Sixfold Content

I’m beyond excited to announce that Sixfold has been officially selected to take part in the 12th cohort of Lloyd’s InsurTech accelerator program, Lloyd’s Lab. 

Lloyd’s Lab was recently recognized as a top-25 European start-up hub by the Financial Times and Statista and ranked the very top insurance-focused accelerator out of 19 countries and +2,000 organizations.

The program will give our team the opportunity to collaborate with Lloyd’s mentors to develop innovative solutions for the world’s leading insurance and reinsurance marketplace. Over the course of the 10-week “fast-track, fast fail” program kicking off in late April, our team will build, test, and iterate innovative solutions that “challenge how we do things and help the Lloyd’s market better serve its customers.”

Sixfold was one of just 22 insurtechs invited to travel to London to take part in Lloyd’s “Pitch Day” event. This year saw the largest-ever application pool for the program—Pitch Day invitees had to be culled down from more than 250 applications submitted by insurtechs spanning 33 different counties. 

For our pitch, I showcased how much we have built to serve underwriters in the past 10 months to 1,000-plus virtual and in-person attendees from across the Lloyd’s market ecosystem. We were selected into the program as one of only 12 teams by a panel of market leaders, mentors, and the Lab team, and we were the only underwriting solution to be accepted into the program! 

Members of Sixfold’s product and design teams will spend 10 weeks working out of the iconic Lloyd’s building in London. We will focus our efforts on accelerating and optimizing triage and risk appetite match capabilities within the scale of Lloyd’s markets. 

We’ll demo the fruits of our accelerator labor to the entire Lloyd’s market in early July—stay tuned for details!

Lloyd’s Lab unique approach to fostering global innovation is helping to tackle some of the world’s biggest insurance challenges. Progressing the Lab’s mission of supporting innovative insurance solutions across the globe, this latest cohort focuses on developing solutions to some of the biggest risks faced by businesses and communities in the Americas such as challenges arising from natural hazard prediction to risks associated with cybersecurity.

Thank you to Lloyd’s for believing in our vision and helping us bring it to fruition. We can’t wait to show the entire Lloyd’s market—and the rest of the world—what we have in store.

This article was originally posted on Linkedin

Today, life & disability carriers have access to more data from more data sources than ever, but the processing of all that data has paradoxically decreased efficiency and limited underwriting capacity. The industry has wisely turned to gen AI to help it reap the rewards of the modern data abundance while minimizing the cost.

In our view, gen-AI-powered underwriting isn’t about conversing with chatbots, it’s about collaborating with a platform that understands exactly what you need and exactly how to get it. But how should a platform like that function? What would it even look like? We started the process of answering those questions by interviewing dozens of underwriters to understand their frustrations, goals, and how other tools have failed them in the past. This input has guided our approach to design built around simplicity, context, trust, and fun. 

Simplicity

Many L&D carriers rely on automated extraction tools and/or offshore teams for their data-processing needs. These result in long multi-page reports, which are typically just ordered lists of data. Underwriters are left to rifle through these tomes to pull out relevant info. Sixfold eliminates these overly manual tasks by injecting AI throughout the underwriting process.

We use cutting-edge AI to independently build a virtual model of carriers’ unique risk appetites, so the platform can understand what constitutes negative or positive risk signals for that carrier. The platform taps a different set of purpose-built AI models to process ingested data and gracefully surface precisely what information underwriters need — and nothing else.

How does this AI wonder translate into design? We’ve leaned hard into “inconspicuously helpful.” The best AI is one that doesn’t make a show of itself, but can be depended upon to solve problems when called into action. More old-school British butler, and less Robin Williams’ genie from Aladdin. From a design perspective, that means elevating minimalism and simplicity.

Underwriters never see the AI hard at work ingesting data from disparate sources (APS files, MIB reports, labs, etc.). And they shouldn’t have to. All they want to see are easily scannable bits of appetite-aligned information in a unified dashboard experience. Leave the sausage-making to us, enjoy your hotdog.

This effortless minimalism extends to every part of the platform. On the case list page, underwriters can peruse a list of applicants with clear visual cues signifying which cases are closely aligned with underwriting criteria and which aren’t. In a previous technological era, underwriters had to invest time in scrutinizing each and every applicant with limited ability to triage, leading to reduced productivity. With Sixfold, they can focus their efforts on the biggest impact.

