AI Unicorn: Our 6-Year Journey to Success

From Basement Startup to Unicorn: The Tractable Story

Currently, Tractable holds a valuation of $1 billion. The company’s AI technology is utilized globally by a vast number of individuals, accelerating recovery processes following vehicle collisions.

Furthermore, Tractable’s systems contribute significantly to automotive recycling efforts, achieving a scale comparable to the number of Tesla vehicles on the road.

Humble Beginnings

However, just six years prior, Tractable consisted solely of myself and Razvan Ranca, our CTO. We were recent college graduates working from a basement.

This account details the journey of our company, outlining the strategies employed and the key lessons learned throughout its development.

Key Takeaways from Our Journey

• The initial phase involved intensive coding and development by a small, dedicated team.
• Rapid growth was achieved through a focus on innovative AI solutions.
• Scalability was a crucial factor in reaching a billion-dollar valuation.

The success of Tractable demonstrates the potential for transformative growth originating from modest beginnings.

Our experience highlights the importance of technological innovation and a committed team in achieving significant business milestones.

Leveraging a Novel Technological Advancement

My introduction to the field of artificial intelligence, specifically deep learning, occurred in 2013, a mere six months prior to its widespread recognition. This journey began with an online course, “Machine Learning with Neural Networks,” offered on Coursera and led by Geoffrey Hinton. The experience was profoundly captivating.

Prior to this, AI existed for me primarily within the realm of science fiction, exemplified by films such as “The Terminator.” However, reports in technology publications indicated a significant revival within the academic community.

This resurgence was fueled by the availability of training datasets that were 100 times larger and computational power that had increased by a factor of 100, achieved through the repurposing of GPUs (graphics processing units). Consequently, a substantial improvement in predictive accuracy was realized in image recognition a year prior.

This advancement signified that computers were beginning to interpret visual content in a manner analogous to human perception.

The subsequent challenge involved translating this technology into practical applications. While studying at Imperial College London, I collaborated with exceptionally talented peers to develop a plant identification application utilizing deep learning.

Demonstrating the app to a professor during a walk through Hyde Park, we observed his delight as the AI accurately identified various plant species from photographs. Such a capability had previously been unattainable.

I dedicated all available time to exploring image classification with deep learning. Despite this intense focus, the technology remained largely absent from mainstream news coverage, even within Imperial’s own computer vision laboratory.

It felt as though I possessed knowledge of a groundbreaking secret.

In retrospect, concentrating on a specific area of applied science experiencing a paradigm shift, one that had not yet permeated the business landscape, proved to be transformative.

Finding Ideal Co-founders: Building Relationships with Future Partners

I had initially faced rejection from Entrepreneur First (EF), a highly regarded incubator program, due to a lack of technical expertise. After addressing this deficiency, I reapplied to the program.

The final stage of the application process involved a hackathon, during which I encountered Raz. He was engaged in machine learning research at Cambridge University, achieved the highest score on EF’s technical assessment, and had authored publications concerning the reconstruction of fragmented documents and the development of poker bots capable of identifying deception.

His simple personal website stated: “I pursue data-driven solutions to challenges currently considered unsolvable.” This resonated strongly with me, and ultimately provided the inspiration for the name Tractable.

Throughout that hackathon, we dedicated an entire night to coding. The following morning, both he and I recognized a unique synergy developing between us. We subsequently decided to live together and spent the ensuing years working in close proximity, around the clock, from morning routines to extended coding sessions.

However, our progress wouldn't have been possible without Adrien (Cohen, president), who became our third co-founder shortly after securing our seed funding. Adrien previously co-founded Lazada, a prominent online supermarket in Southeast Asia comparable to Amazon and Alibaba, which was later acquired by Alibaba for $1.5 billion.

Adrien’s expertise proved invaluable in teaching us the fundamentals of business development, fostering confidence, and attracting exceptional talent to our team.

