Akash
Agarwal

Somewhere in India, a boy watched his father come home late - papers everywhere, forms in piles, every evening a kind of bureaucratic endurance sport. That father was an insurance agent. The boy grew up to be Akash Agarwal, and when he finally sat down to build a company, he didn't aim at a market. He aimed at the pile of papers.

Pibit.AI, the company Akash founded in 2020, does something insurance companies have wanted for decades but couldn't build themselves: it reads the unreadable. Loss runs - those dense, inconsistently formatted documents that log every claim an insured has ever filed - arrive at underwriters' desks in a dozen different layouts, fonts, and structures. Underwriters used to spend hours, sometimes days, just extracting the numbers. Pibit.AI's platform does it in minutes, with 96% less time spent on loss run analysis.

From IIT Roorkee to the Insurance Industry's Core Problem

Akash graduated from the Indian Institute of Technology Roorkee with a degree in Metallurgical and Materials Engineering - a discipline about understanding the structure of things, how components behave under stress, where systems fail. He may not have known it then, but it was a perfect training ground for insurance.

He went from IIT to Playment, a data annotation platform, where he managed complex AI training data projects. That experience - sitting at the intersection of messy real-world data and machine learning pipelines - gave him a blueprint. Insurance documents were just another form of unstructured data. The question was whether anyone would pay to have it cleaned up. The answer, as it turned out, was yes.

In 2021, Y Combinator accepted Pibit.AI into its Winter batch. For a founder building deep-workflow software for an industry known for resistance to change, the YC stamp mattered - not just for the capital, but for the discipline. YC's "build something people want" ethos aligned exactly with what Akash was doing: solving a pain that underwriters felt every single morning.

CURE: Five Modules, One Mission

The product Akash and his team built is called CURE - Centralized Underwriting Risk Environment. The name is deliberate. Insurance has long needed a cure. The platform delivers five integrated modules that cover every step of the underwriting workflow, from the first look at a submission to the final risk decision.

The five-module architecture reflects a systems-thinker's instinct - don't solve one part of the problem and leave the rest broken. DocumentCURE handles the conversion of chaotic, template-agnostic documents into clean structured data. ResearchCURE enriches that data with external sources. RiskCURE turns it into scores underwriters can actually act on. WorkflowCURE stitches it all together so the process doesn't fall apart between tools.

Kinetic, one of Pibit.AI's clients, now handles more than $1 billion in insurance submissions annually through the platform - without needing to scale headcount proportionally. That's the economic argument Akash makes to every carrier and MGA: the throughput goes up, the unit cost goes down, and the underwriters get to spend their time on judgment calls instead of data extraction.

The $7 Million Series A and What Comes Next

In November 2025, Pibit.AI announced a $7 million Series A led by Stellaris Venture Partners, with Y Combinator and Arali Ventures participating. Total funding reached $7.63 million. The company had 150 employees and was generating $9.8 million in annual revenue.

The funding announcement landed in an interesting moment for insurance AI. After years of generic "AI for insurance" promises, underwriters had grown skeptical. Akash's pitch is different - and the skepticism is actually part of the product design. Explainability is not an afterthought. Every output the platform produces is built to show its work, to be auditable, to stand up to an underwriter who asks "why did you score this risk this way?"

The Series A funds are earmarked for product development, expanded advanced risk models, API layer enhancements, and deeper data partnerships - all aimed at the next 12 to 18 months of deployment. The company is a member of WSIA (Wholesale and Specialty Insurance Association) and NAMIC, the two trade bodies that put it directly in the rooms where specialty insurance decisions get made.

Building AI People Actually Trust

There is a version of the AI-for-insurance story that goes: replace the underwriter, automate the decision, reduce the headcount. Akash is explicitly not building that company. The framing matters - both commercially and philosophically.

Commercial underwriting involves judgment about risks that don't fit neatly into spreadsheets: a trucking fleet with a complicated claims history, a commercial real estate property in a flood-prone area, a construction contractor with one large loss that distorts the whole loss ratio picture. These decisions require context, experience, and accountability. What Pibit.AI does is clear the decks so underwriters can get to that judgment faster and with better information.

The results from early customers bear this out. HDVI, a commercial trucking insurer, saved 25% of its underwriters' time. Shepherd Insurance, RMS Insurance Brokerage, and Method Insurance Company are among the growing client roster. The 32% increase in gross written premium per underwriter is the number that tends to end conversations with skeptics - it means the same underwriter writes more business, at better quality, with less burnout.

Achievements

• Forbes 30 Under 30 honoree
• Y Combinator Winter 2021 batch alumni
• Raised $7M Series A led by Stellaris Venture Partners (November 2025), with Y Combinator and Arali Ventures
• $9.8M annual revenue, 150 employees as of 2025
• Platform processes $1B+ in annual insurance submissions (Kinetic alone)
• 85% reduction in underwriting cycle times reported by customers
• 32% increase in gross written premium per underwriter
• 700 basis points (7%) loss ratio improvement for clients
• 96% reduction in loss run analysis time
• Featured speaker at InsurTech Connect (ITC) Vegas 2025
• Member organizations: WSIA and NAMIC