Context

Traditional data processing & ingestion treats all information equally. Taking a daily multivitamin is, for example, given with the same “weight” as immunosuppressants taken after an organ transplant. “Flattened” data reports, no matter how thorough, require underwriter intermediation to pick out the relevant through lines. 

We can do better in 2024.

Sixfold taps a medley of nine AI models to identify relevant data points from disparate sources and “connect the dots” between them. It’s the difference between, say, mentioning an applicant has diabetes versus summarizing how it’s being managed, e.g., “applicant was diagnosed with diabetes in 2015, but it is being adequately managed through diet and insulin.”

We haven’t merely replaced one form of summarization with an AI-powered one. We use gen AI prudently and solely in service of accelerating decision-making. Our goal is to present underwriters with as little information as possible (or, rather, just the right information) so they can focus their energies on making informed decisions. 

Concise auto-generated explanations are placed front and center, but underwriters can always dive deeper into specific information as needed (e.g., the platform may let you know that recent cholesterol readings were acceptable, but you have the option to dive into the readings from the last few labs).

‍Trust

The best underwriters are detail-oriented, and they deserve tools that meet (or better yet, exceed) their standards of thoroughness. Underwriters using Sixfold should be confident that everything has been scrutinized as thoroughly as if they did it themselves—if not more!

We accomplish this by showing our work. Transparency is baked into the Sixfold UX. We provide sourcing of all surfaced risk factors and conclusions, right down to the source and page so underwriters know precisely where a conclusion came from, and why it was arrived at.

And a little bit of fun

We want engagements with our platform to feel like working with your favorite coworker. You know the one we’re talking about right? The one that gets things done, but throws in a little light-hearted banter and joking around to improve your day just a bit.

Going through medical records isn’t what most people consider enjoyable. So we injected some moments of surprise and delight that can bring the joy back to underwriting. Our platform includes small flourishes that we hope provide a little levity.

Throughout the site, we’ve included illustrations of friendly robots collaborating with humans and empowering the humans to do more. The overriding theme of this aesthetic? Clay. Why clay? It feels deeply human, handmade with little imperfections that add character. We believe that AI can liberate humans from rote work (which is better handled by machines anyway) so they can emphasize their uniquely human qualities. The machines aren’t going to take over our jobs – they’re going to free us to bring humanity back to the work we do.

Want to see the platform in action? Watch our on-demand product demo for a live walkthrough of Sixfold’s Life & Disability offerings.

With our latest product update, we’ve sharpened our focus on Life & Disability via a suite of AI-powered features that overcome common underwriting challenges.

Sixfold’s number one superpower is to easily–and quickly–ingest carriers’ unique underwriting guidelines and automatically surface the submissions that match the carrier’s unique risk appetite. Moreover, the platform empowers Life & Disability carriers to streamline the underwriting process by:

🔘 Ingesting data from multiple disparate sources in an instant

🔘 Generating a comprehensive summarization of the applicant's health history and lifestyle within minutes

🔘 Surfacing positive & negative risk signals aligning with the unique risk appetite of each carrier 

🔘 Triaging submissions with an underwriter-facing dashboard for improved resource allocation

We've significantly broadened our Life & Disability offerings by expanding into six key areas, facilitating us to quickly create a comprehensive 360-degree applicant profile in just minutes by: 

✔️ Reducing manual workload through improved document ingestion 

Our technology has been significantly enhanced to process and analyze an extensive history of lab results, diagnoses, and medication records, covering years or even decades.

The platform is proficient in ingesting data from various sources including APS files, MIB reports, labs, applications, supplementals, Electronic Health Records (EHR), and Fast Healthcare Interoperability Resources (FHIR) files. By automating the ingestion of these diverse data types, Sixfold eliminates the need for manual document handling by underwriters.

✔️ Refining risk evaluation with a holistic view of medications, treatments, lifestyle choices, and family history

The platform leverages advanced synthesis of adjacent applicant history information, including pertinent family medical histories and lifestyle attributes, to offer a comprehensive understanding of their broader health habits and disease predispositions.