The Importance of a Strong Founding Team

A successful startup often hinges on the quality of its founding team. Complementary skills and a strong working relationship are crucial for navigating the challenges of building a company.

Finding co-founders who share your vision and values, and who possess skills that you lack, can significantly increase your chances of success. Consider individuals who can contribute expertise in areas such as technology, business development, and operations.

• Technical Expertise: Having a technically proficient co-founder is essential for developing and implementing your product or service.
• Business Acumen: A co-founder with business experience can help you develop a viable business model and secure funding.
• Strong Relationship: The ability to work effectively and maintain a positive relationship with your co-founders is paramount.

Investing time in finding the right co-founders is an investment in the future of your company. Look for individuals who you trust, respect, and enjoy working with.

Identifying Target Customers Early for Market Alignment

Tractable’s journey began at EF with a significant advantage – an existing, revenue-generating client. The initial application focused on the inspection of… plastic pipe welds.

The work itself wasn't particularly exciting. Plastic pipes are commonly used for transporting water and natural gas to residences, and these pipes are joined using welds. This process involves melting and fusing the plastic ends together. Image classification AI was developed to visually assess weld quality, providing tangible value through advanced AI technology.

Despite this, our sole paying client ultimately ceased collaboration, coinciding with our initial fundraising efforts. This presented a challenge. The limited scope of pipe weld inspections wasn't substantial enough to attract investor interest. Consequently, we broadened our exploration to include sectors like utilities, geology, dermatology, and medical imaging.

Success arrived with the automotive insurance industry. This market was characterized by its considerable size and operational inefficiencies, creating a clear need for innovation. The concept was simple: enable individuals involved in car accidents to submit photos of the damage via their smartphones, allowing AI to process insurance claims automatically and seamlessly.

This approach promised a streamlined, hassle-free experience for claimants. It represented a significant opportunity to modernize a traditionally cumbersome process.

The Importance of Early Customer Validation

The initial experience with pipe welds highlighted a crucial lesson. Having a paying customer, even a single one, provides invaluable early validation of a business concept.

However, it also demonstrated the necessity of assessing market size and scalability. A niche application, regardless of its technical success, may not be sufficient to build a sustainable business.

• Early customer feedback is essential for refining product-market fit.
• Market size is a critical factor for attracting investment.
• Diversification of use cases can mitigate risk.

The pivot to car insurance proved the value of identifying a large, underserved market with a clear need for technological solutions. This strategic shift ultimately positioned Tractable for growth and success.

Leveraging the Fear of Missing Out (FOMO) for Funding Success

Our company successfully secured $1.9 million in funding through a strategy that inadvertently capitalized on the principle of FOMO. The initial impetus came from Matt Clifford, CEO of EF, who suggested that impressing a Google early investor, Charlie Songhurst, could be pivotal.

Songhurst initially pledged $50,000. However, as he connected us with further angel investors in the US, securing subsequent commitments of $100,000 or more became progressively easier.

Remarkably, one investor committed $300,000 during a phone call while driving. I have yet to meet this individual in person, even to this day.

While this “angel party round” proved beneficial, Matt Clifford believed that the structured oversight of a venture capital (VC) fund was necessary. This led us to Ash Fontana of Zetta Venture Partners, based in San Francisco.

Fontana had reviewed our pitch video and engaged in preliminary discussions, but was hesitant to commit at that stage. His position shifted dramatically once he became aware of growing investor interest.

The emergence of other potential investors triggered a sense of FOMO. Fontana promptly traveled from San Francisco to London, remaining until we agreed to allow him to lead the investment round.

The finalization of the deal presented logistical challenges. We were in Berlin, residing in a boat hostel lacking printing facilities. We were compelled to locate a print shop and solicit signatures from passersby to validate the investment documents.

Surprisingly, this process required several attempts before completion. Ultimately, we succeeded in securing the necessary signatures.

