Commercial insurance still runs on faxes and PDFs, and co-CEO of MGT Insurance Graham Topol (pictured) thinks that says less about technology and more about priorities.
While 94% of insurers say they’re working to improve operational efficiency, over half admit they’re falling short. That gap points to deeper issues: legacy systems, risk-averse culture and a disconnect between those building the tools and those using them. For brokers, the result is the same: delays, mismatched appetite guidelines and lost business.
The Topols are taking a different approach. Not by automating brokers out of the process, but by building a full-stack carrier designed around their needs.
“We work only with brokers and agents,” he said. “AI automates the boring stuff so brokers can focus on what they do best.”
Topol’s motivation came from firsthand experience. When trying to buy a commercial policy for MGT, he spent hours filling out forms and on the phone with a broker – only to wait a week for a single quote.
Rather than cutting brokers out of the process, as some startups do, it's important to center them. Brokers aren’t obstacles to be bypassed – they’re the primary users. Automation, then, isn’t about replacing people but handling the repetitive tasks that slow them down.
By streamlining broker workflows, it delivers tangible value where it matters most. Topol contrasted his own two-week quoting experience with what MGT now offers: quote and bind in under two minutes. But he was quick to emphasize that speed is only part of the equation.
“We are a fully autonomous, full-stack insurance carrier. So, we're the carrier,” Topol said. “We're underwriting, taking the risk, managing the book.”
It’s a structure that allows for tighter control of the entire insurance process – from product design to claims. Having that control, Topol said, improves both feedback loops and pricing accuracy over time.
“We’ve got a lot cleaner data, which helps us improve our pricing and profitability,” he said.
The decision to become a full carrier, rather than a tech-enabled intermediary, came after a particularly revealing conversation. A CTO of a tech-enabled broker told Topol that one carrier partner still required documents to be faxed. Another needed them to be mailed.
“That was one of those lightbulb moments; why not build a carrier that was purpose-built for brokers?” Topol said.
Still, not everyone is sold on AI’s role in insurance. Some brokers worry automation could edge them out. Topol sees it differently.
“We see AI as a huge unlock for brokers and not a threat,” he said. “In an AI-first world, the broker’s role actually becomes more valuable.”
He offered a specific example: MGT’s appetite checker, which uses natural language processing to instantly determine whether a risk fits within their guidelines. Instead of combing through codebooks or waiting for an underwriter’s response, brokers can move forward – or not – right away.
“They can figure it out and not waste time,” he said.
Still, efficiency isn’t the end goal. Topol insists he isn’t cutting corners when it comes to underwriting standards.
“We’re maintaining our underwriting standards while we do this,” he said. “All of this is about helping brokers win business and reducing friction for the customer.”
The challenges facing incumbents, he said, are less about talent or intent and more about the systems they’re locked into. Many insurers still operate on legacy tech that was built decades ago.
But it’s not just technology that holds legacy players back. Culture plays a role too.
“We’ve hired folks from big carriers,” he said. “They said, ‘I would have never been able to try this.’”
Experimentation should be expected. Each product team is made up of engineers and insurance professionals working side-by-side, avoiding the siloed approach that can stall innovation.
“Just getting the OK to try something makes all the difference,” Topol said.
Still, Topol is clear that he doesn’t treat technology as an end.
“We use it where we can drive outcomes,” he said. “We’re really pragmatic about where and how we apply AI.”
The company’s model, he suggested, is less about trying to be radical and more about improving the tools that brokers use every day.
“We were like, we can serve this business model better,” he said. “Our job is to make those tools as powerful and easy to use as possible.”
