From Fax Machines to AI: What 25 Years in Insurance Taught MIC Global’s Founder About Timing

From Fax Machines to AI: What 25 Years in Insurance Taught MIC Global’s Founder About Timing

Harry Croydon built his first website to sell insurance in 1999. He was certain he’d discovered the future: put policies online, customers click to buy, and the entire industry transforms overnight.

Then he tried to actually sell it.

“I thought, oh, what we do is we build a website, put it on the website and sell it to all the people in America,” Harry recalls. “And when went there to sell it, you know, amazingly, nobody had a. Nobody could access a website. They all wanted it done by fax machines.”

In a recent episode of Category Visionaries, Harry Croydon, Co-Founder of MIC Global, an embedded microinsurance company that’s raised over $13 million, shared what twenty-five years in insurance taught him about the brutal relationship between vision and infrastructure—and how he finally got the timing right.

The State of the World in 1999

To understand how dramatically Harry mistimed his first venture, you need to understand 1999. “When you cast your mind back to 1999, the Internet had just been invented. AI was for boffins somewhere. Mobile phones were sort of penetrating a lot of the world. Some people had dial up still, you know, this is the state of the world in 1999.”

Amazon existed. A few pioneering companies were starting to sell things online. The infrastructure seemed ready. What Harry didn’t account for was the gap between early adopters in Silicon Valley and everyone else.

He entered insurance through cyber insurance, which the industry then called “insuring the Internet.” The irony was perfect: trying to sell insurance for internet risks to customers who couldn’t access the internet.

“So this is how I got into insurance, because I knew a bit about insurance because my technology sort of background, I knew a little bit about technology because of my technology background and I knew very little about insurance,” Harry explains, describing what seemed like the perfect skillset convergence.

The website launched. The vision was clear. The customers wanted fax machines.

The Two-Decade Wait

What happened next is the part most founders don’t talk about: the long, frustrating middle period where you’re right about the future but wrong about when it arrives.

Harry spent the next two decades watching infrastructure slowly, painfully catch up to the vision he’d had in 1999. The journey wasn’t linear. The dot-com bubble burst. 9/11 froze the insurance industry. Financial crises came and went. But underneath all that chaos, the foundational pieces were assembling.

“We’ve gone from this sort of fax era and telephone and posting that stuff in the last 23 years to actually doing things online 25 years later,” Harry reflects. “So, you know, it’s been quite kind of a ride for me at least, you know, going from trying to tell people what the Internet is and how you can email a policy or automatically generate one to people just kind of accepting that’s what you do now.”

The transformation was complete, but it took a quarter century.

The Convergence Thesis for MIC Global

By 2019, Harry had something most founders lack: pattern recognition from watching infrastructure evolve for twenty years. He could see multiple trends converging simultaneously, each one necessary but insufficient alone.

First, AI and machine learning had moved from academic curiosity to practical application. “AI was for boffins somewhere” in 1999, but by 2019, these technologies could actually power insurance products at scale.

Second, big data infrastructure had become accessible. The platforms needed to manage millions of micro-policies across multiple markets existed and were affordable.

Third, the gig economy had created the demand side. “The gig economy, sharing economy had made it, you know, grown a need for insurance by the mile, by the project, by the night, you know, by the job sort of type insurance,” Harry explains. Uber drivers needed coverage for specific hours. Airbnb hosts needed policies for specific nights. Traditional annual policies didn’t fit.

Fourth, mobile phones had achieved true ubiquity. “More and more people obviously had the mobile phone, which, you know, a kind of a distribution network in people’s pockets.”

The Microinsurance.com Coincidence

When Harry approached his future co-founder Jamie Crystal with this convergence thesis, he was essentially pitching the same vision he’d had in 1999—but now the infrastructure existed.

Jamie’s response validated everything: “Well, interesting enough, I’ve just bought microinsurance.com. He’d actually bought that URL literally two weeks before.”

Two experienced insurance professionals independently recognizing the same moment of convergence wasn’t luck. It was pattern recognition. They’d both watched the same infrastructure build for years, and they both saw it finally reach the tipping point.

“So he was thinking about the same thing and which is big coincidence in many ways. So this is kind of really what brought us together thinking about, you know, the technology change, the infrastructure around the world, the sharing economy, the platform sort of businesses, the new platform businesses that are out there and evolving,” Harry explains.

What Changed Between 1999 and 2019

The difference between Harry’s 1999 venture and MIC Global wasn’t just better technology. It was complete infrastructure transformation across multiple dimensions simultaneously.

In 1999, selling insurance online meant convincing customers to trust a new channel for something important. By 2019, embedded insurance meant meeting customers where they already were—in app-based flows they used daily.

In 1999, processing micro-policies would have been economically impossible. The operational overhead of traditional insurance infrastructure made anything smaller than annual policies unprofitable. By 2019, automation and cloud infrastructure made micro-transactions viable.

In 1999, global expansion meant establishing physical presence in each market. By 2019, platform companies went global instantly, and insurance needed to follow the same path.

“We thought this was the right time to start thinking about, you know, putting an insurance company together specifically to focus on that sort of project,” Harry notes, emphasizing that timing was the entire thesis.

The Brutal Lesson About Innovation Timing

Here’s what twenty-five years taught Harry: being right early is indistinguishable from being wrong. The market doesn’t reward prescience. It rewards timing.

In 1999, Harry had the right vision with wrong infrastructure. No amount of entrepreneurial hustle could overcome the fact that customers literally couldn’t access his website. The gap between vision and reality was measured in years of infrastructure development that no single startup could accelerate.

The hardest part wasn’t the waiting. It was recognizing when the moment finally arrived. Many founders from the 1999 era became cynical, assuming insurance would never digitize. They’d tried, they’d failed, and they concluded the market was the problem.

Harry maintained conviction in the vision while developing patience about timing. He stayed in insurance, watching the infrastructure build, learning which pieces were still missing.

Applying the Pattern

The framework Harry developed applies beyond insurance. Before launching any infrastructure-dependent business, map out the dependencies:

What technology needs to exist? In 1999, reliable web access was missing. By 2019, mobile-first infrastructure was standard.

What customer behaviors need to change? In 1999, buying insurance online felt risky. By 2019, consumers bought everything on phones without a second thought.

What complementary markets need to mature? In 1999, there was no gig economy creating demand for micro-policies. By 2019, platform businesses were everywhere.

What regulatory frameworks need to exist? In 1999, digital insurance was unclear legally. By 2019, embedded insurance had regulatory pathways.

When all dependencies exist simultaneously, that’s the convergence moment. Launch too early and you’re building infrastructure that already exists elsewhere but cheaper. Launch too late and competitors have already captured the market.

The Future Harry Sees Now

Looking ahead, Harry sees another convergence moment approaching around AI in insurance operations. “I think in three to five years there’s going to be a massive change in the adoption of technology into insurance,” he predicts, specifically around claims management, customer service, and fraud detection.

But he’s careful not to repeat his 1999 mistake of assuming transformation happens overnight. “For us, you know, seeing, you know, embedded micro insurance really take off because of people are looking at more risk, you know, being risk adverse and looking at new ways to offset their risks.”

The infrastructure is ready now. The customer demand exists now. The regulatory frameworks work now. That’s how you know it’s time.

Twenty-five years from fax machines to AI taught Harry one fundamental lesson: the best startup ideas require patience to wait for the right moment, wisdom to recognize when it arrives, and conviction to move decisively when all the pieces finally align.

In 1999, Harry was too early. In 2019, he was exactly on time. The difference wasn’t the vision. It was understanding when infrastructure catches up to possibility.