Why SOPs are the Foundation for Smarter, Safer and More Scalable Operations

• By: AnneMarie McPherson Spears
• October 1, 2026

What’s the best way to get the work done? It’s a question that everyone in the agency should be able to answer. But all too often, the answers may vary based on personal preferences, half-remembered workarounds and exceptions, or sticky notes scattered across different desks.

Whether your agency is upgrading its agency management system (AMS), onboarding new hires or integrating exciting new artificial intelligence (AI) tools, the answers matter. Without clear, documented standards, discrepancies can hold an agency back in an era where nimbleness and scalability matter more than ever.

That’s where standard operating procedures (SOPs) come in.

“If you want to grow your agency, you must be able to define your SOPs—how you want to operate—to get as close to a consistent process as possible across your organization,” says Brett Cleveringa, vice president of product management at Zywave.

Here’s how agencies can use SOPs to build smarter operations, adopt technology more effectively and put guardrails around AI.

Consistency is Key

Every time a team member escalates a coverage question, a customer service representative (CSR) responds to a client with a billing issue or a producer enters new business data, each action adds up. The sum determines the customer experience, agency E&O exposure, data accuracy and even the agency’s ability to retain talent—for better or for worse.

“It’s really hard to build a business as an agency that has lots of components in its tech stack and different parts of the insurance lifecycle to manage,” says Mike Indigaro, vice president of customer experience for EZLynx. “If you do those things inconsistently, you’re creating poor customer experience.”

An agency that doesn’t have SOPs, or has ones that are sorely outdated, opens the door to human error.

“The bottleneck that I see most of the time is that there’s one or two people that have most of the knowledge rattling around in their brain,

along with having to organize all these tasks and meetings,” says Kyle Findley-Meier, founder of Agency SLM. “That’s really dangerous because we’re all humans and make mistakes.”

In an agency environment, those human mistakes can quickly become more than operational friction, they can become costly E&O claims.

“When processes aren’t standardized, agencies leave the door open for oversights, missed steps and incomplete documentation—and even minor missteps in those areas can create significant E&O exposure,” says Nancy Germond, Big “I” executive director of risk management and education. “By creating, maintaining and enforcing SOPs, agencies can ensure proper E&O risk control is baked into the day-to-day operations completed by every staff member.”

SOPs also solve a problem that becomes more urgent as experienced employees retire, change roles or leave the agency: Brain drain, the loss of institutional knowledge.

“One of the biggest things happening in insurance right now is the generational handover with massive amounts of retiring industry veterans,” Cleveringa says. “If you don’t document SOPs, you’re not going to be able to train your new staff that comes in.”

“Let’s prevent knowledge from walking out the door,” Indigaro agrees. “If you’re an agency recruiting new producers or CSRs, it’s a lot easier to have them rely on documented processes and training materials.”

And for agencies who are growing and looking to scale, “it’s really hard to adopt new technology and tools without proper process documentation,” Indigaro adds.

That groundwork can be grueling, but it is necessary. “Clear SOPs give an independent agency a common language for how work gets done,” says Kasey Connors, executive director of the Big “I” Agents Council for Technology (ACT). “Although defining expectations and documenting steps can take time upfront, agencies will see clear return on investment as their staff gains clarity and efficiency on those day-to-day tasks that can eat away at a team’s energy.”

Perhaps the biggest ROI is that SOPs allow agencies to more smoothly incorporate new technology, such as AI. “AI doesn’t fix a broken process—it amplifies whatever process already exists,” Connors says. “That’s why SOPs are so important. When workflows are clear and purposeful, AI can support the agency’s operations instead of introducing new uncertainty.”

How to Build an SOP

What agency processes need an SOP? The better question might be which ones don’t, says Findley-Meier. “I would say everything except for relationships,” he says. “Especially in insurance, the more organic the relationship we have with clients, the more safety they feel.”

