Insurance companies could lose years of potential pricing gains by making core system replacement a prerequisite for modernisation, according to Earnix principal solutions consultant for agentic and predictive AI Brian Shipley.
While replacing legacy technology can offer insurers a more flexible operating environment over time, large-scale programmes can also absorb budgets, stretch internal teams and delay the introduction of capabilities that could deliver business value sooner. Shipley argues that insurers should reconsider whether modernising pricing and rating systems independently could deliver faster returns while reducing the risks associated with replacing entire platforms.
As core insurance technology becomes more widely available, competitive differentiation increasingly depends on how effectively insurers develop products, set prices and respond to changing market conditions. Rules engines, automated filing, real-time rating and embedded decisioning are changing how insurers manage these activities. However, replacing a policy administration system (PAS) without rethinking the processes behind it risks carrying existing limitations into a new technology environment.
Separating pricing and rating from legacy core systems can allow insurers to introduce more advanced simulation capabilities, adjust rates more quickly and simplify filing processes without waiting for a complete platform replacement. This approach also allows modernisation to progress in stages, rather than making every improvement dependent on a single transformation programme.
Data migration presents another challenge. Insurers may underestimate the amount of proprietary pricing data and embedded business logic contained within legacy systems until a migration is under way. Moving information into platforms such as Snowflake or Databricks does not automatically resolve inconsistent field definitions, undocumented transformations or gaps in data lineage.
These issues can emerge during reconciliation, when pricing outputs from a new environment fail to match those produced by legacy systems. Teams must then establish whether discrepancies stem from data quality problems or differences in the underlying rating logic. Data models designed around existing products may also struggle to accommodate future requirements, particularly as artificial intelligence increases demand for more sophisticated risk segmentation, elasticity modelling and pricing optimisation.
The opportunity cost of a full core replacement can be substantial. Resources committed to migration are unavailable for other initiatives, while capabilities that could improve pricing performance remain tied to legacy infrastructure.
Shipley said that, based on his experience, a full policy administration system migration that includes pricing and rating at a large insurer can take three or more times longer to launch an initial line of business. Converting the entire book of business can extend the programme further.
An alternative is to move pricing and rating into a dedicated engine connected to the existing PAS through application programming interfaces (APIs). This allows insurers to modernise incrementally, with the pricing engine remaining in place even when the underlying policy administration system is eventually replaced.
Large-scale transformations can also place considerable pressure on employees. Staff may need to maintain legacy operations while simultaneously building and testing replacement systems, potentially for years. This can affect morale, knowledge retention and an insurer’s ability to attract and retain talent.
Research cited by industry commentators highlights the broader risks associated with major technology programmes. In a March 2026 article for Carrier Management, Tokio Marine Group group deputy chief information technology officer Robert Pick referenced research attributed to McKinsey and BCG suggesting that more than 70% of large-scale core transformations fail, with nearly a quarter written off entirely.
An Oracle and LumenData white paper citing Gartner reported that 83% of data migration projects fail or exceed their budgets and schedules. Separately, an analysis published by Objectway estimated that around 80% of core banking migrations fail. These figures cover different types of technology projects and should not be treated as directly comparable measures of insurance transformation outcomes.
Shipley recommends that insurance executives assess how much configuration, deployment and ongoing change their teams can manage independently. He also advises testing this through a proof of concept involving actuarial, pricing and business intelligence teams.
For insurers weighing their next technology investment, the decision is therefore not simply whether to replace legacy systems. It is whether the expected benefits justify delaying other improvements, and whether targeted modernisation could deliver value sooner while limiting the operational burden of a full-scale transformation.
Copyright © 2026 FinTech Global
