation (loss ratio analysis, expense provision)
5. Rate adequacy certification
6. State-specific requirements (prior approval vs file-and-use vs use-and-file)
7. Required forms and transmittals
8. Consumer impact statement (if required)
9. Competitive rate comparison
10. Filing status tracking
Format: Compliance memo for internal regulatory team.
Include: Potential regulatory objections and preemptive responses.

Special Underwriting Situations

11. Catastrophe Exposure Assessment

Role: Underwriter assessing catastrophe exposure for a property portfolio.
Geographic scope: [states/regions], peril: [wind/hurricane/earthquake/flood/wildfire]
Input: [Property values by location, construction types, deductible levels, policy limits]
Task: Perform catastrophe exposure assessment.
Analysis:
1. Total insured value (TIV) by catastrophe zone
2. Probable Maximum Loss (PML) — 1-in-100 and 1-in-250 year events
3. Concentration mapping (ZIP code level exposure)
4. Aggregate limits exposure (across perils)
5. Reinsurance recovery analysis (how much is ceded vs retained)
6. Deductible analysis (are deductibles adequate for CAT events?)
7. Construction class distribution (frame properties = higher risk)
8. Year-built distribution (pre-code construction = higher risk)
9. Mitigation credits (shutters, reinforced roofing, elevation)
Format: CAT exposure analysis with maps/tables.
Include: Recommendations for capacity limits, exclusions, or facultative reinsurance.
Flag: Any ZIP codes exceeding aggregate TIV thresholds for retention.

12. Underwriting Authority Matrix

Role: Underwriting manager defining underwriting authority levels.
Task: Create an underwriting authority matrix.
Dimensions:
1. By underwriter level (trainee / staff / senior / manager / officer)
2. By transaction type (new business / renewal / endorsement / rewrite)
3. By premium range ($ tiers)
4. By coverage limits (per line)
5. By risk quality (preferred / standard / substandard)
6. By deviation authority (within guidelines / deviation / non-standard)
For each combination:
- Approval authority (auto-approve / refer / require supervisor)
- Documentation required
- Exception process
Format: Authority matrix grid.
Include: Audit/quality assurance review requirements for each authority level.
Note: Authority levels must align with corporate delegation of authority policy and state licensing requirements.

Best Practices for Underwriting AI Prompts

1. Always specify the line of business — property, GL, auto, WC, marine, and E&S have fundamentally different rating approaches

2. Include the jurisdiction — state insurance regulations vary dramatically. Rate filing rules differ by state

3. Never input actual policyholder data — use placeholders for names, addresses, and account numbers

4. Reference the rating manual — ISO, NCCI, AAIS, or company-specific manuals. AI output must match the manual

5. Verify all rate calculations — underwriting is regulated. An error in rate calculation can trigger regulatory action. Every calculation must be verified against the manual

For more risk assessment resources, see our ChatGPT prompts for credit analysts and AI prompts for cash flow analysis.

How to Use These Prompts Effectively

Getting the most out of AI prompts for financial work requires a structured approach. Here's how to integrate these prompts into your workflow:

1. Start with Clean Data

Before running any prompt, ensure your source data is organized and accurate. AI output is only as good as the input. Gather your trial balances, general ledger exports, prior-period statements, and any supporting documentation. The more structured your input data, the more useful the AI output will be.

2. Iterate and Refine

Don't expect perfect output on the first run. Start with the prompt as written, review the output, then refine. Add context about your specific industry, company size, accounting software, or reporting requirements. The prompts above are templates — adapt them to your exact situation.

3. Always Verify Against Source Documents

AI can make calculation errors, cite outdated tax rates, or miss nuances in your chart of accounts. Every piece of output that contains a number, a tax rate, a deadline, or a regulatory citation must be verified against primary sources. This is not optional — it's a professional responsibility.

4. Build a Prompt Library for Your Organization

Save the prompts that work best for your team. Document what you changed, what input data you provided, and what output format worked best. Over time, you'll develop a customized prompt library that reflects your organization's specific accounting practices, reporting requirements, and industry context.

5. Use AI for Drafts, Not Final Products

The most effective pattern is to use AI prompts to generate first drafts — whether that's a budget model, a variance commentary, a reconciliation summary, or a tax memo. Then apply your professional judgment to review, correct, and finalize. AI accelerates the draft phase; it doesn't replace the review phase.

