Aug 9, 2023
In the realm of startups, there's no definitive roadmap to success. Every venture has its own unique journey, filled with bright sparks of insights, tough lessons, and unyielding pivots. These narratives, often unfolding behind the scenes, are the heartbeat of innovation. Today, we pull back the curtain to share our ongoing journey at Gaya Technologies, where we are now fearlessly navigating the intersection of technology and insurance.
The Inception of Gaya, the team
We are two immigrant co-founders, Carl from Lebanon and Jean-Pierre from Haiti, with a shared dream of building a transformative software company. Our paths crossed at Stanford Business School, a hub of innovation and entrepreneurship that laid the foundation for our collaboration. Initially, we convened to discuss insurance for marketplaces, a subject Carl had immersed himself in, and Jean-Pierre was intrigued to explore. This 30-minute meet up morphed into a 3-hour campus walk, with conversations running deep and ideas bubbling up. We often joke that we probably hold the record for the longest walking meetings at Stanford. We were intentional about getting to know each other, delving into our backgrounds as if it were a date. This foundation of genuine connection and mutual understanding set the tone for our partnership.
Part One: The First Venture: DriveGaya
As students on Stanford campus, moving around without a car was very expensive. An uber round-trip to SF, 30 minutes away, was for $120. It was right post-covid and Ubers suffered from driver shortages. Jean-Pierre needed a car for the summer and he made a deal with his classmate, who was leaving the bay area for an internship, to rent his car. Jean-Pierre took a step back to see if other classmates were in a similar need. He put together a google sheet for the entire business school to connect and share cars. More than 8 cars were listed and shared that summer.
That’s how it started.
We conceived a car sharing platform that is community centric and that scales through communities vs. geographically. We coded it, launched it, spent an insane amount of time with insurance companies, to insure our rides (thank you Statefarm!), and with dealerships, trying to solve the chicken and egg problem by having an initial inorganic supply for our marketplace. We had teaser campaigns spread out across Campus. We were the talk of the town.
But, 6 months in, it was clear that the unit economics would not work. The market dynamics of car rentals, the prohibitive cost of insurance, the limits of community-centricity, and people’s attachment to their cars, made it in our view impossible. We wished Turo and Getaround goodluck and pivoted.
Most importantly, this 6 months worth of work, during business school, was a great proof of how much we enjoy working together as co-founders and how complementary our skill sets are and how adamant on improving we both are. "For co-founders out there, you can have hundreds of hours worth of conversation on what matters most to you and why, but it's in the heat of shared challenges that you truly grasp each other's caliber."
Part Two: Pivoting to Gayafi
Our interactions with insurance companies and dealerships introduced us to a new problem: car financials: car loans and car insurance.
In the US, most people get a loan to buy a car. There are $1.5 Tn in loans outstanding, with 80% originated by dealerships. After looking closer, it appeared to us that profitable dealerships were more in the business of financing cars rather than selling them. The margins were in the financing deal, an opaque transaction. This landscape was rife with predatory behavior.
Americans shop for cars. They don’t shop for car loans. We thought of solving the problem with a D2C company that spreads awareness about the importance of securing a loan before heading to the dealership. We realized early on that this was a losing war as dealerships are the ones that are holding the cars and will always win. They are the world’s best negotiators.
So we thought of refinancing. Americans are very used to mortgage refinancing. Car loan refinancing was not common. We looked at great D2C players in the space and were not impressed with the channel. We had to go about it differently.
While Americans don’t shop for car loans, they surely regularly shop for car insurance. Insurance is sold primarily through insurance agents. We tested the potential of agents bundling car financials together, the insurance and the loan refinanced. To our suprise, that’s what Statefarm used to do. We had the perfect case study.
Statefarm agents, back when Statefarm had its own bank, used to call on their clients who had a car loan to come refinance with Statefarm bank. With an army of more than 18,000 insurance agents, they used to refinance more cars than any refinancing player in the nation, combined. And they did that with a clunky tool and a terrible customer and agent experience. We felt like this was an untapped opportunity, with a 0 to 1 case study already established. We just had to do it 10x better.
So we set out to build Gayafi, a platform enabling agents to bundle both insurance and refinancing in a few clicks, leading to lower car loan payments and higher insurance stickiness. This innovative pivot caught the attention of Pear.VC, who, impressed by our resilience (we were rejected two months before for the car sharing idea) and the market potential, welcomed us into their world-class PearX accelerator program.
