An AI chief of staff can help your insurance agency draft communications, prepare for prospect calls, analyze reports, and organize daily priorities. Better results start with clear instructions, useful reference materials, and human review.
During Catalyit’s GetLYIT 2 session, “Hire Your AI Chief of Staff,” Matt Smith, an AI consultant for Catalyit, demonstrated five practical ways to use Claude in agency operations.
His core message: Teach AI how your agency works, then give it a specific job.
What is an AI chief of staff?
An AI chief of staff is an assistant configured to support recurring work using your agency’s instructions, reference materials, and approved tools.
Smith demonstrated five jobs:
Job
Agency application
Apply agency context
Follow your voice, preferences, and boundaries
Draft communications
Turn renewal notes into client letters
Follow repeatable processes
Prepare prospect research and submission materials
Analyze agency data
Review retention, renewals, and carrier concentration
Organize priorities
Summarize relevant email and calendar activity
Your team supplies direction, checks the results, and makes the decisions.
1. Teach AI about your agency
Tell AI whom you serve, how you communicate, and what it should never assume.
Smith demonstrated this with a fictional client questioning a commercial auto renewal increase. Claude first produced a general reply. After he added agency instructions, the response became more direct and better matched the agency’s voice.
Useful instructions include:
Your location, customer focus, and specialties.
Your preferred tone and response length.
How to handle missing information.
Statements to avoid.
How to present the next step.
An agency might request plain language, short sentences, and a clear action at the end of each email. It might also instruct AI to flag uncertain policy terms rather than guess.
Check every commitment. A draft may promise a callback or review your team has not scheduled.
2. Build a dedicated renewal workspace
A renewal workspace keeps approved instructions, examples, and templates together.
Smith demonstrated a Claude Project called “Renewals Desk.” It included letterhead, sample letters, guidance on explaining rate changes, and a standard letter structure.
The team could then provide notes and request a draft without repeating the setup.
Keep each workspace focused. Separate renewals, submissions, producer coaching, and agency operations when they require different instructions.
3. Turn recurring tasks into skills
A skill gives AI a repeatable process for a specific task.
A project holds the context. A skill defines the steps. Claude supports reusable skills for recurring work. Read Claude’s skills overview.
Smith demonstrated a lead preparation skill that:
Researched a prospect through public sources.
Summarized the business.
Identified potential exposures.
Suggested coverage topics to discuss.
Prepared first-call questions.
Produced a structured preparation document.
The instructions required assumptions to be marked for confirmation. Potential exposures became questions to investigate, not coverage promises.
For your first skill, choose a task you explain to every new hire. Document the steps, expected output, and required checks.
4. Ask better questions of your agency data
AI can help turn an agency report into information your team can investigate and act on.
Smith used a fictional agency management system export containing about 300 policies. He requested a dashboard showing:
Retention by producer.
Premium by carrier.
Renewals due within 90 days.
Accounts with one line of coverage.
These views can highlight upcoming workload, carrier concentration, and accounts worth a broader needs review.
Define each metric before relying on it. Account retention, policy retention, and premium retention answer different questions.
Verify totals, date ranges, and account groupings against the source report. A polished dashboard still needs accurate calculations.
5. Start the day with a briefing
A morning briefing can bring appointments, important messages, and follow-ups into one view.
Smith demonstrated using connected email and calendar information to identify what needed his attention. Claude highlighted messages requiring a response and relevant calendar details.
Get leadership and IT approval before connecting agency systems. Grant only the access the task requires.
Use the briefing to prioritize, then check the source messages and appointments before acting.
BONUS: Put your brief on autopilot. Turn your morning brief into a scheduled task, so it's ready every weekday before you sit down. See how in the Build Brief.
What guardrails should agencies use?
Smith emphasized three rules:
Protect client information. Use fictional or appropriately prepared data for testing. Follow agency rules before uploading information or connecting systems.
Verify coverage language. Have a qualified, licensed team member review statements about policy terms and coverage.
Require human sign-off. Check facts, figures, names, dates, and commitments before anything goes to a client or carrier.
Smith also recommended paid accounts for agency work. Review data handling, permissions, and administrative controls when selecting an account. Payment alone does not authorize the use of client information.
How can your agency get started?
Choose one recurring task and build a workflow your team can review.
Smith suggested five starting actions:
Write three sentences about your agency, voice, and boundaries.
Create a renewal workspace with two approved, sanitized sample letters.
Turn one documented task into a skill.
Test an analysis with fictional or appropriately prepared renewal data.
Try a daily briefing using approved connections.
Track preparation time, corrections, and usefulness. Refine the process before expanding it.
Ready to hire your AI Chief of Staff? Download our five Claude builds that take daily work off your plate, plus a bonus that runs your morning brief on a schedule.
Can AI act as a chief of staff for an insurance agency?
AI can support drafting, research, analysis, and prioritization. It needs clear direction, appropriate access, and human oversight.
What is the difference between a project and a skill?
A project organizes context and reference materials. A skill defines the steps for completing a task.
Can AI write renewal letters?
AI can draft letters from approved instructions and supplied facts. A team member should verify the details and coverage language before sending.
Can AI identify cross-selling opportunities?
AI can flag accounts for a needs review, including those with one line of business. An agent must confirm the client’s circumstances before recommending coverage.
No Comments Yet
Let us know what you think