Underflow raises $3M seed to automate commercial insurance underwriting
What's the deal? UnderflowDealroom has a profile for this one. Try Dealroom → has raised $3 million in seed funding to build an AI assistant that automates commercial insurance underwriting. Founder and CEO Ola Kolade says past technology only "digitised the filing cabinet" without removing the manual work itself.
The startup's product, Aurora, connects to a firm's email, reads submissions and attachments, extracts data, and flags gaps. It then drafts follow-up emails to brokers and routes completed files to underwriters.
Aurora also writes preliminary risk assessments and enriches submissions with public data from property records and business registries.
Why now? Commercial underwriters spend 41% to 43% of their time on admin tasks like data entry, according to CapgeminiDealroom has a profile for this one. Try Dealroom →'s 2024 World Property and Casualty Insurance Report. AccentureDealroom has a profile for this one. Try Dealroom → puts the figure near 70%, a cost of $85 billion to $160 billion across the industry over five years.
A workforce shortage adds pressure. Roughly a quarter of commercial insurance workers are 55 or older, taking decades of expertise with them when they retire.
"You cannot hire your way out of that," Kolade said. "What you can do is capture the repeatable parts of that expertise in a system that scales."
What could go wrong? Insurance adoption cycles are long, and trust in new systems builds slowly. Kolade is betting that plugging into existing email environments lowers the barrier to entry.
The technical bar is also high. Kolade argues older tools failed because optical character recognition can read text but cannot spot that a loss run falls short of carrier requirements or that two forms contradict each other.
"That is the gap," he said. "Not reading. Comprehension."
The signal: Underflow is wagering that the next phase of insurance technology lies in comprehension rather than digitisation — automating the judgement-heavy intake work that prior tools left untouched. With admin tasks consuming up to 70% of underwriters' time and a wave of experienced staff nearing retirement, its bet is that capturing institutional knowledge in software beats trying to hire around the shortage.
Read more: Complete AI Training