“Once you get into any sort of capturing data or back-office work, those are the key areas where SOPs come into play,” Cleveringa says. “Make sure you’ve got that documented.”

Another area to be crystal clear is “escalation paths,” Indigaro says. When a CSR is on the phone caring for a customer with a problem, it’s not the time to fumble around with outdated instructions or a best guess. “You want that clear playbook so they know where to go and who to escalate an issue to,” he says.

Here are six steps to build a useful SOP:

1) Write down every category that needs an SOP. Findley-Meier recommends beginning with a wide-angle view of agency tasks and then moving into the details. “For instance, recruiting, which has more minute details like cold outreach, interviews, hiring—that’s just one SOP,” he says.

At SterlingRisk in Woodbury, New York, extensive SOPs are required to maintain its System and Organization Controls 2 (SOC 2) certification. “We have internal guidance, approval processes, very specific workflows for many areas, including client and policy documentation, AMS and CRM systems, onboarding, legal and compliance review, request for proposal (RFP) and proposal development,” says Stevi Siber-Sanderowitz, senior communications counsel and vice president of insurance analytics at SterlingRisk.

2) Create ownership. The staff member most proficient in the area should create the SOP, Findley-Meier says.

“Once you’ve identified which procedures to document, assign each one to a team member who can own the process of creating the SOP,” Connors says. “The divide-and-conquer approach makes sure no one person gets overwhelmed.”

“Process ownership is always a struggle, especially as we have 330-plus employees and we’ve been doing a lot of acquisitions lately, so a big part of a workflow is understanding who is taking ownership,” Siber-Sanderowitz says. At the same time, “we try not to work in silos,” she adds.

“This isn’t like Dunkin’ versus Starbucks. There are reasons and a lot of thought that gets put into this.”

3) Document each step. “An SOP should provide clear, step-by-step instructions to complete a specific task,” Connors says. “Keeping in mind one of the benefits of SOPs is to ease the hiring and training process, a great litmus test is, ‘Can a new hire follow these steps to successfully complete this task on Day One?’”

The good news is agencies do not have to start from scratch. The Big “I” Sample Agency Procedures Manual is a template designed for both personal and commercial lines workflow that an agency can customize to its own procedures. Learn more about the template, available only for Big “I” members.

How much information is too much information? “It’s a fine line,” Findley-Meier says. “We want to be precise, but if there’s too much complexity, we can create poor business quality because agents don’t absorb the information.”

“Keep it simple, make sure you have context, make sure it answers the why, and make sure it addresses the problems you’re trying to solve,” Indigaro advises. “Case by case, you may have more complex tasks with nuances and implications.”

It also depends “how much risk is in the process and how delicate you need to be,” Indigaro adds.

4) Edit with the end user in mind. “You can get wrapped up in building the perfect documentation for yourself, where you’re writing for the writer, not the reader,” Indigaro says. “It’s a common issue to use too much jargon or assume someone has context.”

Similarly, creating a consistent design template can prevent confusion and help other staff members interpret information more efficiently. “We undervalue visual cohesion in our documents,” Findley-Meier says. “But when we open up a folder later, we realize it would be valuable to be able to intake information in the same way.”

5) Review. After the SOPs have been drafted, send them to a small group of staff members to test. “You may find a few steps are missing or some clarification is needed,” Connors says. “You can even select beta testers from a completely different part of the agency to really put them to the test—have a commercial lines agent test the SOP for a CSR to document a customer complaint.”

Ongoing review is an important part of keeping an SOP useful, too. After all, nothing is really set in stone.

“If you are tying workflows to technology, it’s really important to make sure you’re auditing,” Indigaro says. “A mature agency will feel really comfortable with the fact that if you put a process into place in 2024, you’re going to update that process in 2026 because things have changed.”

6) Ensure the SOPs are easily accessible. Don’t overlook the step of housing and distributing the SOPs in a way where staff can easily find and navigate them, Indigaro says. “If a staff member can’t access it quickly on that customer call and has to put the customer on hold to troubleshoot or ask around, that’s not a great experience.”

