Drawing on more than 45 years in insurance, Beverly Scheider explains why responsible AI adoption begins with workflows, accountability and the relationships agents are trusted to protect.
Artificial intelligence can help an insurance agency respond faster, document work more consistently and reclaim time for clients. It can also repeat bad information, expose sensitive data or accelerate a process that was already broken. Beverly Scheider’s advice is clear: understand the work first, define the result the agency needs and decide where human judgment must remain in control.
Insurance professionals hear plenty about what AI can do. Across her responses, Scheider returned repeatedly to a more useful question: does it improve how work moves through the agency without weakening accuracy, accountability or trust?
Beverly Scheider is President of The Palmer Financial Group, Inc. and Founder of Operational Freedom™. Based in Illinois and Wisconsin, she brings more than 45 years of experience in insurance operations, agency distribution, operational excellence, AI strategy and workflow design.
That background shapes a view of AI that is grounded in operations rather than novelty. For Scheider, the goal is not to add more technology. It is to solve a defined problem in a way employees can follow, managers can measure and clients can trust.
AI Success Begins Before a Tool Is Selected
Scheider learned the difference between installing technology and changing an operation while leading a cross-functional team during the consolidation of two large insurance companies. The organizations brought together different systems, processes and cultures. The technology mattered, but it was not the hardest part. The real work was understanding how tasks moved, where they broke and how employees would adopt a new process.
“A technology project ends when the software is installed. An operational transformation ends when the way work moves through your agency has actually changed. Those are very different finish lines.”
The same distinction applies to a smaller agency considering an AI platform. Scheider describes a hypothetical agency that automates quote follow-up before standardizing how prospect information enters its system. Intake arrives through several channels, records contain missing fields and employees follow different naming conventions. The tool performs exactly as configured, yet it contacts the wrong person about the wrong line of coverage.
“If you buy a tool and drop it on top of a messy intake process, you’ve automated the mess.”
Before evaluating vendors, the agency should map one workflow from beginning to end. Where does the work enter? Who touches it? Where does it wait? Which decisions and exceptions change its path? New-business intake is a useful place to start because it exposes data quality, ownership and follow-up problems that an impressive demonstration may never reveal.
That process map also improves vendor selection. Scheider recommends asking how the platform handles incomplete information, unusual cases, monitoring and data removal. She also cautions agencies to decide who will monitor and maintain the workflow after launch. If the consultant leaves and nobody is assigned to maintain it, the automation can go stale or produce unexpected results. A vendor should be able to explain what happens after implementation and provide relevant examples from businesses of a similar size. The agency needs a system it can operate over time, not a demonstration that works only under ideal conditions.
“If the vendor conversation starts with their tool instead of your workflow, why are you still in the conversation?”
The Best Early Uses Often Solve Ordinary Problems
“The highest-value uses aren’t glamorous. They’re the leaks.”
Those leaks are familiar to agencies of every size: calls that go unanswered when the team is busy, online inquiries that wait too long, renewal reviews that slip past their due dates and routine notes that remain unfinished at the end of the day. AI may help capture an after-hours request, acknowledge an inquiry, schedule the next conversation, prepare a meeting summary or remind the team that follow-up is due. Her suggested starting point is concrete: count last month’s missed calls and unanswered web inquiries. “That number is your business case,” she says. Many of these use cases are explored further in NAPA’s guide to practical AI uses for insurance agents.
The purpose is not to automate everything that can be automated. It is to reduce a specific point of friction. Before using automated calls or texts, an agency should also review applicable consent, disclosure and opt-out requirements with the appropriate compliance or legal resource.
Asked how agencies can separate measurable value from novelty, Scheider returned to the business case. Each use case should connect to a measurable result. The agency might seek to reduce average response time, increase the percentage of inquiries receiving follow-up, complete more scheduled policy reviews or recover staff time for client conversations. Counting how many messages a system sent measures activity. It does not show whether service improved.