Building this platform took 11 grueling months. We constructed a marketplace in partnership with over 30 banks and credit unions. We landed more than 300 agencies to use our platform. Initially, interest among them was high, but their action was too low. They were focused on offering insurance and building up a recurrent portfolio (rightfully so, in hindsight). To remain top of mind for the agents and simplify the submission process, we developed a browser extension that tracks relevant information as the agents are quoting clients for insurance. As a result, agents could pre-qualify their customers with a few clicks and minimal loan application form-filling.
This tactic saw an upsurge in volume, but an unfortunate timing of federal rate hikes led to a conversion rate of less than 5%. Agents, used to converting 30% of quoted clients on the insurance side, found this to be more akin to a lottery business, creating another barrier to widespread adoption. Even the established D2C players had to let go of 50-70% of their lending officers. It was terrible timing.
Part Three: Accidental $500/ month SaaS deal.
Amid this major turmoil we were going through, we had one “lucky” strike.
One agency out of Michigan was impressed with our refi technology and browser extension and asked for us to repurpose it to help them sell more insurance policies vs. keeping it for the refi cross-sell. Inspired by Graham’s famous saying “do things that don’t scale”, we became a dev shop for that agency for one week-end and delivered on a very niche workflow: scraping information from a couple carrier portals and sending it over to Salesforce through an API.
On Monday, we presented our work. The agency owner loved it and decided to start paying us $500/ month. That prompted us to rethink what we were doing. We believed in the Gayafi vision and were doing everything it takes to make it work. But it felt like we were swimming against the current. We decided to take a step back.
We went back to the brainstorming board and reflected for days. We had spent around a year navigating the insurance distribution landscape, we became experts in their workflows, tech stack and processes. More importantly, we had heard tons of complaints from agents and agency owners but rarely did we sit reflecting on their pain points and how to solve for them at scale as we were busy pursuing our own vision.
After a couple of weeks of customer re-discovery with the many insurance professionals we had befriended, it was clear that there is a pressing need to assist insurance agents and brokers with their quoting workflows. The promise of API-driven form aggregators, which aimed to streamline the process by filling a single form for multiple carrier quotes, fell short in practice and lacked comprehensiveness.
Some clients even dabbled in creating their own Robotic Process Automation (RPA) systems, akin to pre-set macros replicating their actions. However, these band-aid solutions are costly, require significant upkeep, and are prone to malfunctions. Over-the-top-automation solutions (as opposed to back-end API-led automation) has always been thought of as fragile, rightfully so.
But not anymore.
Part Four: The rise of Large Language Models
In late 2022, the world experienced a seismic shift when it met ChatGPT, an advanced AI technology that represented the advent of the Large Language Model (LLM) era. The landscape was forever altered as AI asserted its place, not just as a theoretical marvel, but as a practical, groundbreaking tool.
The abilities of LLMs are immense, extending far beyond what one would initially assume. For many, their first encounter with these models was through content marketing, where AI was employed to craft engaging, human-like text. But this was merely the tip of the iceberg, an early exploration into a vast expanse of possibilities. These models demonstrated a keen understanding of context, intricate command over language, and the capacity to generate creative content that was indistinguishable from text written by humans.
Where LLMs truly shined was in the hands of developers - leading to productivity boosts of around 40-50%. Tasks that once consumed hours were now accomplished in minutes. When prompted well, LLMs can generate quality code and can understand code as well, including HTML pages.
Should one revisit the consensus around over-the-top-automation? Can we tie LLMs with broader AI tools including computer vision models to effectively parse HTML elements on complex webpages? Can RPA become antifragile? We spent a couple of months experimenting and our conviction went through the roof. “That’s exactly what we will be bringing to the world, starting with insurance.”
Part Five: Launching Gaya.ai | the co-pilot for insurance brokers
Recognizing the need for a change in the insurance industry, and capitalizing on a generational “why now” with the rise of LLMs, we pivoted once more, giving birth to Gaya.ai, the co-pilot for insurance brokers.