That’s what Findley-Meier calls “the binder problem,” which is when well-written, detailed manuals collect dust because there’s “a gap between the document and the moment someone actually needs the answer,” as he described in an article for Independent Agent magazine.

One solution is retrieval-augmented generation (RAG), Findley-Meier says, which is when an AI platform only uses an agency’s SOPs to answer questions.

“RAG can help agencies turn SOPs from static documents into practical tools employees can use in the moment,” Connors says. “An agency can connect a secure AI assistant to its document storage and ensure the AI only retrieves answers from those approved SOP documents.”

“Then, a staff member can ask the AI a question—for example, say a CSR asks, ‘What steps do I follow when a client reports a billing discrepancy?’—and the tool can point them to the exact procedure, summarize the relevant steps and link back to the source document,” Connors says.

AI for SOPs—and SOPs for AI

AI programs can help agencies create, refine and search SOPs. But the direction of responsibility matters. Agencies should not let AI invent the agency’s process.

“AI can definitely be used to write SOPs, but being in AI and seeing the limitations of AI, I actually think it is the other way around,” Findley-Meier says. “We’re writing SOPs for AI. We have to stand between the AI and our final product all the time.”

Well-written SOPs can help agencies implement AI—especially agentic AI, which uses machine learning models to “accomplish a specific goal with limited supervision,” according to IBM. If the SOPs are incomplete, inconsistent or missing context, the AI will inherit those weaknesses.

“If you don’t have well-documented SOPs, and someone says, ‘Let’s build an AI agent to take care of a process,’ you don’t have any real structure or standardization to use to build an automation or solution,” Cleveringa says. “If the data and steps aren’t clean, you’re not going to be able to leverage AI very well at all.”

The need for a strong SOP foundation is becoming all the clearer as more agencies seek to implement high-powered AI. Fourteen percent of agents said their agency has already implemented an AI tool or solution, more than double the share in 2025, according to the “2026 Independent Agency Growth Study” from Liberty Mutual. Another 42% said their agency is learning about AI’s potential and identifying opportunities. And while only 8% of agents said their agency currently uses AI to analyze customer data, 42% said their agency would consider that use in the next year or two.

Lay Down the Law: AI and Governance

Before an agency starts plugging in AI solutions, it’s well worth the extra time to pause and set some extra guardrails.

“Think of AI as an employee that you just hired that has a great skillset, very personable, doesn’t produce a lot of typos, but hasn’t been doing insurance very long,” says Jason Gobbel, chief solutions officer at Kite Technology Group. SOPs determine what a good agency process looks like, but a governance framework determines whether an employee—human or AI—measures up to those standards, he explains.

Importantly, AI governance is about both “governing the interaction with the tool, but also governing the tool itself,” Gobbel explains. “What is it doing? You need visibility and observability on the backend as well.”

AI adoption is already outpacing formal governance. Catalyit’s “State of Tech Report” found that 48% of agencies are using AI for customer and prospect emails, 35% use AI for data analysis, and 34% use it for policy extraction—but only 19% report having a formal AI policy.

It’s a concerning gap, given the pitfalls of AI. In the Big “I” Agents Council for Technology (ACT) Agency AI Lab session, “AI Governance: Protecting Your Agency & Your Clients,” Gobbel spelled out some of the risks. Data security is a big one, with the potential for sensitive information to be exposed when entered into a nonsecure tool. “There are already laws that dictate what we can and cannot do with certain types of [customer] data,” he said.

Agency E&O exposure is another concern. “If we’re copy and pasting AI output without fact-checking it, and it hallucinated, and it’s wrong, we’re on the hook for it,” Gobbel said.

Additional concerns include laws emerging around bias and unfair treatment, as well as vendor risks. “We think that these tools are validated, and to some degree they are validated, but they’re validated to somebody else’s comfort level,” Gobbel warned attendees. “Does that vendor’s comfort level and faith in its own product match ours?”

