PwC’s Kumarappan Sundaram says insurers should separate rating logic from policy systems, improve data flows and governance, then add advanced analytics and AI.
How quickly an insurer can put an approved rate change into use may reveal more about its pricing capability than the sophistication of its analytics. PwC Managing Director Kumarappan Sundaram says carriers should first measure that delay and address the systems that cause it. His recommended starting point: separate rating logic from the core policy administration system.
The issue matters because pricing assumptions can lose relevance as conditions shift. The source identifies social inflation, higher repair costs, tariff pressure and catastrophe losses that challenge historical patterns. A rate change may be approved, yet still take weeks or months to reach customers while it waits for a policy system release.
That lag can constrain how quickly a carrier responds to new information. Sundaram’s advice is to establish the basic technology and operating practices first, then add advanced analytics and AI where they can inform decisions and support implementation. The sequence is a practical one: understand the delay, free rating logic from system release cycles, improve data flows, set governance, and then expand analytical tools.
Start by measuring the delay
Carriers may not know how long it takes for an approved pricing change to become available in the market. Sundaram recommends calculating that interval before choosing new tools. The result can help show whether the main constraint lies in technology, handoffs between teams, or the process for deploying and monitoring changes.
In many carriers, actuaries develop the pricing logic and pass it to IT teams, which place it in a rating engine inside the policy administration system. When the logic is tightly tied to that system, a change may depend on a new system release. That can turn a pricing update into a broader technology deployment, slowing its arrival.
Decoupling rating logic gives insurers a way to change ratings without waiting for the policy administration system’s release schedule. Sundaram describes this as the foundation for modernization. It can make rate updates more responsive, while still leaving carriers responsible for managing how changes are approved, deployed and monitored.
Build the foundation before adding AI
Once rating logic is separate, Sundaram says carriers should focus on dependable data flows and clear governance. Those elements help teams understand what information informs a pricing decision and how a change moves from analysis into use. Without them, advanced tools may be difficult to apply consistently across the business.
Analytics can then help actuaries estimate the effects of a proposed rate factor change across a portfolio. Sundaram gives a hypothetical example: a carrier with a $10 billion book is considering changing a factor tied to driver age. Advanced analytics could help assess the impact across the whole book; actuaries would use that analysis to decide whether to make the change. The technology can also help carriers deploy and monitor pricing adjustments more quickly and safely.
That example illustrates a supporting role for analytics. It can give decision-makers a clearer view of potential portfolio effects, but it does not make the pricing decision for them. Sundaram draws a distinction between capabilities that are already useful and AI applications whose value remains unproven. For carriers, the practical question is where a tool improves a defined part of the work and whether the underlying data and processes are ready to support it.
Modernization varies by line
The transition is uneven across property and casualty insurance. Sundaram says personal lines are further along, while commercial lines are earlier in the process and still depend heavily on underwriter judgment and rating engines embedded in policy administration systems. That difference shapes what modernization may require: a carrier’s starting point depends on its existing systems and how pricing decisions connect with underwriting.
Modern platforms may also help narrow the longstanding gap between pricing and underwriting. That connection matters because pricing decisions and assessments of individual risks inform related parts of the insurance process. Improving the systems that carry rating logic and data can make it easier for teams to work with current information, though the source does not specify a particular platform or deployment timetable.
For executives, the takeaway is to treat pricing as a strategic capability rather than only a back-office technology concern. Measuring the time between approval and market availability gives teams a concrete place to begin. From there, they can identify system dependencies, strengthen data movement and governance, and evaluate advanced analytics or AI against specific needs.
Related coverage
Other articles featured with the pricing discussion examine continuous underwriting, AI-assisted underwriting, organizational silos, submission handling and the use of technology amid uncertainty:
- Continuous Underwriting Wants to Scale, by Riv Arthur, discusses barriers to scaling continuous underwriting.
- Agentic AI Redefines Real-Time Underwriting, by Rachita Nahata, looks at using current data to help underwriters assess risk faster.
- Breaking Down Silos in Insurance, by Rory Meleniclis, considers product teams organized around shared capabilities such as underwriting and claims.
- How AI Can Transform Submissions for Underwriting, by Prathap Gokul, focuses on prioritizing submissions for underwriting attention.
- Making AI Work in Commercial Submission Intake, by Vital Soupel, discusses where to apply AI and how to scale it.
- Turning Volatility Into Value, by Insurance Thought Leadership, features Dawn Miller, chief commercial officer at Lloyd’s.
The featured sponsor material from PwC concerns AI adoption and workforce change:
The full interview with Sundaram is available at Pricing at the Speed of the Market. Its central question for carriers is concrete: how long does a rate change take to reach the market, and which system dependencies account for the wait?
