How MGAs Can Build an AI ROI Model That Holds Up Over Time

Artificial intelligence is creating significant opportunities for managing general agencies (MGAs) to expand their capacity, tackle friction in workflows, and accelerate their time to market. But as MGAs move from AI experimentation to AI operationalization, the cost of the technology is likely to become a discrete budget item.

Gartner expects spending on AI agent software to reach nearly $207 billion this year, more than double 2025 levels. And that growth is expected to reshape how companies buy and pay for software in the years ahead.

As MGA leaders consider their return on investment (ROI) models for AI, there’s one factor they need to address: using AI isn’t free. The actual cost of AI itself is continually shifting, and MGAs don’t control what those underlying economics may look like tomorrow. In response, MGAs need an AI business case designed to hold up over time, not one that depends solely on AI being cheaper than labor today.

An AI business case built primarily around labor savings assumes that today’s AI economics will hold. If an MGA is comparing an AI agent with offshore labor at $10 or $15 an hour, the pure cost-savings case may already be relatively narrow. If AI costs change, that equation changes with it.

I spent years in telecom working with the fiber, connectivity, and data center infrastructure that supports technologies like AI, so I have a real appreciation for how much that infrastructure costs to build and operate. We have all grown accustomed to inexpensive or even free access to AI tools like ChatGPT, but those prices do not necessarily tell us what the economics will look like over the long term.

Making major workforce decisions based on the assumption that today’s AI costs are permanent can create unnecessary risk. That is why a more durable ROI model needs to measure the business value AI creates beyond labor savings, particularly through increased capacity, higher bind rates, and faster speed to market.

The first AI outcome MGAs should measure is capacity. How can technology help the people they already have accomplish more?

Today, an MGA processing 5,000 submissions a month may rely on people to read emails, review documents, and enter information. That means the volume of business an MGA can handle is tied directly to–and limited by–human capacity.

In today’s model, if submissions suddenly double, that doesn’t necessarily lead to more business if an MGA isn’t equipped to process the higher volume. Hiring and onboarding more people takes time. Current employees can only process so many submissions in a day.

AI can create a more flexible model by decoupling submission volume from headcount and taking on the manual burden, like reviewing documents, gathering and structuring information, identifying what is missing, comparing changes, checking details, and completing repetitive system work. In short, it returns time so underwriters can focus on evaluating risk and making decisions. That’s extra capacity.

Conversely, AI can help determine where an underwriter’s limited time is most valuable. If 1,000 submissions come in, agentic AI can surface the submissions from agents with strong existing relationships or opportunities with other indicators that make them more likely to bind.

The ROI calculation should capture this added capacity, not simply the number of labor hours AI can replace.

Most MGAs know how much premium they write and how much business they bind. But understanding the business they did not win can be just as valuable. Where are bind rates lower? Why are opportunities being lost? And where is the business leaving potential revenue on the table?

AI can help determine where an underwriter’s time is most valuable. If an MGA reviews 100 submissions and binds 10 today, moving that bind rate from 10% to 12% means additional policies generating commission. AI can then surface the submissions from agents with strong existing relationships, or the ones carrying other indicators that make them more likely to bind.

This distinction becomes even more important as AI costs change. The question is not simply whether AI costs less than the process it replaces. It is whether the technology is helping the MGA identify and win enough additional business to justify the investment. That is the kind of value that can keep the ROI case compelling even as the underlying economics of AI evolve.

Capacity and bind rates show how AI can help an MGA get more value from the business already coming through the door. Speed adds another dimension to the ROI equation by helping MGAs move more quickly when new opportunities emerge. This is an incredibly entrepreneurial market, and MGAs are constantly looking for where they can go next.

Think about two independent MGAs going after the same opportunity. Maybe they both want to launch a new fire insurance program in a particular market. They can’t start writing that business until they have the systems and processes ready to support it. If one MGA can use AI to get there significantly faster, that becomes a real competitive advantage.

That ability to move faster changes the AI value equation. The return is not simply the cost of the work AI can automate. It also includes the value of getting a program to market faster, placing risk sooner, and capturing business that might otherwise go to a competitor. For MGAs, the opportunity to move faster and grow can create a much stronger business case for AI than cost savings alone.

As AI economics continue to evolve, MGAs need an ROI model that evolves with them. Cost efficiency remains part of the equation, but it should not be the only yardstick for AI investment. A stronger business case looks at what AI enables the organization to accomplish, from increasing the capacity of existing teams to improving bind rates and getting new opportunities to market faster.

Ultimately, the question is not simply whether AI costs less than the work it replaces today. It is whether AI creates enough business value to remain a worthwhile investment as the underlying economics change. That requires a shift in how we think about ROI. Instead of asking only what AI can replace, MGAs should also ask what AI makes possible that was not possible before.

That is where the bigger opportunity lies: giving existing teams the capacity to pursue more business, respond faster when opportunities emerge, and grow without expenses increasing at the same rate.