BlueSkyRisk

Jacob Shewbert

Insurance Innovation Authority

The MGA market has no shortage of technology. New policy administration systems, ingestion tools, AI copilots, analytics platforms, and data vendors enter the conversation constantly. But the question that matters is simpler, are we using technology and analytics to create a better experience for brokers, insureds and the underwriting teams serving them?

Having overseen operations across multiple MGAs, I have learned how easy it is to mistake digitization for modernization. Technology is not the strategy. Data is not the strategy. Better service is the strategy.

The Problem: Complexity

Every MGA likes to say its business is unique. That is true. A coastal homeowners program should not look like a commercial property warehouse program. Risk profiles differ and some data elements will always be line-specific.

But the industry has allowed that truth to justify too much fragmentation. Submission intake, clearance, triage, underwriting review, quote, bind and issuance and servicing may vary by class, but the operating model often rhymes. Too many MGAs still treat each new program as if it requires an entirely bespoke workflow and data model. That approach is expensive, slow and hard to scale.

The market talks a great deal about fast onboarding, including the race to 90- day onboarding. But fast onboarding that simply recreates legacy fragmentation is not modernization. A program is not truly on-boarded because it has been loaded into a policy administration system. It is on-boarded when it can quote, bind, service and report on a common operating model with usable data and repeatable controls.

A second problem is that too much valuable insurance data still lives inside documents. Submissions arrive in emails, ACORDs, PDFs, spreadsheets, schedules of locations and loss runs. In many organizations, those documents still drive the workflow. That is backwards.

The Solution: A Core Operating Model

The answer is not one-size-fits-all process design, and it is not bespoke everything. It is a common operating model with room for underwriting nuance. In practice, that means building workflow templates for major lines of business on a unified platform. The goal is to standardize the parts of the process that repeat while preserving the underwriting nuance that creates value.

When programs sit on one platform, each lesson learned becomes reusable. A workflow improvement in one program can be reflected in the next version of the template and deployed more broadly. That is how operating leverage is built.

“The right model is not blind automation. It is underwriter-guided learning. A system can validate submission data, summarize location schedules, identify missing information, and recommend coverage or pricing ranges.”

It also means solving the data problem at the beginning of the process rather than at the end. The real opportunity in insurance is to get submission data into a usable structure as early as possible. Once that happens, the data can drive triage, underwriting review, quoting, servicing and reporting. Analytics stops being a downstream reconciliation exercise and starts becoming part of the operating system.

This is also where AI can be genuinely useful. The industry does not need more vague claims that AI will “transform insurance.” It needs practical use cases. Two stand out: speeding movement through the various checkpoints in a policy life cycle and extracting usable data from unstructured documents.

Just as important, underwriters need to remain in the loop. The right model is not blind automation. It is underwriter guided learning. A system can validate submission data, summarize location schedules, identify missing information and recommend coverage or pricing rang

The Impact: Trust at Scale

None of this matters if it only produces more data outputs. Analytics initiatives succeed when they start with a business decision, not a dataset. Early in my career, leadership requests could take a week to answer because the data lived in disconnected systems and had to be reconciled manually. Even today, many experienced program leaders still run their businesses partly on feel. Dashboards should not replace judgment. They should validate assumptions and provide a pulse check.

That is why I believe it is better to deliver five or six core metrics consistently and accurately than fifty metrics no one trusts. Once that trust is established, organizations can add drill-downs by state, occupancy, account size, or location characteristics and create a more nuanced view of the portfolio.

The payoff is broader than internal efficiency. Brokers get faster, more consistent service. Insureds get a smoother experience with fewer duplicate requests and better-informed decisions. Underwriters spend less time rekeying data and more time on risk selection, broker relationships and portfolio management.

As organizations become more data-centric, technology itself will become a weaker differentiator. AI will lower the cost of capability, and access to data will continue to expand. What will matter most is how well firms deploy both. In that environment, relationships become more important, not less. The winners will be the MGAs and insurers that use technology to reduce friction, improve decision-making and make it easier for brokers and insureds to do business.

That is the future I am optimistic about: not a market with more software, but one where technology and data strengthen judgment, improve service and make it easier to do business.