Why These Prompts Matter for Finance Teams

Finance and accounting teams are under increasing pressure to do more with less. Month-end close cycles are compressing. Reporting requirements are expanding. Audit expectations are rising. And the demand for real-time financial insight — not just historical reporting — is growing across every organization.

AI prompts like these address that pressure by automating the repetitive, pattern-based work that consumes hours of professional time each week. A variance analysis that used to take three hours can be drafted in twenty minutes. A budget model that used to start from a blank spreadsheet can begin with a structured framework. A reconciliation narrative that used to be written line by line can be generated from structured input.

The goal isn't to replace financial professionals — it's to free them from mechanical work so they can focus on analysis, strategy, and advisory roles. The most successful finance teams using AI are the ones that treat it as a junior analyst: capable of producing solid first drafts quickly, but requiring review and guidance from senior professionals before the work is finalized.

For more finance prompts across specialties, explore the full Skillent Finance Prompt Library.

Common Mistakes to Avoid with Finance AI Prompts

After working with hundreds of finance teams adopting AI tools, we've identified the most common — and most costly — mistakes professionals make when starting to use AI prompts in their workflow.

Mistake 1: Treating AI Output as Final

The single most dangerous mistake is treating AI-generated financial analysis, tax calculations, or reporting narratives as final without review. AI can produce output that looks correct — proper formatting, confident tone, plausible numbers — while containing subtle errors. A misplaced decimal, an outdated tax rate, a misapplied formula, or a misinterpreted accounting standard can turn a useful draft into a professional liability. Always treat AI output as a first draft that requires professional review.

Mistake 2: Inputting Sensitive Client Data

Entering client names, social security numbers, full financial statements, or proprietary business data into an AI tool without understanding the tool's data handling policy is a serious risk. Some AI tools retain input data for training purposes. Some may expose data through API responses. Before using any AI tool with real client data, verify the tool's privacy policy, data retention practices, and whether you can disable training on your inputs. When in doubt, use anonymized or placeholder data.

Mistake 3: Using Generic Prompts for Specialized Tasks

A prompt designed for general financial analysis won't produce good output for a specialized task like transfer pricing documentation or ASC 842 lease accounting. The prompts in this guide are tailored to specific finance functions. Using a generic "analyze this financial data" prompt when you need a specific deliverable — like a variance commentary for a board meeting — will produce generic, unhelpful output. Match the prompt to the task.

Mistake 4: Ignoring Context

AI tools don't know your company, your industry, your accounting policies, or your reporting requirements unless you tell them. A prompt that says "analyze this balance sheet" without context will produce generic analysis. A prompt that says "analyze this balance sheet for a SaaS company using ASC 606 revenue recognition, with a fiscal year ending June 30, reporting to a Board of Directors" will produce targeted, useful analysis. Context is the difference between a parlor trick and a professional tool.

Mistake 5: Not Iterating

The first run of a prompt rarely produces the best possible output. The most effective users run a prompt, review the output, refine the prompt based on what was missing or wrong, and run it again. This iterative process — two or three rounds of refinement — typically produces dramatically better results than the first run. Build time for iteration into your workflow.

Mistake 6: Forgetting the Human Review Layer

AI is a tool, not a team member. Every piece of output that will be seen by a client, an auditor, a regulator, or a board member must be reviewed by a qualified professional who understands the context and takes responsibility for the final product. This isn't just best practice — in many financial roles, it's a regulatory requirement. The human review layer is non-negotiable.

Quick Start Guide: Your First Week with Finance AI Prompts

If you're new to using AI prompts in your finance workflow, here's a practical week-one plan:

Day 1-2: Experiment with Low-Stakes Prompts

Start with prompts that produce output you can immediately verify. Try the variance analysis prompt with last month's actual vs. budget data. Try the reconciliation summary prompt with a small subset of accounts. The goal is to get comfortable with the prompt structure, see the quality of output, and calibrate your expectations.

Day 3-4: Integrate into a Real Workflow

Pick one recurring task — maybe the monthly variance commentary, or the weekly cash flow summary — and use the relevant prompt to produce a first draft. Then complete the task as you normally would, comparing the AI-assisted version to your traditional approach. This will show you where AI helps, where it falls short, and how to integrate it effectively.

