Motor insurance in India is facing a structural profitability challenge.

For every Rs 100 that insurers earn in motor premiums, payouts and operating costs can exceed Rs 110– Rs120 in several segments. This imbalance reflects a deeper issue in how risk is assessed and priced.

Underwriting continues to rely significantly on static vehicle attributes such as engine capacity, age, and location, particularly in regulated segments. These variables help estimate repair costs, but they do little to predict the likelihood of an accident. Two identical vehicles can carry very different risk depending on how, when, and where they are driven.That gap between pricing inputs and real-world exposure is now forcing a reset.

For a long time, motor insurance underwriting in India was built on broad classification. Vehicles were grouped into categories, and pricing was applied at a portfolio level. This approach worked when data availability was limited, but it came with an inherent constraint. Individual driving behaviour and other measurable risk indicators remained outside the underwriting lens.

Even after pricing flexibility was introduced in segments like Own Damage nearly two decades ago, underwriting continued to rely largely on these static constructs. What is now emerging is a more refined approach that incorporates how a vehicle is actually used, not just what it is.

This shift is visible in the growing adoption of usage-linked constructs such as Pay-As-You-Drive. These models introduce a more direct relationship between usage and premium. A vehicle that is driven less, or used in a more predictable manner, can be priced differently from one that is exposed to higher usage and variability.

Beyond usage, insurers are also beginning to incorporate a wider set of customer-linked parameters into underwriting. This includes factors such as driving patterns, historical behaviour, and contextual usage signals. The objective is not just segmentation, but better alignment between risk and pricing.

Intelligence Reshaping Underwriting

Artificial intelligence and machine learning will enable this transition at scale. Traditional underwriting relied on a limited set of variables and relatively simple models. Today, insurers can process larger datasets and identify patterns that more closely reflect real-world risk.

What is emerging is more precise underwriting, where relevant, behaviour-linked variables carry greater weight than broad static inputs. This improves the ability to differentiate between risks that would otherwise appear identical.

The impact is visible across the value chain. Better risk selection improves portfolio quality. Advanced analytics also strengthen fraud detection by identifying anomalies that are difficult to detect manually.

Underwriting, as a result, is gradually shifting from a periodic exercise to a more continuous and responsive process.

Regulation as a Catalyst

This transition has been supported by a regulatory approach that has encouraged measured innovation. Sandbox frameworks introduced by the regulator allowed insurers to test new underwriting models in controlled environments. These experiments provided early validation for usage-linked constructs in the Indian market.

The subsequent approval of Pay-As-You-Drive products marked a clear step toward mainstream adoption. At the same time, the broader structure of motor insurance remains intact. Mandatory third-party coverage under the Motor Vehicles Act continues to serve its core social objective of protecting accident victims. This balance between innovation and stability is critical as the industry evolves.

The Compounding Value of Better Data

One of the defining characteristics of robust, real-time data-led underwriting is the advantage it builds over time. Each additional data point improves the ability to assess risk more accurately. Better risk assessment leads to more precise pricing, which in turn improves portfolio quality.

This creates a reinforcing cycle. Insurers that invest early in building data capabilities are likely to benefit from progressively better outcomes. Unlike traditional inputs, behavioural and usage-linked insights become more valuable as they accumulate.

At its core, this shift is changing what motor insurance represents. The emerging models are more adaptive, with pricing increasingly reflecting how a vehicle is used over time.

This creates the possibility of a system that is not only more accurate, but also more equitable. It aligns cost more closely with usage and behaviour, rather than broad categorisation.

Moving toward data-led risk assessment is about building a system that reflects how risk actually plays out on the road. The future of underwriting will rely less on static proxies and more on measurable usage and behaviour. That shift is already underway.

Paras Pasricha is the Head of Motor Insurance of  Policybazaar.Views expressed are the author's personal.