By integrating details from submitted records—such as family diagnoses ("father was diagnosed with melanoma at 63, but was successfully treated")—with insights into exercise routines ("engages in moderate-intensity aerobic exercise and weight lifting"), hobbies ("applicant scuba dives several times per year"), and substance use ("consumes a few beers every few days"), Sixfold provides a holistic view of an applicant's health.

This comprehensive approach enhances the precision of assessments, enabling more informed decisions regarding risk.

✔️ Improving risk decision precision with in-depth analysis of health condition progression 

We are now able to aggregate data related to a condition, including adjacent factors like medication, to chronologically track the comprehensive progression of the condition across multiple data sources. By utilizing detailed health data and its evolution over time, we enable more informed and accurate underwriting decisions.

This approach provides a unique layer of detail, incorporating crucial health information over time, allowing underwriters to quickly grasp risk with relevant context, thereby informing more precise rating and pricing.

✔️ Enhancing fraud detection by identifying inconsistencies across sources

Underwriters are tasked with synthesizing and managing a vast amount of information from lengthy documents, including Attending Physician Statements (APS), self-reported data, laboratory results, and more. An important aspect of analyzing these documents is to identify inconsistencies or discrepancies that could arise from oversight or fraud. Instead of solely relying on underwriters to detect irregularities across diverse documents, Sixfold proactively identifies and flags these discrepancies to the underwriter.

The platform's automated capability to identify problematic areas empowers underwriters to make informed decisions, leading to more accurate pricing of premiums and greater reliability of applicant information, safeguarding both insurers and applicants.

✔️ Expanding traceability with full document and page number sourcing

Our recent achievement of SOC 2 Type 2 certification underscores our commitment to being a responsible AI solution.

In this release, we've taken traceability to the next level by ensuring underwriters have access to complete information sourcing, pinpointing the exact document and page for increased transparency and accuracy.

✔️ Boosting underwriting capacity with upgraded triaging functionality 

Sixfold's latest update introduces a streamlined dashboard experience, designed specifically to empower underwriters to efficiently prioritize applicants who meet their risk tolerance and gracefully set aside those who do not. This effectively addresses the common 'front door issue' in Life & Disability underwriting, which involves managing an overwhelming influx of submissions by automating the pre-processing of applications. Within minutes, Sixfold accurately identifies and aligns applicants with the carrier’s risk criteria, significantly easing the burden of manual sorting and enabling underwriters to focus on the most suitable cases.

Our newest updates equip life & disability underwriters with a complete, clear and accurate health snapshot for every applicant, pinpointing key data points without sacrificing our commitment to compliance and data privacy.‍

Curious to see it all live? 👉 Watch our 20-minute product demo.

Dive into our Q&A with Brian Moseley, Co-Founder and CTO at Sixfold. We're chatting about his tech journey, from the early days at Sega to his recent tenure at American Express, and what led him to Sixfold. Brian also shares insights into the team's 'chill, pragmatic, and adaptable' culture and talks about his passion for Warhorn, his own gaming community project.

What’s your career path and how did it lead you to becoming Sixfold's CTO?

My software career began in college when I joined my friend’s community and media startup; we were among the first to experiment with web-based advertising. Through this venture, I met many tech professionals in San Francisco, who got me excited about the emerging web scene there. Ultimately, I decided to leave school and move across the country to work at Sega of America, focusing on database administration for their website.

From there, I joined several early-stage startups, including Critical Path, where I was the 17th employee. I experienced growth through its IPO, becoming an engineering manager and software architect. Later, I worked at the Open Source Applications Foundation, implementing a number of obscure internet protocols as the lead of an open source calendaring server.

As my career progressed, I found myself being the first engineering hire at a startup and then the first employee of another, so the logical next step was to become a cofounder and the CTO of Hoodline, a hyperlocal journalism company. Unfortunately, we did not succeed in reinventing local news, so I decided to try my hand at a big company and started working for American Express. 

I made the leap to full-time engineering management at Amex and eventually spent five years as the firm’s Head of Developer Experience, serving a global crew of 10,000 software engineers. Working in a big company was an incredibly educational part of my career, but I increasingly had the itch to build something from scratch again. So, Alex’s pitch got my attention and ultimately led to me cofounding Sixfold AI.

