Newfront’s AI Hackathon: Unleashing the power of generative AI to de-risk human progress
By Gordon Wintrob | Published June 29, 2023
World-class data scientists. Top technology engineers. Industry-leading product managers. And some of the deepest expertise in commercial insurance. All converged on San Mateo at the end of April to push the best APIs in artificial intelligence to their breaking point. It was a weeklong deep dive into the transformative advancements in generative AI and Large Language Models (LLMs). Maybe you’ve heard about it? There were late-night debates on scalability and volume limits, integration arguments, demands for step-level infrastructure changes, and careful reviews of data sets to train GPT. Passion bleeds into obsession, but that’s where the genius comes from.
What was the occasion? Just Newfront’s annual hackathon.
How can AI help you?
For clients and prospects wrestling with business insurance needs, bringing the power of AI to bear on policy management is a matter of extreme urgency. Not a nice to have—a must have. And not tomorrow. Yesterday. Why? The world is becoming riskier and less certain, and attracting and retaining the best talent continues to create challenges.
Think of it this way: say you’re a CFO at an emerging tech company, or the company’s risk manager, or an IT exec. Just consider what you’re faced with when it comes to insuring your business:
First, there are hundreds of pages of over-involved proposals and contracts with complicated language and arcane terminology. Not what a CFO wants to spend her Monday morning stewing over.
But it’s not just that. There are multiple policies, each with different premiums with different due dates that all seem to be up for renewal soon, at least according to the reminder emails that land in your inbox every 12 hours or so. (It’s okay. Hit delete and have another swig of your quad latte.)
Oh, and your boss keeps reminding you to keep costs in check. Only the coverage we need, he says. Do you tell him about the explosion of ransomware attacks and cyber extortion that has cyber premiums surging, or the rise in natural disasters that has property coverage growing pricier by the day? Does he watch the news?
And don’t forget the policy changes that ought to be implemented now that the market has done a 180° in the last six months. Above all, don’t ignore that small but nagging fear that you might get hit with a class-action claim once you go public. If you don’t have the right D&O coverage, it won’t just be the C-Suite that loses their shirts.
Now flow all of that into old-school PDFs, byzantine email threads, stacks of messily printed contracts, certificates, and bills, and even the occasional fax.
Then get to work. The ELT wants a recommendation by EOD.
Crazed, collaborative, and pretty amazing
Hence the hackathon. Engineers were divided up in cross-functional teams. Bringing together this collaborative group of diverse thinkers is a big departure from the threadbare conventions of the insurance industry.
Each team had a double mandate: either break the chain of repetitive busywork that plagues insurance professionals across every industry—from transportation to telemedicine; or, imagine how to deliver hyper-personalized communications and insights for all of our clients. Ideally, do both.
It’s not enough to feed clients the popular cliche of “just trust us.” For Newfront, we stand by our value of Deliver Results.
Maybe now you understand why we started Newfront in the first place. Seems like an industry ripe for a revolution, no? We may not be wearing bonnets rouge, but we’re bringing some real change to this business. Here are some of the key client pain points we tackled in just a week of (admittedly obsessive) focus:
Claims – Auto-ingesting hundreds of PDFs and restructuring the loss runs to help build submissions and run claims analytics
Quotes – Automatically extracting structured quote data from lengthy documents, enabling comparative analysis and eliminating time-consuming data entry
Contracts – Creating AI-powered contract review tools that compare requirements of new contracts within in-place policies, identifying coverage gaps
Benefits – Developing conversational chatbots available 24/7 to answer employee questions and help them better understand their benefits package
Showcasing the results
Illustrious hackathon voting committee members joining me included: My fellow Co-Founder & CEO Spike Lipkin, Vice President of Engineering Lin Yuan, and Chief Product Officer Aaron Forth.
As a voting committee, we delivered an unexpected response to the half-dozen demos the teams presented in various states of sleep deprivation and slack-jawed astonishment. In a move unprecedented in the annals of Newfront Hackathons, we declared all of the projects winners, reflecting the exceptional quality of each submission, all which, in the words of one anonymous judge, “just killed it.”
Spike called the event “an inflection point” for the company, while Aaron said, “The cost of intelligence just became 1000 times cheaper. It's clear that large language models will redefine our roadmap, but also transform this industry and shape the future of work itself.”
Our insurance professionals were no different. They raved on Slack. They high-fived in meetings. They posted on LinkedIn. They said the innovative solutions presented represented major breakthroughs in client service.
Pushing the proverbial envelope
Even though we’re excited by the potential of what we witnessed earlier this year, Newfront has been diligently building a data schema and workflows for the past six years. We are already transforming the business with our client dashboard, which is like a central nervous system for policy management: submitting quotes, reviewing contracts, simplifying renewals, automating billing—everything is easier and faster with the platform. But it’s not just saving time. Using machine learning and generative AI, Newfront is also delivering data-driven insights into coverage and market trends that allow clients to make more informed risk decisions based on real-time data. No small thing.
And yet, the success of the hackathon has, to put it mildly, transformed the product roadmap. Newfront's Technology team is already bringing many of these AI-driven products to colleagues and clients.
Stay tuned. Or better yet—get in touch. Unless you enjoy the old way of doing business. In that case, we’ll let you get back to your filing systems and spreadsheets and putting more paper in the printer, which looks like it's on the blink again. Better call IT.
Newfront is transforming the delivery of risk management, employee experience, insurance, and retirement solutions by building the modern insurance platform. Transparent data delivered real-time translates into a lower total cost of risk and greater insights. Newfront makes insurance work for you. Headquartered in San Francisco, Newfront has offices throughout the country and is home to more than 800 employees who serve clients across the United States and globally.
Co-founder & CTO
Gordon Wintrob is the Co-founder and CTO of Newfront, where he leads the development of Newfront's platform and operating system. Gordon attended MIT and worked at Morgan Stanley, Blackstone, and Goldman Sachs during his time there. Prior to Newfront, Gordon founded StackLead, a company that built automated technology to research sales leads. He sold that company to LinkedIn and it became LinkedIn's jobs product, aggregating millions of postings across the web.Connect with Gordon on LinkedIn