“Measure outcomes, not activity.”
Her test is practical: write down the measure, its current baseline, the target and the date when the agency will evaluate the result. The review should include errors caught by employees and what happened to the time the system saved. As Scheider puts it, “Faster is only better if it buys something.”
AI Can Prepare the Work, but People Must Own the Decision
Insurance is built around professional judgment and accountability. AI can draft, summarize, remind and prepare. It should not independently recommend coverage, offer claims guidance, make binding decisions or provide other advice that requires licensed professional judgment.
“AI output is confident by default. Confidence is not accuracy.”
That is why Scheider recommends review requirements that rise with the risk of the task. An internal meeting recap may need a quick accuracy check. A personalized client communication needs a named human reviewer. Material involving coverage, claims or a policy decision should be verified by an appropriately licensed professional against the relevant source documents before it reaches the client.
"If it requires a license to say, it requires a licensed human to approve.”
AI can support that review without replacing it. A preliminary check may flag missing disclosures, prohibited wording or unsupported statements before a person evaluates the draft. “The human still decides,” Scheider explains. “The machine just makes the review faster and more consistent.” For every meaningful AI-assisted output, she recommends assigning responsibility to a specific person rather than a general team.
Sensitive Data Requires Written Answers
The review boundary extends to client data. Insurance agencies may handle financial details, health information, driving records and information about a client’s home. Before any of it enters an AI-enabled platform, Scheider recommends obtaining written answers about where the data is stored, who can access it, whether it is used to train a vendor’s models, how long it is retained and whether it can be deleted.
The agency should also confirm that the vendor’s agreements and security controls are appropriate for the information involved. Personal accounts or free consumer tools should not be used for confidential client work. An agency cannot meaningfully supervise AI use if employees are quietly choosing their own platforms.
A Short AI Policy Is More Useful Than an Unread One
Her answers on data, review and accountability reflected a consistent preference for practical rules employees can use in daily work. A short AI policy should identify approved tools, state what information may enter them, define review requirements by risk and name the person accountable for the agency’s AI use. It should also cover client disclosure, employee training, incident reporting and periodic monitoring.
Shared prompts can be managed like other agency templates. Common prompts for quote follow-up, renewal reminders or meeting recaps should be named, versioned and improved by the team. This helps make the approved process the easiest process to follow.
These controls may also matter when an agency explains what happened after an error. AI can reduce missed follow-up while creating a different risk: AI-assisted miscommunication delivered quickly and at scale. Written procedures, review records and communication logs support accountability, but they do not prevent every claim or determine whether coverage applies. Coverage depends on the applicable policy terms, conditions, limits and exclusions. Agencies should discuss material AI uses with their insurance professional and other appropriate advisers.
Automate the Wait, Never the Relationship
AI can improve a client’s experience at the agency’s front door. It can capture an after-hours call, acknowledge an inquiry or help schedule the right conversation. Scheider believes transparency should focus on the moments that matter. When a client interacts directly with an AI voice or chat system, the agency should make that clear and provide an easy route to a person.
“Be transparent about the moments that matter, not every keystroke.”
The boundary changes when a conversation involves loss, fear or professional judgment. A client calling after a house fire needs empathy, not a barrier between the client and the agency. A rate increase, coverage decline or complex recommendation also calls for a professional who can listen, explain and accept responsibility for the guidance provided.
“Automate the wait, never the relationship. Use AI so the human conversation happens faster, not so it doesn’t happen.”
The same principle applies to retention. Generic outreach does not become valuable because AI produced it faster. Scheider recommends using approved information already in the agency’s systems to prepare a relevant draft for human review. A renewal message that reflects a prior conversation can give an agent a timely reason to reconnect. The system helps remember and prepare. The professional decides what is accurate, appropriate and worth saying.
