Layak Singh, Founder & CEO, Artivatic Data Labs in conversation with Amit Singh shares that while India is still at a nascent stage of AI adoption and maturity; Artivatic is actively working towards developing self-reliant AI systems
Where do we stand today with respect to AI adoption and maturity of the technology in India viz-a-viz globally?
India stands at a nascent stage in terms of AI adoption and maturity viz-a-viz globally. While in the developed economies AI has been actually implemented in various cases in an effective manner, India has just started experimenting with the technology.
In India, many of the senior executives want to use AI in their businesses but do not know where to use and how. Also, the biggest challenge is of experimenting without investment. They fear to invest in building something new.
However, chatbots have seen good adoption in India in segments like travel, e-commerce, logistics, healthcare, and education. Computer vision and recommendation technologies are few other AI applications seeing growth. We expect that over the next few years AI will see reasonable adoption in sectors like supply chain/logistics, agriculture, construction, engineering design, manufacturing, finance, BFSI, healthcare, education, and security.
At present, lots of use cases are being built on AI but it may take 2-3 years to attain mid-level of AI adoption and maturity in India.
How do you expect AI to transform sectors like manufacturing, retail, banking, healthcare, automobile, and logistics?
We expect AI to be most aggressively used in overhauling the following processes:
Mundane tasks – Tasks which are repetitive in nature like data-entry from documents, document matching, customer support, and answering to ticket/calls.
Product or service customization – New product offering, new services, and new product customization as per the customer need in sectors like retail and banking.
BPO/KPO related jobs – Most of these jobs will be overhauled by AI due to a lot of repetitive processes.
Maintenance/regular checking – Service-related checkups for maintenance in segments like manufacturing, automobile and oil and natural gas, will be done by AI-enabled sensors.
Operation or decision-related processes: Most of the operation-intensive processes like account opening, loan approval, pathology, payment processing, claims settlement, home interior designing, automotive manufacturing, and personalized education will be done by AI.
In addition, in the manufacturing segment, AI will help enable processes like automated welding, surface finishing, product designing, and assembling. Robotics will make this possible without any human intervention with great scale and high precision.
In retail, AI will help in consumer risk experience, customized products, computer vision-based product matching, brick-n-mortar store-focused customer experience, auto payment processing, real-time product recommendations, automated inventory processing, vendor matching to buyers, catalog quality check, warehouse optimization, and product description quality check.
Further, in banking AI will enable tasks like automated account opening, cheque processing, fraud detection, alternative data intelligence for loan, credit or insurance, claims processing, payment processing, KYC/document check, and risk profiling.
In healthcare, AI will help in diagnosing the diseases better, X-Ray/MRI report check, pathology processes, preventive health, complex operations and so on. AI will enable logistics industry for driver behavior analysis, accidental stoppage, route planning, cargo planning, and fuel and cost saving.
Similarly, AI will impact many such industries in a big way over the next few years.
Reliability of AI systems depends primarily on quality and quantity of the data. As the availability of relevant data is still a challenge, how are you treading this challenge to develop a self-reliant AI system, which works without active human intervention?
That’s true. Reliability of an AI system depends on quality and quantity of data but most importantly the reliability depends on the contextual learning for the particular industry.
Availability of data is surely crucial as the winner will be the one who will have a large amount of high-quality data. At the moment, these problems are being solved through public data, which are being shared by research institutes or enterprises where small PoCs are developed and then the same system is trained with massive data on the business premise to make it more self-reliant. Large enterprises have enough data and working effectively for AI-related solutions. Governments are also opening new data options which are helping to develop these cases.
Segments like retail, finance, automobile, and manufacturing have plenty of data available. We are working with our clients to develop self-reliant AI systems with the help of a massive amount of past data in a step-by-step process.