From Entrepreneur First to US$1B: The Story of Tractable

Posted: 17 June, 2021

“It feels like it’s a crazy thing to do,” recalls Alex Dalyac.

“You find yourself thinking ‘how the hell is this six-month project going to be worth millions of dollars.”

Alex is speaking to an audience of aspiring founders, all of whom have tuned in to hear his story as the founder and CEO of Tractable. 

Just 6 years earlier, he and his co-founder Razvan Ranca were in their position. What they were attempting at the time probably did seem crazy.

They had joined Entrepreneur First, then a relatively new talent investing platform. They were going to build a startup based on a little-known area of artificial intelligence. Alex had a year’s work experience, Razvan had finished his Master’s earlier in the year. They knew very little about business, and had only just met each other.

What they did have was ambition, and a conviction that the technology they were working on would turn their industry on its head.

Now, Tractable, which uses AI to assist in accident and disaster recovery, has been valued at $1B.

It has over 200 employees across three continents, and customers include industry heavyweights Ageas, Covéa and Tokio Marine – some of the largest insurers in the world. The company today is well into 8-figure revenue, and has grown 600% in the past 24 months. 

Its AI is used to visually assess damage to vehicles and property – which means it can accelerate insurance claims and help recover people’s livelihoods more quickly, as well as analyse photos of end-of-life vehicles to determine the parts that can be recycled.

This is currently helping recycle about as many cars as Tesla has put on the road in 2019. 

The acceleration of the business appears extraordinary. But this kind of journey doesn’t happen overnight.

In this interview, Alex shares with us his story, from his first application to our platform, to now, as the CEO of a unicorn – and what he learned along the way.

Finding Your Unfair Advantage

If you thought that Alex’s application to Entrepreneur First was a breeze, it wasn’t. His first application to Entrepreneur First, when he was a third-year econometrics undergraduate at London School of Economics, was unsuccessful. 

His determination to join was sparked by a talk with Entrepreneur First’s co-founder Matt Clifford, whose words had resonated deeply with Alex.

“He said: ‘never before in time has it been possible for someone to build a product by themselves, in their bedroom, that could serve millions of people. But with the advent of the programmable computer and the internet, that is now possible’. 

That’s what I’d always wanted to do. There’s an intense sense of freedom and adventure that comes with being a founder. EF seemed like the right opportunity.

However after my interview they told me: you can hustle, but you don’t know anything about technology.

That rejection changed my life.” 

EF asks founders joining the platform to identify their ‘edge’ – their unfair advantage in solving a problem compared to others. Most who take part have a ‘technical’ edge, meaning that they have deep expertise in a particular technology field.

Alex became obsessed with creating this edge for himself. Therefore, he went to Imperial College London to take a conversion Master’s course in Computer Science.

It was there that he encountered the technology that would become the foundation for Tractable.

“Now, everyone working in tech knows about deep learning – but back then only a handful of labs were working on it. That’s how I got a head start.”

Deep learning is a form of AI that replicates the functions of the human brain to process data and make decisions. It can be used for a variety of tasks, including detecting speech, translation, and recognising objects in images.

The team now deploys this technology to inspect damaged vehicles – however at Imperial, Alex applied it elsewhere.

“I was working with a company that inspects plastic pipe welds to make sure they’re done properly, and discovered that you could generate a lot of value, such as preventing bad welds, by automating the visual inspection. The domain was not exactly cool, but it was a real world use case for the technology I’d become obsessed with.”

This idea gave Alex the vision to use image classification to build something that ultimately resulted in Tractable.

When the time came to interview for EF again, Alex also had a PhD offer from Imperial College to pursue his research in deep learning. He’d given himself the technical edge he needed, and proven his grit to build something game-changing. He was accepted and began EF in 2014.

Alex’s advice to other aspiring entrepreneurs looking for their competitive edge is to turn to academia, and give yourself a headstart.

“Don’t follow the news for ideas. If it’s in the news it’s too late. Follow the academic fields; read academic journals; speak to academics. Find out where, across all areas of science and engineering, the breakthroughs are happening.

You need to follow academia for breakthroughs, get very technical to understand them intimately, and be the first one to work out which industrial use cases are going to be enabled by this academic breakthrough.”

EF's third cohort

Finding Your Co-Founder

While Alex had already developed some technical skills to enhance his ability as a prospective CEO, he knew he needed a strong CTO to execute his vision. 

“I had this idea that I wanted the best technical mind in the cohort – and there was a lot of talent in that room.”

Among them was Razvan, a recent postgraduate from the University of Cambridge. He was quiet, and technically brilliant. 

Having moved from his native Romania to complete his undergraduate degree at the University of Edinburgh, Razvan had persuaded the UK’s top machine learning professor to take him on as a student.

He’d heard about Entrepreneur First from an old flyer on the lab notice board. At the time, EF used tests to assess applicants’ technical ability. Most couldn’t complete them; Razvan aced every one.

“I learned about his credentials and thought this is the guy,’” Alex recalls. 

The two were placed on separate teams at a hackathon at EF’s kick-off weekend, but Alex left at 2am to find Razvan. They wound up coding all night to heavy metal – Razvan’s soundtrack of choice. 

The team soon realised that Alex’s comfort with risk and business savvy, Razvan’s outstanding technical ability, and both of their intense work ethics made them the right fit to work together. 

“I’m personally someone who ideates profusely, so Razvan’s calm competence, combined with EF’s structure, gave us a methodology to make this work,” says Alex.

“I’m also competitive – we had Magic Pony Technology in our cohort [who would later exit to Twitter extraordinarily quickly for $150M], and I wanted to race against them. I think that really helped us push ourselves.”

