Predictive risk management, predictive analytics is the way forward

Developing and providing innovative technology solutions can significantly impact various industries. Technology-driven businesses can create software, hardware, or services that enhance the efficiency and productivity of sectors like manufacturing, agriculture, healthcare, and finance.

India Business and Trade spoke with Liz Thomas, Director, RoadE Labs Pvt. Limited on how predictive analysis and AI can analyse large datasets, identify patterns, and make data-driven predictions. The company’s core strength is Data & AI with Risk Analytics which involves the use of data, statistical models, and technology to assess, improve processes, bringing in efficiency and mitigate risks across different aspects of a business. The director says that process improvement with AI Learning, predictive risk management and predictive analytics are the way forward for businesses in India to stay relevant and ahead of the competition

 

Photo Source: Pixabay

IBT: Share an overview of RoadE Labs, and their core missions and explain some of the innovations that you have introduced in the market through the company.

Liz Thomas: RoadE Labs was incorporated in December 2017 and I joined the company in 2021. When the company was incorporated, its founders had this really good vision to be cutting-edge pioneers for Artificial intelligence and Data, especially predictive analytics. They were incubated in the Government of Kerala startup incubation centres and they also offered solutions to the Government of Kerala projects. One of the things was at that time GST was not there and they were one of the first people to actually use Edge Computing to try and process GST via e-way. They were able to use provide proof of concept on use of high-speed cameras to capture commercial vehicles number plates and categorize them as commercial. I would say RoadE Labs as an organization is still dedicated to innovations in Data and AI.

We are predominantly very focused on predictive analytics and artificial intelligence, especially edge computing. What our vision and mission is to help the industry, specifically manufacturing and automotive, and give solutions which were accurate in creating and maintaining efficiency in the industry. We also wanted to make sure that we would be the first promoters of digital fluency. Digital fluency for us, is saying that if you have data and if your process needs optimization and if there is any reason and chance that we can use your data to work for you, we would love to give a solution optimized for you at a reasonable cost.

Skill wise, I am into data and risk analytics and my co-founders Renjith Viswanathan is core AI, machine learning and deep learning and Banarji Balakrishnan is into predictive analytics. We are hands-on in our company’s strategy, specifically the technical side and Anoop Ramakrishnan (COO), who is a mechanical engineer, comes from a hardware background. He brought in expertise on IoT and connected networks. We are core into AI and we are into generative AI right now.

IBT: As of now you are catering to the automotive industry and you are present in the international market. In the coming years, what are the plans for expanding your foothold in the international market and entering any other industry segment?

Liz Thomas: We do cater to BFSI via Risk Analytics also. I have come from IBM and Roade Labs partner with IBM for their product implementations and I am a Risk analyst and a Solution Architect by domain & profession. So GRC, i.e. Governance Risk and Compliance is a niche segment that came into focus 15 years ago and proliferated in the last 10 years. What we do in this domain is we advise industries on operational risk management and we also advise banking, finance, insurance and security about the regulatory compliance management. Regulatory risk management has become big. Risk Analytics is one of our biggest domains expertise in RoadE Labs.
RoadE Labs got into Risk Analytics for GRC in 2021. GRC is all about what is the CEO worried about, what keeps the CEO awake at night and what is the health of the company.

You can see that 100-year-old companies are now rare. There is no brick and mortar organizations, which says that I will live forever. You have to be worried about the competition, environment and the changing trends of consumers.

We also proliferated in other related domains. Wherever we could go and propose Risk Analytics, AI and Predictive Analytics followed seamlessly and vice versa. We were able to provide solutions for Predictive Analytics with AI, that self-learned from the data of the organization and any processes that generated data. RoadE Labs has also been awarded as the Top 4 Predictive E-Care Startups for Electrical Automotive Worldwide in 2017. In summary, We do business predominantly in BFSI, automotive, and manufacturing for process improvement and IoT.

IBT: Please share some insight on the mechanism of digital twin works and how is RoadE Labs utilizing it for maybe predictive analysis or risk analysis.

Liz Thomas: Digital twin works great with large factory setups, machineries and manufacturing. We had this opportunity only once. The people in the organization are much more physical in nature and they might be slightly averse to technology on the floor.

