Small commercial capacity is evolving – and not a moment too soon, according to Athula Alwis. “If you go back five to 10 years, not only was it inefficient, but there wasn’t enough change happening,” he said. That stagnation is what motivated the launch of AllDigital Specialty Insurance, aimed at overhauling underwriting efficiency and modernizing distribution in a sector that had grown complacent.
For Alwis, the core challenge remains unchanged: legacy insurers have been slow to innovate. “Most of the people there have day jobs and don’t have time to truly innovate,” he said. While large insurers are resource-rich, their structures often stifle transformation. In contrast, MGAs have embraced new technologies and alternative distribution models – momentum that is starting to reshape the market.
Central to that change is digital infrastructure. “The biggest change is building digital bridges – digital tunnels to connect with your partners and insurance carriers,” Alwis said. Increasingly, those tunnels are powered by artificial intelligence. But AI, he stressed, does not replace underwriters. “It enhances selection, speeds up decision-making, scales the underwriting process, and cuts costs and errors,” he said. Efficiency, particularly in a compliance-heavy space, is no longer optional.
Since launching in January 2021, AllDigital has deployed two AI systems, including one for its Lloyd’s operation. The company has written over $100 million in premium – a scale Alwis attributes to the advantages of cloud computing. “AllDigital wouldn’t exist without the cloud,” he said. Cloud-based architecture flattened the playing field, letting newer entrants operate without the overhead of legacy systems.
But scale, he added, is only part of the equation. The real test is precision – especially in risk placement. “The AI system we built is a machine learning system trained with data and by very experienced underwriters,” he said. That training took six to nine months and continues through a human-in-the-loop process. The result: speed and accuracy. “We offer bindable quotes in two to three minutes in self-service mode,” Alwis said – a dramatic improvement from the days-long turnaround times that once defined the sector.
Digital efficiency also enhances alignment. “Everyone has a shared view of the record, eliminating disputes over who has the ‘true’ version,” he said. Carriers, brokers, and administrators operate on a unified platform supported by technologies like blockchain and Salesforce.
Even amid market volatility – such as the 2022 spike in employment practices liability claims – the AI system has delivered. The company avoided sectors where the data lacked integrity or the model lacked confidence. “We avoided segments where we didn’t feel confident or the data didn’t support the model,” Alwis said.
He also noted that brokers closest to the customer – especially retail firms tied to large wholesalers – outperform others. “AI is an enabler, but proximity to the customer remains a competitive advantage,” Alwis said.
As for embedded or usage-based products, AllDigital is steering clear. “We’re not in the embedded insurance segment,” he said. The economics often don’t work. Low premiums and high integration costs don’t align with the firm’s focus on higher-margin, data-rich opportunities. “There’s a significant technology cost involved, so before allocating capital, we ask: what is the premium volume and what are the margins?” he said.
AllDigital currently works with 17 to 18 broker partners – a number that reflects a deliberate growth model. Rather than expanding headcount, the company is investing in deeper relationships and what Alwis calls “AI-driven decision-making” across its ecosystem. That strategy, he said, differentiates them in a market still working out how to scale without eroding profitability.
Ultimately, AllDigital’s approach reflects a shift in how specialty commercial insurance is done. Faster quotes and scalable underwriting are only part of the picture. The firm’s competitive edge lies in its use of cloud computing, machine learning, and permission-based blockchain to reengineer the mechanics of underwriting itself.
