PUBLISHED DATE: 2026-04-25 04:07:36

How Regulated Enterprises Modernize Legacy Systems Safely

Real case studies showing how AI reduced risk and increased speed for financial services, healthcare, energy and commodities industries

Table of contents


Industry case studies

Can you automate change without increasing risk?

Regulated organizations do not delay modernization because they lack ambition. They delay because failure carries severe consequences: regulatory findings, customer or member harm, security exposure and operational disruption.

In financial services, health and energy, legacy modernization is not just a technical exercise. Teams must also be able to prove—to auditors, regulators and internal risk committees—that system behavior, data handling and controls remain intact after change. Speed without proof increases risk. And historically, proof has been slow, manual and fragile.

This playbook documents how large, regulated enterprises can modernize mainframe and legacy application platforms with less risk for the first time with the help of AI.

The cases below are based on real-world, AI-enabled modernization programs where regulatory scrutiny was constant, and tolerance for error was low.

The core insight: slower isn’t safer

Across every story, risk went down when large technology systems became more observable, more testable and more governable before change.

AI’s purpose in modernization was not just “coding faster.”

Its true value was:

The five biggest risks of traditional, manual legacy modernization

In each case, you’ll see that AI was applied specifically to reduce one or more of the five biggest real-world factors threatening large enterprises during a “status quo” legacy modernization:
  1. Unintended rule changes
    Regulated business rules are reimplemented without shared understanding of the real-world system logic, leading to unintended changes in claims, payments, coverage or eligibility.
  2. Undocumented dependencies caused downstream failures
    Hidden system and data dependencies trigger outages, rollbacks and loss of confidence in future change.
  3. Extended timelines increased exposure
    Partially modernized systems remain in production too long, prolonging regulatory exposure and delaying return on investment.
  4. Security and data-handling exposure
    Refactoring code introduces new vulnerabilities, expanding audit scope and regulatory exposure.
  5. Lack of audit-grade traceability
    Teams cannot demonstrate how requirements, code and tests align, forcing manual reconstruction of compliance evidence.
For each industry, reducing these risks through governed automation delivered measurable business outcomes: faster returns on new technology, reduced subject-matter expert (SME) dependency and millions in avoided technical debt. This playbook highlights how Sapient Slingshot, Publicis Sapient’s enterprise AI platform for software development, made these outcomes possible and how executives should scope pilot projects in their organization to do the same.

Industry case study 1: financial services

U.K. retail and commercial bank

Converted half a million lines of code to verified specifications in eight weeks

Industry context

The company: Large U.K. retail and commercial bank modernizing legacy banking services under regulatory oversight

What was at stake

Core services are classified as “important business services” under U.K. operational resilience requirements. A small defect introduced during modernization that affects reporting, payments or customer processing could trigger:

Status quo approach

Traditional code-to-spec efforts rely heavily on manual, human analysis and SMEs. This approach is slow, inconsistent and very difficult to scale. Specifications (i.e., descriptions of what a system is supposed to do) in large enterprises are often incomplete, out-of-date or disconnected from the underlying code, increasing the likelihood of missed logic and rework during later phases of the program, as well as the risks stated above from slowing down the initiative.

Why AI was necessary

The bank could not safely modernize ~500,000 lines of code tightly coupled with mainframe logic using manual interpretation without risking unintended rule changes or increased compliance exposure.

AI-enabled modernization approach

In this case, Sapient Slingshot, Publicis Sapient’s enterprise AI software development platform, was applied specifically to extract the rules and behavior buried in code, and document them clearly. Because of this, the bank was able to:
  1. Restore visibility before change
    The platform rapidly analyzed a large volume of existing production code to extract hidden business logic and automatically generate traceable, reviewable specifications.
  2. Establish understanding of system behavior
    In just eight weeks, Slingshot drove code-to-spec efforts for 237 programs and 327 feeds, amounting to nearly half a million lines of code across mainframe batch feeds for financial and data products, along with the bank’s payments module.
  3. Design from validated specifications
    Slingshot produced business specifications, designed a modern target-state architecture and revamped the underlying data model.

Outcomes

Industry case study 2: financial services

Large Middle East bank

Stabilized 30+ systems and cut release risk under regulatory scrutiny

Industry context

The company: Large Gulf-region retail and commercial bank operating under regional banking oversight

What was at stake

A vendor’s exit left the bank with 30+ fragmented services, a weak security posture and less than five percent of planned scope delivered—all under the scrutiny of regional banking regulators. Long release cycles risked turning fixable defects into regulatory events, and a months-long remediation gap is exactly what regulators would flag as “failure to remediate a control weakness in a timely manner”, triggering:

Status quo approach

Without automation, modernizing 30+ interconnected banking services is linear, manual and prone to human error.
  1. It starts with manual reverse engineering of code, architecture discovery done by human experts, manual backlog rebuilds and tech debt compounding.
  2. The team, scope and costs expand, increasing integration defects.
  3. At the end of project, teams can’t guarantee “doneness,” and risk slightly out-of-sync, crucial systems that are still reliant on the legacy core.

