Agentic AI Could Deliver Up to 90% Productivity Gains in Insurance Core System Modernization
Agentic AI — autonomous or semiautonomous software agents that can interpret legacy systems, generate documentation, validate configurations and coordinate complex workflows — could improve productivity by 10% to 90% across various stages of insurance core system modernization, according to an April 2026 report from McKinsey’s Financial Services Practice.
The greatest gains come in testing, reconciliation and defect cycle compression, where the report estimates improvements of 15% to 90%, while discovery and reverse engineering of legacy systems could see 20% to 50% productivity improvement.
Why Core Modernization Has Stalled
Insurance core systems often comprise decades’ worth of sparsely documented business rules, batch processes, custom interfaces and data semantics, the report said. Executives have long recognized the need to upgrade these platforms, but structural costs and risks have repeatedly undermined their motivation to act.
Among the most persistent challenges are underdocumented product logic and actuarial settings, semantic gaps that surface late in migration projects and drive rework, and cutover risks that force conservative sequencing, according to McKinsey. These dynamics create costly “double-bubble” periods in which insurers pay to maintain their legacy systems while simultaneously funding the modernization program, the report said.
A critical but often overlooked factor is where effort actually concentrates during migrations. Rewriting code or configuring the target platform represents only a small portion of the work, the report found. A “disproportionate share of time and investment” typically goes toward understanding and configuring rules, data conversion, quality control, reconciliation, operational readiness and post-migration stabilization, McKinsey said.
How Agentic AI Differs From Existing Tools
Unlike developer copilots that assist users moment by moment, agentic AI systems are designed to pursue a goal, break it into tasks, use tools and context, and iterate based on feedback and controls, according to the report. In a policy administration migration, this distinction matters because the biggest bottlenecks are rarely writing code but rather “the loops of discovery, mapping, testing, reconciliation, and cutover,” McKinsey said.
Some of the most significant gains come from agents’ ability to decode outdated programming languages that few workers can still understand, the report said. Agents can read code written in archaic languages, reverse engineer the logic and convert it into plain English. In many cases, an agent can accomplish within days what would take a trained subject matter expert months or even years to complete, according to McKinsey.
Once agent capabilities are established, the incremental cost of modernizing additional products and systems can fall quickly because the same agents, patterns and context layers can be reused across waves and domains, the report found. This creates what McKinsey described as “a portfolio option that insurance technology leaders have not had before.”
Three Shifts for Capturing Value
The report identified three critical moves for insurers looking to apply agentic AI to modernization efforts.
First, McKinsey recommended building modular agents — treating them as “a library of atomic capabilities, each with clear inputs, acceptance criteria, and escalation paths to humans” — rather than deploying monolithic solutions. This approach improves control, makes outputs auditable and enables reuse across discovery, data, testing and cutover phases.
Second, the report urged a shift from viewing modernization as a single large-scale migration to managing it as a coordinated portfolio of opportunities. With a reusable agent stack in place, the incremental effort to modernize additional products or adjacent applications “has the potential to fall materially,” McKinsey said, allowing leaders to evaluate platform migrations alongside selective rewrites of long-tail legacy applications.
Third, McKinsey emphasized redesigning roles, governance and risk management for agentic execution. Because insurance modernization is regulated and operationally sensitive, agentic workflows need built-in controls such as human-in-the-loop approvals at stage gates, traceability from requirements to configuration to test evidence, and clear model-validation practices, the report said.
Read the full report here. &