What originally attracted you to Sixfold?

It was a combination of three things:

1️⃣ I’d been looking for some time for an opportunity to build a company, team, and product from scratch.

2️⃣ I had a great time working with Alex briefly back in the early 2000s, so the prospect of cofounding with him was super appealing.

3️⃣ It seemed like exactly the right moment to go after the opportunities unlocked by generative AI in vertical business domains. Everything I saw was focused on LLM infrastructure and tooling, but nobody was talking in public about using generative AI to solve problems for business users like insurance underwriters.‍

‍How is Sixfold different from other players in the field?

We aren’t encumbered with legacy technology, organizations, or ways of working. We’re building our team and product from the ground up, which means we get to experiment freely and iterate quickly.

Another aspect that sets us apart is our commitment to ensuring our product is something users look forward to, not something they complain about with each use. The complexity and challenging nature of many enterprise software interfaces can be daunting. That's why we focus on user-friendliness and intuitive design.

We value user delight! Sixfold is not only effective; it's also fun to use.

What aspects of Sixfold and its technology are you most passionate about and find particularly exciting?

I love all of my children equally 🙃
‍
That said, as you might expect, I’m engrossed in the rapidly developing craft of generative AI engineering. What we’re doing goes way beyond basic “chat with your document” scenarios. We have a whole new set of primitives to assemble into complex systems - and they are evolving so fast it’s hard to keep up! It’s particularly interesting to apply these techniques in near real time as part of the live user experience. 

Our AI pipelines are hybrids of batch and streaming, with the work distribution and scheduling needs of batch processing and the resiliency and throughput of a real-time system. “Notebook AI” solutions don’t scale or meet our availability and performance requirements without a lot of additional engineering.‍

How would you characterize the dynamic and strengths of the team?

‍Our founding team has so much experience starting and scaling companies, we’ve all done zero to one many times. That doesn’t mean we don’t make mistakes, but they tend to be new mistakes, not many of the ones we’ve seen before. And my co-founders are teaching me everything there is to know about insurance.

The core of our tech team is a group of veteran startup hands with deep experience in their functional areas who can all flex into team building and leadership as we grow. We built on this core with a group of 'smart and gets things done' engineers who can build fast and iterate quickly, especially the closer they get to the AI layer of our stack.

Ultimately, our tech team is able to build, test, and learn quickly while keeping the quality bar high. And our go-to-market team keeps us oriented, fusing an understanding of what our customers need with a compelling vision for the future of underwriting.

How do you contribute to the team's growth and development?

As a hands-on-keyboard software engineer for nearly 30 years, I lead our product and platform development from the front. I pair with our engineers to help them develop their core software skills, which include architecture, design, coding, troubleshooting, and operating in production. This also provides an opportunity for them to build mentorship skills, as there is much I can learn from them as well.

I strive to create opportunities for our team members to step outside their comfort zones. Most of our engineers arrived with a specialization in some technical area, whether it be front-end development, infrastructure, or AI. However, our team is most effective when any member can work on any feature, regardless of where it falls in our stack. Therefore, I encourage each engineer to become proficient in all our languages and frameworks. For example, an AI engineer might make changes to our core domain model in Ruby, or a front-end engineer could add a new step to our AI pipeline using Python.

What is an achievement you are particularly proud of that the team has accomplished since the launch?

I’m most proud that we’ve been able to make the product work the way we envisioned it. It’s not just about our accuracy, which is great and getting even better. Remember the first time you experienced the magic of ChatGPT? Underwriters have that same feeling when they use Sixfold AI. They don’t have to read a hundred pages of tedious documents anymore just to find the three hidden nuggets that show the risk in an insurance case. They can sit back and let Sixfold find the nuggets for them in a matter of minutes. Magic!

In three words, how would you describe the team's culture?

Chill. I don’t mean we’re not working hard or that we don’t have high intensity about our venture. But you can tell that most of us have been in startups before because people handle the pressure really well. Nobody’s freaking out or creating negative vibes. We all have each others’ backs.
‍
Pragmatic. We aren’t striving for the perfect user experience or the most scalable systems right out of the gate. We’ll get there over time as we prove our right to be in business by delivering things customers want to pay for. We push ourselves to make “two-way door” decisions a little bit faster knowing that we can always iterate or reverse course later.

