The Playbook: What to Do

How to Do It

Treat every AI output as an unverified first draft. Cross-reference policy language, regulatory citations, actuarial data, and industry statistics against primary sources. Never submit, share, or act on AI-generated content without verification.

Step-by-Step

1. Run your prompt through the AI tool and review the complete output.
2. Identify every factual claim, regulation reference, policy citation, and data point.
3. Verify each against primary sources: policy documents, regulatory databases (NAIC, state DOI sites), industry reports.
4. Check that any statistics or market data cited actually exist and are current.
5. Confirm that the AI's interpretation of policy language matches industry-standard interpretation.
6. Document your verification process for compliance records.

Recommended Tools

How to Do It

Use AI to generate initial summaries, identify relevant regulations, survey market trends, and draft research outlines. Then deepen, verify, and refine with primary sources and your professional judgment.

Step-by-Step

1. Define your research question clearly before engaging AI.
2. Use AI to generate an initial overview or summary of the topic.
3. Ask AI to identify relevant regulations, industry standards, and key considerations.
4. Use the AI output as a roadmap — it tells you where to look, not what to conclude.
5. Conduct your own verification and analysis using primary sources.
6. Apply your professional judgment to form conclusions and recommendations.

Recommended Tools

How to Do It

Draft and implement a comprehensive AI use policy that covers approved tools, data handling requirements, verification procedures, disclosure obligations, and accountability frameworks. Review and update the policy quarterly.

Step-by-Step

1. Audit current AI use across your organization — who is using what, and for which tasks.
2. Identify data sensitivity levels: what can and cannot be entered into AI tools.
3. Define approved AI tools and platforms based on security, privacy, and compliance requirements.
4. Establish verification protocols for different output types (internal analysis, customer-facing, regulatory submissions).
5. Create clear accountability: who is responsible for AI-assisted decisions?
6. Set training requirements for all staff who use AI tools.
7. Schedule quarterly policy reviews to address new tools, regulations, and lessons learned.

Recommended Tools

How to Do It

Identify all regulatory disclosure requirements in your jurisdictions. Develop standard disclosure language. Integrate disclosure into your workflows so it happens automatically, not as an afterthought.

Step-by-Step

1. Research AI disclosure requirements in every jurisdiction where you operate.
2. Identify which business processes use AI in ways that may require disclosure.
3. Draft clear, plain-language disclosure statements for different contexts.
4. Integrate disclosure into underwriting, claims, and customer communication workflows.
5. Train staff on when and how to disclose AI use.
6. Document all disclosures for compliance records.

Recommended Tools

How to Do It

Use AI to create initial drafts of common insurance documents. Always review, refine, and personalize the output. The AI writes the first draft; you write the final draft.

Step-by-Step

1. Identify the document type and its audience (internal, customer, regulatory).
2. Write a detailed prompt using the CRAFT framework: Context, Role, Action, Format, Tone.
3. Generate the first draft using AI.
4. Review the draft critically: verify facts, check policy references, ensure accuracy.
5. Refine the language, add your professional judgment and specific case details.
6. Have the final version reviewed by a colleague or supervisor as appropriate.

Recommended Tools

How to Do It

Use AI to perform initial document review, extract structured data from unstructured documents, and identify key provisions or anomalies. Always review AI-extracted data for accuracy, especially for coverage-critical provisions.

Step-by-Step

1. Identify the documents to review and the specific information you need extracted.
2. Use AI to extract key data points: dates, amounts, coverage limits, exclusions, conditions.
3. Ask AI to summarize the document and flag any unusual provisions or potential issues.
4. Cross-reference AI-extracted data against the source documents.
5. Use AI to compare documents (e.g., current policy vs. renewal, competing quotes).
6. Document your review process and any corrections to AI-extracted data.

Recommended Tools

How to Do It

Start collecting and refining prompts that produce consistently good results for your work. Organize them by task type, insurance sector, and use case. Share successful prompts with your team.

Step-by-Step

1. Identify your 10 most repetitive insurance tasks (the ones you do weekly or daily).
2. Write an initial prompt for each using the CRAFT framework.
3. Test each prompt multiple times and refine based on the results.
4. Save the refined prompts in an organized, searchable format.
5. Include notes on what works, what doesn't, and any context needed.
6. Share your library with colleagues and incorporate their feedback.
7. Review and update your library monthly as AI tools improve.

Recommended Tools

How to Do It

Dedicate time each week to staying current on AI developments relevant to your insurance practice. Follow trusted sources, participate in industry discussions, and regularly reassess your AI tools and workflows.

Step-by-Step

1. Subscribe to 3-5 trusted sources covering AI in insurance (see recommendations below).
2. Dedicate 30 minutes per week to reading AI industry updates.
3. Follow regulatory developments in your jurisdictions (NAIC bulletins, state DOI updates).
4. Attend at least one AI-focused insurance event or webinar per quarter.
5. Reassess your AI tools and workflows quarterly — what worked three months ago may be outdated.
6. Share relevant updates with your team and integrate learnings into your practice.

Recommended Tools

How to Do It

Develop and implement a structured AI training program for your team. Cover fundamentals, hands-on practice, compliance requirements, and your organization's AI policy. Make training ongoing, not a one-time event.

Step-by-Step

1. Assess your team's current AI knowledge and comfort levels.
2. Develop a training curriculum covering AI fundamentals, your AI policy, and practical skills.
3. Start with hands-on exercises using approved tools on real (or realistic) insurance tasks.
4. Cover compliance requirements specific to your jurisdictions and lines of business.
5. Create mentorship pairs: team members with AI experience paired with beginners.
6. Schedule regular knowledge-sharing sessions where team members share AI tips and lessons learned.
7. Track adoption metrics and adjust training based on what's working.

Recommended Tools

How to Do It

Begin your AI adoption with low-risk, easily verifiable tasks. Build your skills progressively before applying AI to high-stakes insurance decisions. Document what you learn and share it with your team.

Step-by-Step

1. Identify 3-5 low-risk tasks where you can experiment safely (internal summaries, research, drafting).
2. Use AI for these tasks alongside your normal workflow (do it both ways to compare).
3. Evaluate: Was the AI output helpful? Accurate? Time-saving?
4. Refine your approach based on what you learn.
5. Gradually expand to moderate-risk tasks with appropriate verification.
6. Only apply AI to high-risk tasks (regulatory submissions, coverage decisions, customer-facing outputs) after you have strong verification processes in place.

Recommended Tools