The subsequent $8 million Series A round proved more challenging, as we hadn't yet achieved the $1 million in recurring revenue that was becoming a standard expectation. Fontana’s continued advocacy and diligent follow-up with investors were instrumental in overcoming this hurdle.

Securing and Publicizing Prominent Clients

Even after completing our Series A funding round, the viability of the business remained uncertain. Securing a one-month pilot program with an insurance company required a full year of effort, and the resulting revenue scarcely covered associated travel expenses. We even considered acquisition offers from major technology companies, one reaching $25 million, but fortunately, those negotiations were unsuccessful in retrospect.

A pivotal moment arrived with the acquisition of our first client contract valued at $1 million. This allowed us to demonstrate to an insurer the potential for our AI to deliver $50 in savings and efficiency gains for each of their 1 million claims processed.

Just twelve months prior, I had concealed my reaction to Adrien’s request for $30,000 per month from a potential client. Subsequently, I was instrumental in securing $1 million in new business to support the growth of Tractable.

Achieving this milestone three times propelled us into the esteemed category of companies experiencing “10x year-on-year growth.” Consequently, raising our Series B funding proved significantly easier, ultimately being led by the prominent investment firm, Insight Partners.

A crucial lesson learned was the importance of publicly announcing partnerships with high-profile clients to accelerate growth within the enterprise sector.

Often, large organizations are reluctant to embrace innovative solutions from new companies, even when those solutions represent the leading edge of technology. However, success with a major player, coupled with a public announcement, can create a significant shift in the market. Competitors will then feel compelled to adopt the technology to avoid being left behind.

We first observed this “fear of missing out” (FOMO) in Japan, where we now collaborate with all of the nation’s leading insurance providers. We are currently replicating this strategy in France, Poland, and, crucially, the United States, the largest insurance market globally.

Key Takeaways for Growth

• Secure Large Contracts: Demonstrate significant value with substantial client agreements.
• Public Announcements: Leverage client wins for market validation and competitive pressure.
• Focus on Major Markets: Prioritize expansion into key geographic regions like the U.S.

Defining Our Core Purpose

Initially, our focus was on rapidly scaling a business leveraging advanced AI technologies. However, a recurring question from the team prompted us to define our underlying mission: how do we contribute to a genuinely positive change in the world?

This led us to broaden our perspective, considering not only growth, technological advancement, and value creation, but also the potential for positive impact. We discovered that vehicles deemed too costly to repair often end up being sold for salvage at online auctions.

Our AI solutions are now utilized to identify vehicles suitable for recycling, in collaboration with LKQ, a leading automotive recycler. Each recycled vehicle contributes to a reduction of approximately half a metric ton of carbon dioxide emissions. By analyzing images of damaged cars at auction, our technology enhances the recycling process and increases the value of recoverable parts.

Furthermore, we recognized the opportunity to accelerate recovery efforts following natural disasters through AI-powered home damage assessment. The increasing frequency and intensity of extreme weather events, such as typhoons and hurricanes, pose a significant threat to homes and displace countless individuals.

This autumn, we anticipate our AI for home damage appraisal will assist a thousand families in Japan with rebuilding their homes more quickly. Recognizing that not all homeowners have insurance coverage, we are establishing an AI disaster recovery fund.

In partnership with Georgian, our investors, this fund will deploy our AI to those most in need, irrespective of their ability to pay. The funding source is unique: early in Tractable’s history, an intern utilized unused computing resources to mine Ethereum. This initial investment, held for six years, has grown substantially and is now available for a more impactful purpose.

Reflecting on our journey, I am grateful for our early involvement in the AI revolution. We benefited from the power of FOMO in securing funding and driving customer adoption, and we are pleased we resisted premature acquisition offers.

Achieving the first billion in revenue presents considerable challenges, but we are determined to push boundaries. Our ambition is to become the provider of a visual expert accessible to everyone, simplifying the management of car, home, and other asset-related issues, anytime and without complication. We are striving to make this vision a reality.