Day 5: Customize and Save

Take the prompt you used during the week and customize it for your specific situation. Add your industry, your accounting software, your reporting format, your typical account structure. Save the customized prompt. This becomes the starting point for next week's work — and the beginning of your personal prompt library.

Ready to access the full library of 190,000+ professional AI prompts? Get Skillent Pro for $9/month and unlock prompts for every finance function.

Advanced Pro Tips for Finance AI Prompts

Chain Prompts for Complex Analysis

For complex deliverables, chain multiple prompts together. Start with a data extraction prompt to pull key figures from source documents. Feed that output into an analysis prompt. Then feed the analysis into a formatting prompt that structures the output for your final deliverable. This three-step chain — extract, analyze, format — produces better results than trying to do everything in one prompt. Each step can be refined independently, and intermediate output can be reviewed before proceeding.

Build a Prompt Version History

As you refine prompts over time, keep a version history. Save the original prompt, note what changes you made and why, and track which versions produced the best output. This creates an institutional knowledge base that your entire team can benefit from. When a new team member joins, they can start with your refined prompts rather than the generic originals.

Create Industry-Specific Variants

The prompts in this guide are written for general finance and accounting use. But a CPA serving manufacturing clients needs different prompts than one serving SaaS startups. Create industry-specific variants of each prompt — add industry-specific tax considerations, accounting standards, reporting formats, and common transaction types. Industry-specific prompts produce dramatically better output than generic ones.

Pair AI with Spreadsheet Skills

The most powerful workflow combines AI prompts with advanced spreadsheet skills. Use AI to generate the analysis framework, the commentary, and the narrative. Use Excel or Google Sheets for the actual calculations — where formulas are auditable and verifiable. Copy AI-generated commentary into spreadsheet cells alongside the numbers. This hybrid approach leverages the strengths of both tools: AI for language, spreadsheets for math.

Document Your Review Process

For audit and compliance purposes, document your AI review process. Note which AI tool was used, which prompt was used, what the output was, and what changes you made during review. This documentation demonstrates professional diligence and provides a clear audit trail. Some firms now require this documentation as part of their quality control procedures.

Real-World Applications: Finance AI Prompts in Practice

Month-End Close Acceleration

A mid-market manufacturing company reduced its month-end close from 12 days to 7 days by integrating AI prompts into its close process. The controller used AI to draft variance explanations for each major account, generate the monthly close memo, and prepare the first draft of the consolidated income statement commentary. The AI-generated drafts were reviewed and finalized by the accounting team, but the drafting time was cut by approximately 60%. The key was using structured prompts that included the company's chart of accounts, materiality thresholds, and standard commentary format.

Audit Readiness for a Growing Startup

A Series B startup facing its first external audit used AI prompts to prepare. The finance team generated a complete PBC (prepared by client) list using an audit preparation prompt, then used reconciliation prompts to clean up balance sheet accounts before the auditors arrived. When the auditors requested supporting schedules, the team used AI to draft the schedules from source data. The audit partner noted that the startup was "unusually well-prepared" — a direct result of using AI prompts to structure the preparation process.

Tax Season Volume Management

A solo CPA practitioner serving 200+ individual and small business clients used AI prompts to manage tax season volume. Client intake questionnaires were generated by AI. K-1 matching prompts processed partnership distributions. Tax research memos for unusual situations were drafted by AI and reviewed by the CPA. The practitioner estimated that AI prompts saved 15-20 hours per week during peak season — equivalent to adding a part-time staff member without the overhead.

Cash Flow Crisis Management

A retail company facing a liquidity crisis used AI prompts to build a 13-week cash flow forecast, stress-test it against multiple scenarios, and prepare a presentation for its bank. The AI-generated forecast identified a week where the company would breach its loan covenant — three weeks before it happened. The company was able to negotiate a covenant waiver with the bank in advance, avoiding a default that would have triggered an interest rate spike.

Budget Season Efficiency

A nonprofit organization with 15 departments used AI prompts to streamline its annual budget process. Each department head received a prompt template to input their assumptions, prior-year actuals, and planned initiatives. The finance team used AI to consolidate the departmental inputs, identify gaps, and generate the first draft of the organization-wide budget. The budget was completed in 3 weeks instead of the typical 8, freeing the finance team to focus on scenario analysis and board presentation preparation.