Adaptable. We develop in quick sprints and frequently stop to check in on what’s most important right now. Some weeks, that’s building features. Others, it’s improving resilience or shoring up security. At this stage of the company, every next customer engagement brings new problem statements and requirements. While we’ve introduced a measure of longer-term planning, we’re still keeping our heads on a swivel (as my high school football coach would say) to make sure our vision, roadmap, and near-term development plans evolve in the face of constant market discoveries.

How do you stay updated with the advancement in AI and LLMs, any particular newsletters or blogs that you like to follow?

I read 'Ben's Bites' daily, a resource recommended by Alex (our CEO) when I first joined. It's even included in our welcome email to all new team members. In addition, I follow 'The Information', particularly their newsletter 'AI Agenda', which I find very insightful and full of links to deeper reading.

Our team is also a significant source of discovery. Team members are constantly tuned into various niches and aspects of our industry, sharing relevant links and resources in Slack. We’re reading each other's stuff all the time.

Could you share some of your favorite tech tools or frameworks that you rely on heavily?

iTerm and Visual StudioCode - This is where I write and run my code. Others love their all-in-one commercial IDEs, but I do just fine with these tools.

Ruby and Rails - I’ve been building Rails apps and writing scripts and command line tools in Ruby since 2007. It just keeps getting better! I spend a lot of time with JavaScript as well, and it’s fine, you can’t really get away from it in the modern world, but I’ll always reach for Ruby when I can.

React - Again, I have been doing it since near the beginning, and it still scratches the itch. Haven’t had any reason to embrace the Hotwire part of the Rails omakase experience or to dabble in any of the post-React web frameworks. Maybe somebody out there will finally make me see the light?

Linear - Modern project management for product development that strikes an almost-perfect balance between the “all batteries in the universe included” and “bring your own batteries” approaches of pretty much everything else out there.

Mermaid - Being able to create a diagram by writing code has 10 x'ed my output of diagrams, which improves the clarity of my engineering dialogues.

Do you have a favorite AI solution that you would like to highlight?

Not yet! The gen AI ecosystem is still immature and changing rapidly. The popular tools and frameworks don’t always prioritize qualities we require for production usage - stable APIs, for example, or the ability to bring your own observability tools or work distribution systems. We’re having a better experience stripping down to the basics and devising our own abstractions over AI primitives (model and vector store APIs, for example) and in-house utilities.
‍
Ask me again in 12 months - I may well have a different answer!

Any fun tech projects that you're working on at the moment (non-Sixfold-related)? 

Yeah, there are two main things I'm focusing on outside of work. Firstly, I've recently picked up World of Warcraft again after about a year's break. It's been a fun way to spend time outside of building our company.

More to the point, I started a website back in 2001 called Warhorn, which I'm still running today, more than 20 years later. It's essentially a marketplace for tabletop role-playing games, board games, collectible card games, miniatures, word games, and more. I originally created it out of my love for tabletop role-playing, particularly tournament-style games like Dungeons and Dragons. It was born from the need to streamline the logistics, scheduling, and signups for large gaming events. 

The site was initially written in PHP in 2001, then I rewrote it in Ruby on Rails in 2011 and added a React front end around 2016 or 2017. Warhorn has been my laboratory for learning new technologies, such as GraphQL. It's a SaaS offering, which has helped me familiarize myself with various cloud providers and vendor systems. This experience has directly influenced my decisions at Sixfold and at previous jobs.

Warhorn is not just my tech lab but also my way of contributing to the gaming community. It's a long-standing project with over 25,000 monthly active users. I've kept it running as a passion project, not for commercial purposes. It's a creative outlet for me, similar to painting on canvas. It's an enjoyable aspect of my life where my profession and hobby intersect, providing a way to relax and be creative.

Warhorn seems absolutely awesome! Thanks, Brian, this was a lot of fun. 

‍

Are you excited about the opportunity to work at Sixfold?

If you’d like an opportunity to work with us, check out our vacancies.

