Startup mantra: Damage claims now get AI video check
Inspektlabs reduces the cost of inspection and time of insurance and automotive companies by almost 95 per cent, with their differentiating factor being the ability to detect not only the object, but also the state of the object in question
PUNE Have you ever paid for damages for a rented car? Or had your insurance claim for the said damages rejected?
Devesh Trivedi, an IIT Delhi graduate, faced a similar problem a few years back, but he did not let it go. He decided to find a solution, using technology, and reducing human intervention.
Trivedi, along with Sanchit (who only goes by his first name) founded Inspektlabs, an inspection-as-a-service software startup in 2018.
Their software uses artificial intelligence and machine learning to automate photo and video-based inspection of assets like cars.
Inspektlabs reduces the cost of inspection and time of insurance and automotive companies by almost 95 per cent, with their differentiating factor being the ability to detect not only the object, but also the state of the object, such as a scratch on it.
Trivedi is an Indian Institute of Technology (IIT) Delhi 2008 alumni. He worked as Senior Data Scientist at Inductis for 3.5 years and then did his MBA from the Indian School of Business in 2013. Trivedi has also worked with McKinsey as a managing consultant for 4.5 years.
Sanchit is a computer vision specialist. He is also an IIT Delhi 2009 alumni and has finished his MBA from the Indian School of Business in 2014. He worked with the Indian Oil Corporation, Wipro and IBM in his initial career for six years. He then co-founded Curill, an artificial intelligence-technology powered healthcare startup in 2016. Sanchit then joined Zomato in 2018 as assistant vice-president, analytics, Zomato Gold.
Sanchit and Trivedi have known each other for the last 10 years. Trivedi says, “After I left my job in February 2018, I was looking for some startup ideas. Once, I rented a car, but when I returned it, I was charged heavily for damages which I was not responsible for. That is when I started looking deeper into this issue. I realised then, that there is no automated way to detect damage to a car. I had worked with some insurance companies during my previous two jobs. I had some connects there, and so I worked on this idea.”
Concept of Inspekt Labs is born
Trivedi says, “We are a computer vision tech platform. Computer vision means applying artificial intelligence and machine learning on images and videos to extract insights from that data. We are presently focussing on automated car inspection where the end-client is the insurer and the automotive industry. Our team of data scientists, PHDs, mechanics, and adjustors combine AI and inspection expertise, for example in a car crash, to automatically detect the damage.”
Overcoming a cold start
On the initial hiccups, Trivedi says, “We found a way to collect data and built a very large damaged-car image library in-house. As a result, we do not require or store client’s images for training and production. Also, we mask all personal information before processing images or videos, and do not store any personal data. We now have data of millions of damaged assets in photos and videos and have filed international patents for select damage categories.”
“Our first customer was Liberty General Insurance. To raise funds, we needed to have good traction. Initially, I put in my own money. Later, in July 2020 we raised $0.6M in a pre-series A round led by Better Capital of Vaibhav Domkundwar and Titan Capital of Kunal Bahl and Rohit Bansal and other angel investors.”
Future plans
“We are initiating a series A funding round. We are in talks with multiple VCs. Our broader vision is to include different product categories like detecting damages on cell phones, bikes. We are working with several automotive players outside India wherein the focus is multi-fold. For example, we are working with an OEM where we capture the video of the car before and after it enters a workshop. It helps us determine whether there was damage, what was repaired and what was not repaired, or whether there was any incremental damage or not. Currently we have clients in Japan, India and US and we are now expanding to other countries mostly Europe.”
Kashish Nagpal, an employee at Inspekt Labs, says “It has been an amazing journey so far. I started out as an intern two years ago, to focus on computer vision aspects of the product. It has been a great to see the company scale quickly across multiple countries.”
Founders’ speak
“We are a computer vision tech platform. Computer vision means applying AI and machine learning on images and videos to extract insights from data,” said Devesh Trivedi, founder
“We are in talks with multiple VCs. Our broader vision is to include different product categories like detecting damages on cell phones. We are working with automotive players outside India,” said Sanchit, founder.