As we shared in our introductory blog, Gaya.ai is the AI-assistant insurance agents and brokers have been waiting for, taking away all tedious and laborious data-entry tasks and liberating them to focus on caring for their clients and growing their business. Gaya is designed to propel agents' productivity and accuracy across their core activities:
quoting: Gaya will autofill directly everything you have just inputted on any carrier portal into any other carrier portal, dynamically interacting with a clunky carrier portal page.
re-marketing: Gaya will recall all your interactions with a client and automatically go on your behalf to the right carriers to get a quote, compare them and reach out to your client.
selling: instead of having to dig through carriers’ documentation, or wait hours for help from support staff, you can simply ask Gaya’s chat bot and immediately receive an answer with a link to where the answer was taken from. You will be selling faster.
servicing: just like you would go today to ChatGPT and type “generate a Job Description to recruit a licensed agent in Texas”, you would endorse a policy simply by going to Gaya.ai and typing “change lienholder information for John Smith’s Honda Civic to XYZ Credit Union, endorse policy and send email to email@example.com”.
More than 50% of an agent’s time is spent on quoting new clients. For that, our first drop is a browser extension that acts as Google Autofill on Steroid, helping agents and brokers quote 10x faster. It is a lightweight plug-in that requires no change of systems and processes and no training whatsoever. This will give Gaya an unprecedented go-to-market wedge from which we can further verticalize and assist in the other core activities.
Customer reactions have been great so far. Across the board, every agent is echoing the exact same problem. The insurance industry is in a turmoil and household carriers are pulling out of entire states, refusing to renew policies. Agents are being hammered by renewal quoting activities, a highly manual and repetitive data-entry process. API-led raters are not enough. An AI-powered automation solution is the way to go. Without even launching, we have 50+ customers in our waitlist, more than 5 major insurance agency influencers watching for us and wanting us on their podcasts, and several agency owners have even invested in the company, in the very early phases of the product! It is one thing to gain users and another to convert them into paying customers, but having them invest in us at such an early stage? That is a powerful testament to the value and potential they see in Gaya.
Winning in insurance is not a pure function of the best technology. Just like the famous saying of “insurance is not bought, it is sold”, we believe that “broker-tech is not bought, it is sold”. Sure there will be a subset of early adopters like in every industry, but to win at scale, we had to roll from the inside. Adoption is not a pure function of productivity gains and ROIs. It is a function of building trust, caring for your clients, and being their ally. For that, we have proudly joined the 101 Weston Labs, an insurtech accelerator born out of the Independent Insurance Agents of North Carolina Trade Association, the most progressive insurance brokerage state in our view. They are introducing us to insurance agents across the tech adoption spectrum. As early stage startup founders, we are naturally surrounded by early adopters who are excited about shiny objects, are curious, helpful and want to share what is painful for them. But that’s not a holistic representation of our customer base. 101 Weston Labs is making sure we see beyond the early adopters so that we breathe and speak to the pain points faced by agents across the spectrum and reflect from the get-go on what prohibits different agents from adopting new tech despite its clear rational benefits.
We are also already getting interest outside of the insurance industry. Real-estate brokers and mortgage lenders for example share a similar problem of spending 40% of their time on data entry and form filling. They have also adopted various API-led automation tools (e.g., Zapier) but that is not enough. They also want Gaya and we do recognize the potential for a broader horizontal application of our technology.
We see Gaya evolving into the infrastructure of choice to transform any webpage into an API. By enriching CRMs with real-time data captured from the core websites where the actual work is happening, we are poised to tackle the significant chunk of time that knowledge workers invest in data-entry and form-filling. Furthermore, Gaya has the potential to revolutionize the technology landscape by dismantling the barriers created by legacy systems. No longer will businesses remain tethered to outdated solutions due to the daunting challenges of data migration; with Gaya, we aim to make transitions seamless, ushering in an era where data fluidity becomes the norm.
But every journey begins with a single and focused step. Ours starts with the insurance industry. Our commitment is to first establish a profound and tangible impact within this domain. Only then, fortified by our experiences and learnings, will we venture into broadening our horizon, bringing Gaya’s value proposition to a myriad of industries."
In the grand scheme of things, Gaya.ai represents a waypoint in our journey rather than a final destination. We have learned that the trajectory of a startup is not a straight line from conception to success, but a winding path brimming with lessons, pivots, and immense growth opportunities. Each step along the way has fortified our resilience, cultivated wisdom, and stoked our drive to persistently innovate.
At Gaya.ai, our commitment to transparency is unflinching. We believe in sharing our narrative, our struggles, and our triumphs. Our story is not one of perfect execution but of relentless learning, adaptability, and an unwavering resolve to deliver value to our customers.
We welcome you to stay tuned as we continue to innovate and navigate the twists and turns of our journey. Each chapter we write serves as a testament to our dedication to disrupt, innovate, and redefine the intersection of technology and insurance. Cheers!