AI Business recently caught up with Alex Dalyac, CEO and cofounder of Tractable.
After beginning his career as an Entrepreneur-in-Residence for Lazada Group, Alex worked as a quantitative researcher for Toscafund, a London-based long-short equity fund. He spent the following year at Imperial College London, leading the Computing department’s first industrial application of deep learning, consisting in automating the visual inspection of infrastructure and machinery in uncontrolled environments. The project was spun off into Tractable, cofounded with Razvan Ranca in 2014. Tractable raised its first institutional round of venture capital financing in June 2015.
Today, Tractable is focused on automating expert visual recognition tasks, with a current focus on the property and casualty insurance industry and industrial inspection.
Alex will bring his entrepreneurial character and AI expertise to The AI Summit in San Francisco on 28-29 September, where he will deliver his keynote on ‘Deep learning computer vision business cases: how to identify and develop them’.
Alex Dalyac of Tractable
We last interviewed Alex before he spoke at The AI Summit in London, so we took time to reflect on the event and discuss his key takeaway:
“The key takeaway was that enterprise is truly waking up to the opportunities that AI offers. Innovation departments are growing in seniority and influence, with senior executives seeing a move to it as a strategic internal career move; and AI is increasingly ending up at the top of their priority list. This makes sense: the tasks that AI can introduce or automate will grow quickly in breadth, enabling the AI or innovation department of a firm to grow into others, propelling the managers to the top of the organisation”.
Just as the enterprise has woken up AI, Alex tells us how business at Tractable has been accelerating since the London Summit five months ago:
“We have closed a seven-figure partnership deal with a major US automotive insurance claims player, and are running paid pilots of our AI audit solution with major European insurers. These pilots are proving that our AI can analyse the imagery from millions of auto collision repair estimates, automatically identify unnecessary repair operations, and deliver instant savings of $40 to $60 per claim. Our enterprise strategy for insurance is now shifting from proof of concept to growth”.
Considering the impressive savings made by the AI audit solution, it is perhaps not surprising that Alex positions this as Tractable’s “current focus” within insurance. The statistics he has reveal the massive opportunity:
“There are 20 million auto claims in the US per year in personal lines, another 12 million in commercial lines, and 100 million worldwide. The impact of this solution alone is therefore considerable”.
There is a second solution that comes from Tractable’s auto claims AI, though, which works in very different way – by place repair estimating in the hands of the customer. Alex explains:
“As soon as the accident occurs, by capturing images of the damage, our AI will produce a repair estimate for them. This will open up a wide range of possibilities: shortened cycle times, early warnings of total loss, third party claim capture, and on-demand cash settlement”.
And the potential here is even greater; Alex says “this market is estimated to be three times the size of the aforementioned audit market”.
Outside insurance, Alex explains that Tractable’s fastest growing sector is industrial inspection. The proposition is familiar, however:
“We are developing solutions to, once again, insert an AI expert eye into workflow processes to visually assess infrastructure and direct repairs appropriately. In the case of water infrastructure, we are finding that $100s per repair job are lost due to incorrect visual appraisal of the situation, and incorrect communication of visual information between repair gangs. The beauty of it is that images are already captured throughout these processes; all that is missing is a reliable, scalable, real-time expert eye to exploit the information from these images as soon as they are captured”.
With new AI start-ups emerging daily and expanding the ecosystem at an unprecedented rate, the competition in the marketplace is fiercer than ever. So what sets Tractable apart? Alex picks out two key areas and explains what makes his company unique:
“What sets us apart is our vertically integrated go-to-market and our proprietary interactive machine learning technology”.
“The former means that we build and own AI products that are inserted into the flow of data within the industry, allowing us to build proprietary datasets. In insurance, thanks to the 600 million image dataset coming from our partnership, we are now the only company in the world with both the training data and deep learning capabilities to develop AI auto collision repair estimating”.
“Our interactive machine learning technology is a set of proprietary algorithms packaged into an interface that enables a non-technical human to rapidly transfer his expertise to the machine. For example, without this interface, it would cost $30 million of manual labour to get insurer appraisers to teach AI to perform collision repair estimating. In general, this core technology lowers the tech barrier to entry into new visual inspection verticals. And our longer term plan is to extend it to speech and NLP”.
At The AI Summit in San Francisco on 28-29 September, Alex will deliver his keynote on ‘Deep learning computer vision business cases: how to identify and develop them’.
He will be joined at the event by fellow CxOs from the world’s leading enterprises and the most exciting AI software developers, gathering to explore the huge opportunity that AI presents all industry verticals.
To find out more, visit: theaisummit.com