At the same time, “AI is not a separate universe,” Siber-Sanderowitz says. “It’s just a technology that affects every existing risk that we have as an agency from an E&O perspective. It’s just moving faster and things are new.”

“We got in pretty early on, so we’ve already addressed AI in certain process-specific standards,” she continues. “Particularly in corporate communications and drafting workflows, we have an approved internal AI assistant. Our approach is that employees need to use approved tools, and those approved tools have established company guardrails.”

In the ACT Agency AI session, Gobbel outlined the AGENCY Framework, a five-week task list for creating a governance framework:

Week 1: Assess AI use. Agencies should start by creating an inventory of what AI tools are currently being used and how. “No judgment, we just need to know what’s going on,” Gobbel said. “This isn’t about right or wrong. This is about what’s missing.”

Week 2: Group use case by risk. Take use cases and group them by low risk, medium risk and high risk. “We can’t treat every AI use the same,” Siber-Sanderowitz says. “Something like internal, non-confidential notes is going to be treated differently than analyzing client information, interpreting policy language or making a coverage recommendation.”

Week 3: Establish roles and responsibilities. The agency needs to decide which specific humans own the AI governance. “It can be a team, but the team needs to have names,” Gobbel said. “Who approves the tools? What does that process look like? Who’s going to review the vendors? Who’s going to train the staff?”

Week 4: Name approved tools and data rules. An approved AI tool list is a must at this stage, along with an acceptable AI use policy, data classification rules, human review requirements and a vendor review checklist, Gobbel said.

“I know there are some agencies that use AI to independently find coverage, but we don’t want AI independently binding coverage or making a final coverage determination or other consequential decisions,” Siber-Sanderowitz says. “Vendor review is a big thing, so any AI products we use, we are extensively evaluating them for security, confidentiality, contractual terms, data retention, model training practices and integration with agency systems.”

Further, “all of our employees are required to sign an AI corporate governance agreement that says they must use SterlingRisk approved platforms,” she says. “I said to someone recently, ‘This isn’t like Dunkin’ versus Starbucks. There are reasons and a lot of thought that gets put into this.’”

The Big “I” Office of General Counsel and ACT issued a memo and sample acceptable use policy for AI to assist agents in developing their own internal controls and governance practices. Big “I” members can log in to learn more.

Week 5: Control high-risk activities. An agency must establish extra oversight for the most sensitive AI tasks, including client-facing outputs, coverage-related summaries and claims-related work, Gobbel outlined. This is when agencies outline high-risk procedures, such as human review, source verification, approval and documentation.

Human accountability is a key part of AI governance at SterlingRisk. “In the end, the employee using AI is responsible for the final work,” Siber-Sanderowitz says. “AI is not the final author. It’s not the lawyer or the broker. Verify, verify, verify.”

“And this is all built into our SOP,” she adds.

Yearly review and continuous improvement. Like SOPs, governance is an ongoing process. While yearly review is the goal, “every quarter for the first year, we want to be reviewing this stuff start to finish,” Gobbel said.

“Things are changing so quickly,” Siber-Sanderowitz says. SterlingRisk has held multiple staff training sessions on appropriate AI usage, as well as provided customized instructions and explanations to different departments. Part of the education process is explaining “this is why we need to do this,” and reinforcing that “this is important,” Siber-Sanderowitz says.

Ultimately, “the wonderful thing about being an independent insurance agency is you can do things however you want,” Gobbel says. “That is both a blessing and a curse. Within any agency of any size, you have producer and account executive preferences. And these AI tools may be pre-trained on some processes, but you have to align them with what good looks like for your agency.”

“And that can’t be Bob’s way and Sally’s way and Tom’s way and Sue’s way,” he adds. “It has to be the agency’s way.”

AnneMarie McPherson Spears is IA news editor.