It’s impossible in 2024 to be an insurance carrier and not also be an AI company. In this most data-focused of sectors, the winners will be the organizations making the best use of emerging AI tech to amplify capacity and improve accuracy.

This is a challenge and opportunity that Sixfold is uniquely suited to address thanks to our decades of collective industry and technological experience. We know insurers’ needs—intimately—and understand precisely how AI can overcome them. 

In previous posts, I’ve described how Sixfold uses state-of-the-art AI to ingest data from disparate sources, surface relevant information, and generate plain-language summarizations. Our platform, in effect, provides every underwriter with a virtual team of researchers and analysts who know exactly what’s needed to render a decision. But getting there is the rub. Training AI models (these “virtual teams”) to understand what information is relevant for specific product lines is no small task, but it’s where Sixfold excels.

To use AI is human, to create your own unique AI model is divine

Underwriting guidelines aren’t typically encapsulated in a single machine-readable document. They’re more likely to exist in an unordered web of internal documents and reflected in historic underwriting decisions. Distilling a diffuse cultural understanding into an AI model can take months using a traditional approach, but with Sixfold, it can be accomplished—and accomplished well—in as little as a few days.

Sixfold’s proprietary AI captures carriers’ unique risk appetite by ingesting a wide variety of inputs (be it a multi-hundred-page PDF of guidelines, a loose assortment of spreadsheets, or even past underwriting decisions) and translating it into an AI model that knows what information aligns with a positive risk signal, a negative one, or a disqualifying factor. 

With this virtual wisdom model in place, the platform can identify and ingest relevant data from submitted documents, supplement with information from public and third-party data sources, and generate semantic summaries of factors supporting its conclusions—all adhering to the carriers’ unique underwriting approach.

Frees human underwriters to do uniquely human tasks

It can take years for a human underwriter to master underwriting guidelines and rules, but that doesn’t mean human underwriters are no longer needed–quite the opposite. By offloading the administrative bulk to AI, underwriters can use their increased capacity to prioritize cases that align with their unique risk appetite. 

Consider a P&C carrier that prefers not to underwrite businesses that work with asbestos. When an application comes in, Sixfold’s platform processes all broker-submitted documents and supplements it with relevant data ingested from public and third-party data sources. If Sixfold were to then surface information about “assistance with obtaining asbestos abatement permits” from the applicant’s company website, it would automatically mark the finding as a negative risk signal (with clear sourcing and semantic explanation) in the underwriter-facing case dashboard. With Sixfold, underwriters can rapidly discern the applications that are incompatible with their underwriting criteria and quickly focus on cases aligned with their risk appetite.

Automating these previously resource-intensive data processing workflows allows carriers to obliterate traditional limits on underwriting capacity. The question for the industry has rapidly moved from “is automation possible?” to “how quickly can we get it implemented?” Sixfold’s purpose-built platform empowers customers to leapfrog competitors relying on traditional approaches to training AI underwriting models. We get you there faster–this is our superpower.

‍

New York City, January 9, 2024

‍Sixfold, the Generative AI exclusively built for insurance underwriters, today announced that the company has joined Guidewire’s Insurtech Vanguards program, an initiative led by property and casualty (P&C) cloud platform provider, Guidewire (NYSE: GWRE), to help insurers learn about the newest insurtechs and how to best leverage them.

Jane Tran, Co-founder & COO at Sixfold, expressed, “Guidewire stands as the industry's foremost policy vault, embodying the definitive source of truth. Collaborating with Guidewire empowers us to advance our enterprise-grade generative AI solutions tailored specifically for underwriters.”

Insurtech Vanguards is a community of select startups and technology providers that are bringing novel solutions to the P&C industry. As part of the program, Guidewire provides strategic guidance to and advocates for the participating insurtechs, while connecting them with Guidewire’s P&C customers. 

Sixfold seamlessly handles the ingestion, routing, classification, and summarization of submissions, and provides trustworthy, data-driven policy recommendations to underwriters in a user-friendly format.

About Sixfold

Sixfold brings the power of generative AI to the underwriting process. The platform significantly reduces manual workload for underwriters and amplifies confidence in every underwriting decision with improved accuracy, transparency, and capacity.