Let me give you a scenario. Let’s say that the startup that was born 10 years back has now grown into a big company. They have factories across India, multi-storeyed factories, that may or may not be dependent on each other as a chain of processing. This is the physical scenario. Now, here RoadE Labs come and proposition to the CEO to say that we will improve your process, they might not be interested in investing huge amounts unless they understand the ROI, so we venture to create a digital twin to simulate process improvement virtually, without great cost on the ground. We do assessments floor by floor and ask the floor owners about the improvement prospects. And then we make a digital copy of the infrastructure and process in computer of the physical entity. This, we may call a digital twin of the factory set up of the organization. Now, I can simulate scenarios and do scenario analysis on this digital twin to see what would happen, if any changes are made on a floor. I can put a report to say that if you change three or four things it may not cost you anything except, change the place of people or change the sequences of process, and this would shave off 1-2 INR per product you are manufacturing, to a manufacturer who churns out 1000s of products per day, that is a significant cost cutting, at almost no expense.

Sometimes it makes common sense, but we don’t know how to do it, as we didn’t have the whole picture and the whole effect. But if you put it as a blueprint it will be more relatable and financially risk-free before the actual changes come into place. Digital twin actually helps with scenario analysis and I would say stimulate things for process improvement

IBT: In your view, how does predictive analysis bring out the efficiency of machinery or automotive?

Liz Thomas: It all depends upon the data. Traditional analytics are two types in nature. Whether you want to understand descriptive analytics which means what happened, and then there is detective analytics which actually works on descriptive analytics and says why it happened. That is not going to go away. Normal Analytics and Predictive Analytics is are two different things.

But predictive analytics today is going to combine normal analytics and analytics by extrapolation. For example, in automobiles, predictive analysis can predict the likelihood of when your car battery may die. If extrapolate one million cars, you may get a greater accuracy in prediction or probability. But if you extrapolate four cars then the prediction is not that good. So data has some more important stand here. First, predictive analytics will say that if I extrapolate it, I may get a probable time when it can happen and I can see patterns, and I may predict trends. It is also important to prep and clean the data before applying analytical models. This is more important in automotive manufacturing data. It is also more important in financial data where if you have about 20 years of market data, you may predict trends of the market moments. This is why data is very important.

AI predictive analytics never comes with one single product or hardware. It’s not that, If you buy two of them, you can get everything. No, we have to get them integrated.

IBT: RoadE Labs has a product called garbage detective. If you can tell me a little bit about it.

Liz Thomas: It is interesting that the company is headquartered in Kochi and Kochi once upon a time had a garbage problem. The district municipality came and explained about tackling the garbage disposal, which was right next to their government office. They tried to put a security guard for 8 to 12 hours. Then they had to put three security guards around the clock, it was difficult to prove in court when culprits were caught, because sentencing after incidents were not immediate

Those who want to litter, are very devious. They know how to actually analyse gaps, times and opportunities to dump garbage. So, the company founders set up a camera on a coconut tree and then they caught their first culprit. The camera caught the visuals and the municipality got the video and they were able to identify the culprit.

This is where we realised that this is a sustainability problem and human is the only creature that can generate waste in such quantity. We generate more waste than the world can handle. Now Sustainability is become a very important thing and a company such as ours can help. Building such technology also requires precision because we had to eliminate everything else like the rain, the sunshine, the shadows, the hail, and flood. If the water actually picked up garbage and ran with it, the water is a culprit also. The cat who foraged for food in the garbage dump and carried away a plastic bag is too. The municipality started receiving multiple notifications. Then we improved on it and then there was a solid and good product that gave correct notice or notification to their phone. Then we found out that it is not just the municipality that needs it very much. Apparently, it is the people who live abroad, who have empty and vacant lands, which was becoming a garbage dump. Now the UAE person who is sitting in their office will get a video and he can send it directly to the police station and they can nab the culprit because there is recognition and evidence captured. Then it became interesting that this became popular, not just as a sustainable problem, but also for people who wanted their surroundings clean. That is the origin of garbage detective.

We have proliferated into intruder detection, restricted area detection for FMCGs, defence and fire detections. Everything related to companies and factories would give targeted alarms and this is going to be a generic product where people can buy modules of whatever they want, one or more. Hotels decided to say that they will have multiple cameras with multiple channels to watch multiple locations and they will get dedicated notifications. All the things actually evolve from the garbage detective model. What started as a personal idea and now it is a product that is repurposed for multiple functions.

---

Liz Thomas is the director of Roade Labs Pvt. Limited since 2021. She has previously worked at IBM for over a decade in risk analytics. She completed her B.Tech in Information Technology from Cochin University of Science and Technology in 2000 and later earned an MBA degree from the Indian Institute of Management, Kozhikode in 2018.