Finding Your Industry

The pair then explored how Alex’s idea of automating visual inspection tasks could have an impact in industry. While they knew the technology could work for plastic pipes, the market was small. “There were only 300,000 pipe welds in the UK every year. We needed something bigger,” he recalls.

They were looking for three features in an area to work in: a narrow image classification task, a large market, and a lot of training data for their algorithms.

They explored areas from dermatology to geology and medical imaging, but struggled to find enough data for their applications. Then, they landed on car insurance – and the data they’d been looking for:

“Every year billions of photos are generated of damaged cars by repairers who send them to insurers,” explains Alex. “You have this huge visual data digital trail, which is basically a deep learning gold mine.“

When the team looked up the numbers, they were floored; they’d gone from tens of millions to billions in market size.

They had a company.

Finding Your Pitch

Following months of customer research, programming and sleepless nights (Alex and Razvan even moved in together), Tractable were ready to pitch at Demo Day.

“5 months ago, a technology called deep learning reached human accuracy in large scale vision”, Alex boldly stated as he walked onto the stage. “This is a historic moment for artificial intelligence.”

Over the next few minutes, Alex talked the audience of investors through the technology that he and Razvan had built – their journey from plastic pipe welds, to dermatology, to their ‘billion dollar idea’, car insurance. 

The picture he painted was of two founders – young, ambitious, and very smart – ready for a technological arms race. And they had a headstart on the giants they were about to go up against.

Pitching at Demo Day

Finding Solutions to Problems

Alex and Razvan were joined by a third co-founder after Demo Day; Adrien Cohen, an experienced business executive, signed on after their Seed round. 

“We were young, excited and ambitious – we definitely had fire behind our eyes”, recalls Alex.

“However, big corporates often like to also see a ‘seasoned pair of hands’. That was Adrien. He knew how to talk with executives. He was invaluable in building those critical relationships and being a source of accelerated supervised learning.”

Adrien also supported the other two founders to build the strongest team possible for their vision. 

“When you’re still young and inexperienced, it’s hard to understand why someone with 5, 7, 10 more years of experience than you would come and work for you. 

But Adrien had a conviction that we were going to find the best – and that the best would join.”

The first two years after Tractable’s launch saw Alex devote some time to programming and mostly to finding the company’s first pilots. This time was one of the most challenging in Tractable’s history – and Razvan’s resilience proved a vital component of the company’s internal engine.

“Our first ever proof of concept with a UK insurer failed. We lost one of our best researchers in the process. There was so much doubt in many of our heads,” says Alex.

“But Razvan took one look at the situation and said ‘are you kidding me? Of course it isn’t done. This is what we need to do to fix it.’ As someone who’s emotionally driven, it really pays to have someone like Raz keeping you even-keeled on the bad days.

Another challenge for the business came when it was time to raise their Series A. The industry imposed a brutally slow sales cycle, and it took them a year to get a first partnership agreement. By then their Seed funding was running low.

“It didn’t look good- investors want your revenue to be way higher than ours was by the time you’re at Series A. So many people said no, and I had to fine tune my pitch to minute detail. We didn’t know if we had a business.

Several months before the Series A we’d been approached by a big tech company for an acquisition. Given the context, we’d fallen prey to a risk averse mindset and were ready to sell for $25M. Even that didn’t work out. Looking back, thank God it didn’t; we kept going.”

Finding Success

The years that followed their Series A saw the company shift their focus to sales. 

“Thanks to our academic breakthrough-driven timing, we were still the only ones in the industry to master deep learning. That gave us a huge advantage when it came to getting our first training data partnerships, because we were the only ones who could wield this technical magic.

Once they partnered with their first leading player and announced it, the company hit a turning point. Suddenly, the big corporations were interested.

“When you’re young and unproven, the big companies don’t want to take a risk on you. However, once you sign their competitor, suddenly they’ve got a different kind of risk – they’re about to be left behind. So they will be more inclined to adopt you.”

Once they had signed their first large customer, contracts came in faster and faster. 

The company now count leaders in the insurance industry as customers, including some of the US’s largest insurance companies, as well as the leading insurers in Japan, France and Poland.

Founding The Future

Tractable is also working in another area where the founders believe they’re going to have a huge impact.

With climate change sadly making natural disasters worse and more frequent, Tractable is taking their AI into the sphere of disaster recovery, and even prevention. 

“When a typhoon hits the shores of Japan, tens of thousands of people can be left without a proper roof over their heads. The spike in home insurance claim volumes is hard for insurers to manage with a human workforce. People can be left waiting for months for the on-site appraisal required to get their insurance claim payout. 

Our vision is that with Tractable, people will be able to submit photos from their phone and get an AI-accelerated claim payout within days of a natural disaster to start rebuilding their home. We’ve set ourselves the goal of helping a thousand families rebuild their homes faster after typhoon season this Autumn.” 

There can be no doubt that Alex and Razvan’s lives are very different to when we met them. The two excited and ambitious founders, both barely out of their Master’s, are now the leaders of the world’s first computer vision unicorn for financial services, and the UK’s first computer vision unicorn in general.

We asked Alex how he thinks he’s changed – and not – in that time.

“I’ve learnt to be a better leader, and to have personal conviction in the science behind what I’m working on, and to know when to default to personal instincts versus those of others. I believe that a comfort with high upside bets even if they’re harder to achieve, a strong work ethic and ambition are the best tools a founder can have.

Take two people who are just as intelligent, just as hard working. One is more cautious and just takes low-risk, low-reward bets. One takes on high-risk, high-reward bets. In the startup world, the latter will almost systematically attain a far higher outcome. To be a successful founder, you have to aim for things that seem unreasonable.

That’s what we did. And that’s what we got. And we changed our industry while we were at it.

Entrepreneur First tell you: ‘you’re here to build billion dollar companies’. I remember thinking ‘that’s ridiculous.’

I guess it isn’t.”