Insurance Core Platforms in the Age of Agentic AI - Part 2

Our Solutions

Case Studies

Resources

About Us

Book a Demo

Home

Products

Services

Customers

Blog

Pricing

See a demo

May 09, 2025

Insurance Core Platforms in the Age of Agentic AI — Part 2

Modernization Approaches to Prepare for an AI-First Future

Introduction

In Part 1 of this two-part series on Insurance Core Platforms in the Age of Agentic AI, we explored technological success factors and requirements of core insurance platforms for an agentic AI future (see Illustration 1 for an overview)

The Digital Opportunist archetype refers to insurers that have recently embarked on their digital journey. This may include insurers that are interested in expanding their digital footprint, are mindful of staying current with the digital trends, and may have initiated small tactical plays.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

ILLUSTRATION 1 : Essential insurance core platform capabilities to scale Agentic AI

In this Part 2, we discuss key challenges faced by insurers with regards to their core systems and explore different modernization approaches and no-regret moves to meet the requirements set out to successfully scale AI agents. We contrast the benefits and challenges of following a “hollow out the core” approach with a middle-office and addressing core modernization more holistically by adopting an end-to-end, AI-ready core platform.

Do you want to evaluate the AI-readiness of your existing core platform or a new vendor solution, please see this executive checklist for insurance leaders: Is My Core Insurance Platform Ready for Agentic AI? – A Checklist for Insurance CEOs and CIOs

Recommendations

Assess your current systems and potential new vendor solutions against the requirements set out in Part 1 of this paper and your own business requirements to identify key gaps.

Assess the change complexity based on the gap analysis and your internal change capacity to decide on the right modernization approach (i.e., middle-office vs. holistic core modernization).
Keep continued legacy dependencies and duplications in mind when deciding on the right modernization approach.
Systematically evaluate potential solution providers against the requirements set out in Part 1 (and the separate checklist) and rigorously test providers’ marketing promises.
Think in terms of long-term optionality and credible provider roadmaps, putting more emphasis on the future-proof architecture than pure functional requirements.
Accept that no solution meets all requirements fully out-of-the box; be ready to co-innovate together with other insurers and solution providers, adopting a product and not a project mindset.
Prioritize modular solutions that can grow with your needs and avoid throw-away investments into pure middle-office solutions.
Insurers with operations across countries: Strive for solutions supporting multi-tenant and multi-country deployments to multiply benefits by making it easier to build reusable AI agents and train and improve them on larger data sets.

The Current Challenge of Legacy Technology

Many insurers continue to rely on mainframe-based core systems originally deployed decades ago. These legacy systems frequently fall short of meeting the requirements outlined in Part 1. Primarily, these platforms are monolithic, not designed for real-time data exchange, nor equipped for the API and microservices interoperability essential for agentic AI.

True microservice architectures remain rare exceptions rather than standard practice. Typically, core insurance systems depend on batch processes, instead of real-time transactions. Many processes and functionalities depend on graphical user interfaces (GUI) and are designed solely for human interaction. Most core systems lack the scalability and face severe performance limitations under heightened concurrency demands.

Additionally, insurers typically operate multiple core systems for different product lines and functional modules. The resulting fragmented data repositories and inconsistent taxonomies hinder the unified, comprehensive views that AI agents require. Consequently, poor data quality emerges as a critical obstacle—reinforcing the adage, "garbage in, garbage out."

Given their age, many legacy systems have undergone extensive customizations over time, complicating efforts to retire or replace them. Many insurers have pursued incremental modernization, adding layers or modules to legacy cores for functions like digital interfaces rather than comprehensive replacements. Even when full system replacements have occurred, the technology stacks behind these vendor solutions were often outdated upon implementation.

Most notably, many core platform vendors still do not offer genuine microservices architecture, despite longstanding claims of progress. Gartner® highlights this issue clearly, noting, "However, this has been a common theme in vendor discussions for a number of years, and little tangible progress has been made on releasing true microservices." ¹

This underscores the necessity of considering modern, AI-ready core solutions, while choosing the right modernization approach.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Choosing the Right Modernization Approach

When addressing the current shortcomings in legacy insurance core systems, two primary approaches can be considered:

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Hollowing Out the Legacy Core with an AI-Ready Middle-Office

This approach involves gradually replacing or encapsulating legacy functionalities with modern layers, such as an API-based middle-office or orchestration solution. The legacy core is effectively relegated to a system of records, while business logic and functionalities reside in the new middle-office layer.

