Solving multiple problems with a single solution

Having a conversation with Ram and Raj at their office in IIT-M Research Park is full of anecdotes. One fills in for the other. Both avid bikers once, the two met when they were part of a Royal Enfield Bullet club. Their professional careers took different paths, but when Ram wanted someone to handle technology for a venture he had in mind, he connected with Raj.

Raj, who enjoyed software development and did not quite like getting into management work, felt he would be the CTO till he could identify one for Ram, but that never happened. That was in December 2014. The work they did together morphed into Zasti, an AI-based modelling platform that is meant to help businesses improve efficiencies by predicting risks.

Clinical trials: An eye opener

Ram narrates an interesting incident in his life that got him interested in clinical trials. He says he suddenly lost sight in both eyes and even top ophthalmologists and other specialists were not able to diagnose the problem. He was being treated at one of the premier eye hospitals in Chennai, where a chance meeting with some visiting specialists saw Ram going to the US for treatment. He was part of clinical trials being conducted on patients and, luckily, he regained sight in both the eyes. Apparently, the blindness was caused by something called travellers’ virus. That, Ram recalls, got him interested in clinical trials.

When they started out on their own, their idea was to dabble in Artificial Intelligence (AI). While most companies were using AI to solve one particular problem, build a workflow on that and then sell it as a product. “Nobody is selling AI as an offering,” says Raj. The idea behind Zasti was to see if by building an algorithm, whether they would be able to solve 5-6 similar problems. “Why should it solve only one problem? Why can’t we build a set of algorithms that can solve more problems, but similar in nature,” says Ram.

Raj chips in and says large companies such as Airbus and Maersk, for both of which Zasti is doing some work, will have numerous problems that can be solved using AI, but they can’t engage with different vendors for each one of them. That is where Zasti’s platform approach will come in handy.

Similar sets of data can be used to solve multiple problems. Raj points out that in the early days of software, companies used different software vendors for different functions and all of these got consolidated to form an ERP. “Somewhere this is bound to happen for AI. This is what we are trying to push,” he adds. According to them, they have built a platform where a company can come, provide its data, mention the problem and ask for a solution. The platform will analyse the data and come up with the best models that will work. Before they launched Zasti, they worked on the idea for more than a year to make sure that the platform model will indeed work. “What would have taken a year-and-a-half is now going to take just 6-7 months because we have something that can fast-track the process,” points out Raj. They are doing work on glaucoma, because of which they got incubated at the MedTech Incubator. Zasti is part of the first cohort of Lloyd’s Lab of London and the venture has also won the MICCAI AI challenge on pancreatic cancer survival prediction.

One of the prestigious partnerships they have, according to Ram, is with Paris-based Fondation A.R.CA.D, which is a custodian of clinical trial data for colo-rectal cancer. Zasti will help in finding which of the clinical trials have a greater chance of success. This will not only substantially reduce the time taken for clinical trials, but also will push trials in the proper direction.

Data privacy

According to them, all the data they use is anonymous, which ensures privacy even when there is a breach. The company gets the data only in numbers and does not know whom or what it belongs to. Another project Zasti is working on is to help a bank in Europe model its short-term loan products for customers, depending on the customers’ profile, earning and spending habits.

Zasti’s is a B2B product and gets its revenues from companies that use its platform. It is a Platform-as-a-Service model and can get its income either through a subscription model or a recurring fee. Since MedTech was the first focus, the healthcare segment will continue to account for a bulk of its revenues.

Pact with varsities

An interesting aspect of Zasti’s strategy is to tie up with universities, since AI requires a lot of mathematics and algebra knowledge. This is why it has opened an office in Croatia and tied up with the University of Algebra. The company has offices in Chennai, Bengaluru, Mumbai, Zagreb and Washington, DC. Next year, Zasti will look to raise about $5 million, by which time its 25-member team will more than double.