Why AI was necessary

Manual modernization of 30+ interdependent services made it impossible to move faster and still maintain strong oversight.

AI-enabled modernization approach

Implementing Sapient Slingshot, Publicis Sapient’s enterprise AI software development platform, the bank was able to:
Stabilize all legacy services
Identify design flaws, improve architectural visibility, map and clarify service dependencies and reduce defect carryover.
  1. Increase test coverage and control maturity
    Achieve 80 percent+ unit test coverage, automate regression generation, reduce defect rate by 30 percent and enforce compliance and security rules.
  2. Compress review and release time
    Shorten review cycles, streamline approval cycles, reduce testing bottlenecks and improve traceability across the whole lifecycle.
  3. Improve delivery efficiency
    Streamline the backlog-to-deployment cycle, meaning less rework and fewer manual bottlenecks.

Outcomes

Industry case study 3: health

U.S. health insurance company

Compressed 10-year claims modernization to ~three years without compliance drift

Industry context

The company: Leading U.S. health insurer processing billions of claims annually under CMS and HIPAA oversight

What was at stake

Claims adjudication logic was embedded across 10,000+ pages of COBOL on a legacy mainframe. At this scale, minor defects can trigger major compliance events—improper denials, provider underpayments or PHI exposure.

Potential consequences included:

Status quo approach

Traditional modernization focuses on rewriting code—not proving equivalence. After three years, only ~10% of features had been converted manually.

Risks included:
This forced a defensive delivery posture: Teams slowed progress to reduce error risk, extending the modernization timeline and prolonging exposure.

Why AI was necessary

Proving that the new system behaves the same way as the old system (behavioral equivalence) across 10,000+ pages was necessary, yet infeasible through manual review alone.

AI-enabled modernization approach

By implementing Sapient Slingshot, Publicis Sapient’s enterprise AI software development platform, the organization was able to:
  1. Complete discovery and rule extraction before change
  2. Phase modernization with continuous validation
    Modern Java and React microservices were generated from validated specifications.

    For every migrated feature, the team:No feature was complete without proven behavioral parity.
  3. Embed security and control into delivery

Outcomes

Industry case study 4: health

U.S. pharmacy benefits manager

Modernized 100TB financial system in under three years without breaking contracts or compliance

Industry context

The company: Large U.S. PBM managing billions in rebates and federal/state reporting requirements

What was at stake

The risks were technical and financial:
A misapplied contract term or subtle shift in eligibility logic could distort an entire quarter’s invoicing. A downstream allocation error could affect thousands of employer groups. The organization could not afford unintended rule changes and a five- to seven-year dual-platform transition.

Status quo approach

Why AI was necessary

Manual lifecycle reconstruction of rebate logic across 100TB of data was operationally impractical.

AI-enabled modernization approach

Sapient Slingshot, Publicis Sapient’s enterprise AI platform for software development, did the opposite:
  1. Discovery and rebate rule extraction complete up-front
  2. Financial lifecycle sequencing
  3. Financial regression and governance

Outcomes

Industry case study 5: health

Medicare enrollment platform

Modernized a critical Medicare enrollment platform without risking coverage loss

Industry context

The company: U.S. health care organization managing Medicare enrollment and billing

What was at stake

The risks were technical, financial and regulatory:
A subtle eligibility logic shift could result in wrongful coverage termination. The organization could not risk unintended rule changes—or a five- to seven-year dual-platform transition.

Status quo approach

Why AI was necessary

Manual reconstruction of enrollment and eligibility dependencies risked coverage gaps at scale.

AI-enabled modernization approach

Sapient Slingshot was deployed inside the client’s regulated environment, integrated with their AI Studio and certified LLM models, enabling:
  1. Eligibility and workflow discovery
  2. Enrollment workflow sequencing
  3. Behavioral regression and governance

Outcomes

Industry case study 6: energy and commodities

Energy infrastructure

Revived a 25-year-old black-box application without rebuilding it from scratch

Industry context

The company: European energy producer operating gas and generation assets where system reliability directly impacts operational continuity and financial performance

What was at stake

For a European energy producer, modernization missteps in core operational systems didn’t just risk project failure—they risked grid stability, regulatory compliance and continuity of service. If an application fails or remains out-of-date, an energy organization might face:
However, rebuilding an application from scratch without fully understanding its original business logic risks recreating the same errors.