However, significant drawbacks exist. Dependency on legacy systems remains, adding complexity and ongoing maintenance costs. Even if the middle-office provides APIs or even MCPs (or other future AI-specific protocols) to facilitate interaction with AI agents, the middle-office still needs to be integrated with the legacy systems, which lack the integration capabilities.

Configurations and customizations often need duplication across the middle-office and legacy systems, multiplying efforts and costs. Moreover, middle-office solutions may encounter performance limitations and transaction constraints inherited from the legacy backend, while real-time data synchronization remains a challenge with batch-based processes in the legacy backend.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Replacing the Legacy Core with an AI-Ready End-to-End Core Platform

This comprehensive approach replaces outdated legacy systems entirely with a modern, AI-ready insurance core platform. Modernization can proceed horizontally (by product line across the full functionality), vertically (by functional capability across all product lines), or rarely as a complete ‘big bang’ transformation.

Opting for such comprehensive modernization allows insurers to establish a future-proof architecture—without the challenges of middle-offices outlined above. It eliminates dependencies and duplicated efforts associated with maintaining legacy systems.² Insurers can gain significant benefits in efficiency, product innovation, agility, accelerated value realization, resilience, and security even before fully implementing agentic AI.

Nevertheless, this approach demands a higher initial investment, substantial organizational commitment, and a longer timeline. It is inherently riskier, especially during the migration of legacy products, rules, and calculations, which are often poorly documented (particularly, for   life insurance).

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Considerations for Choosing the Right Approach

The appropriate approach varies based on an insurer’s context and strategic priorities. A middle-office can temporarily alleviate the symptoms without addressing the root cause.

Gartner mentions that "However, buyers should note that a middle office does not remove the need to modernize the core." As such, Gartner recommends: "Buyers may want to evaluate a middle-office option for more immediate needs, but should do so with the longer-term goals and objectives of their modernization strategy in mind." ³

An investment into a middle-office may be the right tactical choice, but it should not be made as a throwaway investment. Once an insurer inevitably initiates a full core modernization, the investments into the middle-office should be re-usable and not simply discarded. As such, the middle-office should not be a dedicated standalone solution but rather individual microservices of an end-to-end, AI-ready cloud platform. Such insurer could gradually adopt more and more functional capabilities (i.e., additional microservices and modules) of the core platform. This way, it becomes more of a gradual vertical modernization across functional modules, making a middle-office strategy a no-regret move.

Many insurers have been shying away from core modernizations due to the complexity of migration. Particularly for life insurers, migration is steeped with complexity and risk of migrating often hundreds of poorly documented product versions (often even requiring code re-engineering to understand business logics and rules). However, AI-assisted migration tools promise to lower overall risk, shorten timelines, and enhance accuracy in system migrations. The true benefits at scale are yet to be proven.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

ILLUSTRATION 2 : Comparison of modernization approaches

Build, Buy, or Partner?

Traditionally, insurers have relied heavily on custom-built, in-house solutions or extensively customized vendor solutions. Such highly tailored systems often become challenging to maintain and upgrade. However, in a rapidly evolving world—whether it is AI, other technological, environmental or societal changes—insurers require a core tech stack that evolves with such needs. So, it is not just about building a core platform once but maintaining and consistently enhancing it. As a single insurer (or even a large group), this is difficult and expensive.

AI-first insurers require a modern tech stack capable of evolving continuously. Most insurers do not have the talent for modern tech stacks. Not because they have unskilled talent, rather because they have talent highly skilled for the legacy world. If an insurer’s legacy core runs on COBOL, they need a COBOL specialist. If the insurer runs an on-premises mainframe, it does not need and does not have cloud experts.

All this makes it challenging for an insurer to build AI-ready applications. Transitioning to AI-ready platforms would demand significant investment in both new technology development and workforce up-skilling or extensive new hiring to maintain legacy operations simultaneously.

Given these challenges, acquiring a robust vendor solution typically represents a more viable path than building solutions in-house. Nonetheless, given rapid innovation and evolving requirements, no off-the-shelf solution is immediately perfect. Insurers can share the cost of research and development with their peers, while focusing on innovation on top of the core platform to achieve differentiation.