Institutional Knowledge Should Not Live in One Chair
Some of an agency’s greatest operational risks are not recorded in any report. They live in the memory of the experienced employee who knows a carrier’s preferences, understands an unusual billing problem and remembers the workaround used every March. When that person retires or leaves, the agency may discover that a critical process was never documented.
AI can help an employee describe how the work is performed and convert that conversation into a draft procedure. Scheider stresses that capture and approval are separate steps. The employee should confirm that the draft reflects current practice. An owner or manager should then determine whether that practice is correct, current and consistent with carrier or compliance requirements. Otherwise, the agency may turn one person’s habit into an official procedure without ever deciding whether it is the right one.
“An SOP with no date is a rumor with formatting.”
Approved procedures should identify an owner, effective date and review schedule. Once maintained, they can support onboarding and reduce dependence on a broker or senior employee for every routine question. They also give the team a common standard that can be updated as requirements change.
Employee involvement is essential. Scheider recommends starting with work the team dislikes, such as repeated data entry or the pressure of unanswered calls after hours. The people who perform the work should help map and test the new process. Their participation improves the design and makes the purpose of the change easier to trust. “Consistency follows trust,” she says. “It never precedes it.”
Documented operations can also strengthen continuity and succession. A business that depends entirely on the owner’s memory is difficult for another qualified professional to operate. Systems that can be understood, monitored and improved make the agency less dependent on any one person. In Scheider’s view, documented and automated workflows also create transferable value. A buyer can more readily operate an agency whose systems do not depend on the owner’s memory.
“The owner who is the system has nothing to hand over but their own calendar.”
Questions Every Agency Should Answer Before Using AI
Scheider’s responses point to a practical readiness test. Before an agency automates a workflow, its leaders should be able to answer these questions clearly:
- What operational problem are we trying to solve?
- How does the work move today, including delays and exceptions?
- Which measurable result should improve and by when?
- What client information may enter the system and under which written protections?
- Which outputs require human review or licensed professional judgment?
- Who owns the workflow, monitors its performance and responds when something goes wrong?
If the agency cannot answer those questions, it is not ready to automate that process. The next step is to understand the work more clearly, not to select another platform.
AI Can Create More Room for Professional Expertise
Scheider expects AI to change the administrative layer of an insurance professional’s role most significantly. Quoting preparation, document handling, summaries and routine follow-up may require less manual effort. That shift does not make professional experience less important. It increases the value of the abilities technology cannot own: judgment about risk, empathy during a difficult moment, accountability for advice and trust built over time.
Responsible adoption therefore has a more human goal than the word automation suggests. A sound workflow, appropriate safeguards and measurable expectations can reduce the friction surrounding an agent’s relationships. They cannot replace those relationships.
“The agents who win will be the ones who used AI to make more room for that, not less.”
Explore the NAPA Expert Perspectives Series
Strong operations can reduce preventable mistakes, but no workflow or technology eliminates every professional liability risk. Insurance professionals reviewing how operational changes may affect their broader risk-management approach can explore NAPA errors and omissions insurance for insurance professionals. Any coverage response depends on the applicable policy terms, conditions, limits and exclusions.
This spotlight expands on themes introduced across NAPA’s wider Expert Perspectives series. Continue with What Top Insurance Pros Say Separates High Performers in 2026 to explore how experienced professionals approach operations, client relationships, documentation, technology and growth.
About Beverly Scheider
Beverly Scheider
President, The Palmer Financial Group, Inc. | Founder, Operational Freedom™ | IL & WI
Expertise: Insurance Operations • Agency Distribution • Operational Excellence • AI Strategy • Workflow Design • 45+ Years of Experience
Throughout her career, Beverly has led underwriting, billing, customer service, operations, and agency distribution teams before transitioning into independent practice and operational consulting. Today she helps organizations improve efficiency, communication, workflow design, and AI adoption while maintaining the human relationships that define exceptional client service.
After centuries of cautious management, insurance faces a tech wave threatening the role of human agents. But will AI reshape the industry as expected? Find out in our free whitepaper.