Status quo approach

This application exists only as compiled binaries, basically machine code (binary instructions made of 0s and 1s) that a computer’s CPU can execute directly. Therefore:

Why AI was necessary

Traditional modernization would require weeks of manual reverse engineering by senior engineers with no ability for leaders to measure success or completeness.

AI-enabled modernization approach

The team applied a structured five-step process accelerated by Sapient Slingshot, Publicis Sapient’s enterprise AI platform for software development.
  1. Decompilation and recovery
    Using open-source AI tools, binary files were converted back into readable Java source code to restore the foundation for modernization.
  2. Environment rebuild
    A modern runtime environment (Java 17 and PostgreSQL 16) was created so the application could operate on current systems.
  3. Refactoring and cleanup
    Sapient Slingshot was used to restructure and modernize the recovered codebase, reducing over 9,000 lines to approximately 5,000 clean, readable lines with updated syntax and standards.
  4. Business logic extraction
    AI automatically generated entity-relationship diagrams and data-flow mappings, revealing the application’s functional logic, something no one in the entire company could previously access.
  5. Documentation and testability
    Inline documentation and standalone artifacts were automatically generated, transforming the app from opaque code into a maintainable system.

Outcomes

Industry case study 7: energy and commodities

Energy & utilities

Migrated more than 400 APIs without losing regulatory lineage

Industry context

The company: A multinational energy and utilities company operating regulated electricity and gas infrastructure in the United States is subject to strict federal and state oversight, including:
In this environment, system traceability, data integrity and operational continuity are regulatory obligations—not business preferences.

What was at stake

The company relied on a large, aging API estate to move operational and renewable-generation data across IT and OT systems supporting grid operations, renewable asset ingestion, regulatory reporting and security workflows.

Modernization carried significant risk: breaking generation and reporting integrations, losing required audit lineage and weakening visibility into how regulated data was transformed. The organization needed to scale renewable data ingestion without increasing compliance exposure.

Status quo approach

In large enterprises, APIs are deeply embedded across legacy systems, reports and partner integrations—often with incomplete documentation. Changes require manual coordination and heavy oversight to avoid disruption.

However, a manual approach often means:

Why AI was necessary

The volume of APIs and interconnections made it impractical to manually trace data origins, transformations and dependencies across regulated systems.

AI-enabled modernization approach

The organization applied a modernization model supported by Sapient Slingshot, Publicis Sapient’s enterprise AI platform for software development, that ensured:
  1. A defined scope for the project
    A bounded API domain supporting regulated operational data flows was selected.
  2. Controls before change
  3. Governed migration

Outcomes

What makes a successful pilot for AI-enabled modernization?

In every case documented in this playbook—whether it was a bank, health insurance company or energy producer—the organization began with a deliberately constrained pilot designed to reduce risk before increasing speed:
  1. Pilot scope is intentionally narrow
    A pilot for AI-enabled legacy modernization should focus on:
  2. Controls are established before code changes
    Before any code refactoring or platform migration:
  3. AI is governed, not autonomous
  4. Evidence is produced continuously
    As the pilot progresses, teams generate:This allows risk, compliance and audit teams to engage early with evidence.
  5. Success is defined by confidence, not speed
    In these pilots:
A successful pilot of AI-powered modernization delivers three things for the technical side and the business side.

How Sapient Slingshot systematically reduces risk

The decision facing technology leaders in regulated industries

The high-level outcomes from modernizing with Sapient Slingshot vary across projects, from millions in projected ROI due to faster technology upgrades, to seven years of an engineer’s time saved through mapping thousands of system and data dependencies. But in each scenario, this is exactly why and how utilizing Slingshot achieves safer, faster and better outcomes long-term:

These real stories prove that modernization will not get safer by waiting. In regulated industries, modernization only gets safer if it becomes more observable, more testable and more governed. The organizations profiled in this playbook did not have to accept more risk to move forward. They engineered risk down through automation, and speed naturally followed.

What actually matters

How business rules are handled
When risk is discovered
How compliance proof is created
Speed vs. safety trade-off
Understanding of system dependencies

About Sapient Slingshot

Sapient Slingshot is the only legacy modernization platform that automates the entire software lifecycle end-to-end. Unlike point AI coding assistants, Slingshot pairs a persistent enterprise context graph with specialized SDLC agents to modernize and deliver accurate, fast and governed software. This enables organizations to achieve up to 50 percent savings in modernization cost, 99 percent code-to-spec accuracy and 40 percent productivity gains in new software delivery.

To learn more about Sapient Slingshot, the platform used to accelerate legacy modernization across industries, visit:
https://www.publicissapient.com/platforms/slingshot