Strategic partnerships thus emerge as ideal, where insurers partner with solution providers willing to co-innovate and jointly enhance the platform. Such collaboration must naturally happen on product-level and not project-level.

When selecting a technology partner, insurers should prioritize demonstrated expertise across the success factors and requirements outlined in Part 1 of this series, credible technological roadmaps (including a history of delivering upgrades), and willingness to engage in co-investment and co-development. Strategic alignment ensures sustained technological relevance and agility in adapting to future AI-driven advancements.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Excursion: Regional Modernization and Scalable Agentic AI Strategies

Investments into agentic AI, whether the actual agents or the supporting core platform, can be significant—particularly for insurers already struggling with increasing IT budgets at the rate required to keep pace with technology change and inflation. The best investments are those that can be shared and re-used across entities and borders.

Historically, most insurers—even those belonging to large multinational groups with significant centralized functions—have struggled to realize meaningful synergies across borders, whether in products, operations, or technology. This is not just a regulatory and organizational topic. It is about disparate tech stacks across entities and countries that prevent synergies and scalability.

To enable a synergistic multi-country set-up, products and processes should be harmonized
(to the extent possible to meet local customs and regulations). This harmonization becomes significantly easier if the tech stack and data governance are also harmonized. As such, the insurance core platform needs to support multi-carrier, multi-currency, multi-language, multi-time zone, and multi-regulatory requirements.

Configurations should be easily transportable from one country entity to the next—without the need to rebuild everything from scratch. Future insurance groups may have a global insurance product repository from which local businesses can pick and choose the relevant products and apply localizations via configuration. This requires a platform that runs on a single regional or global code base.

In a non-AI context, multi-tenancy already drives significant value. Not only does it ensure
that all country tenants run on the same version, facilitating the transferability of propositions, workflows, and integrations from one country to another. It also drives more efficient management and operations, as not only can the cloud infrastructure be shared but also technical and operations staff.

In an AI context, this value can be multiplied. Different countries running on the same tech stack and a similar data architecture will not just make it easier to build reusable AI agents but also train and improve them on larger data sets. Many operations, including the AI agents, may be centralized as regional harmonization accelerates.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Investments into agentic AI, whether the actual agents or the supporting core platform, can be significant—particularly for insurers already struggling with increasing IT budgets at the rate required to keep pace with technology change and inflation. The best investments are those that can be shared and re-used across entities and borders.

Historically, most insurers—even those belonging to large multinational groups with significant centralized functions—have struggled to realize meaningful synergies across borders, whether
in products, operations, or technology. This is not just a regulatory and organizational topic. It is about disparate tech stacks across entities and countries that prevent synergies and scalability.

To enable a synergistic multi-country set-up, products and processes should be harmonized (to the extent possible to meet local customs and regulations). This harmonization becomes significantly easier if the tech stack and data governance are also harmonized. As such, the insurance core platform needs to support multi-carrier, multi-currency, multi-language, multi-time zone, and multi-regulatory requirements.

Configurations should be easily transportable from one country entity to the next—without the need to rebuild everything from scratch. Future insurance groups may have a global insurance product repository from which local businesses can pick and choose the relevant products and apply localizations via configuration. This requires a platform that runs on a single regional or global code base.

In a non-AI context, multi-tenancy already drives significant value. Not only does it ensure that all country tenants run on the same version, facilitating the transferability of propositions, workflows, and integrations from one country to another. It also drives more efficient management and operations, as not only can the cloud infrastructure be shared but also technical and operations staff.

In an AI context, this value can be multiplied. Different countries running on the same tech stack and a similar data architecture will not just make it easier to build reusable AI agents but also train and improve them on larger data sets. Many operations, including the AI agents, may be centralized as regional harmonization accelerates.

Peak3 and Graphene: Your Enablers  for the Agentic Future

Graphene is Peak3's intelligent, connected and AI-ready cloud core platform. Its end-to-end but modular capabilities span the entire insurance value chain—across Property & Casualty, Life, and Health lines.

Graphene delivers the AI-ready capabilities outlined in Part 1 of this series on Insurance Core Platforms in the Age of Agentic AI. Furthermore, it includes a built-in AI agent orchestration platform, allowing insurers to centrally configure and manage AI agents, integrate with foundational models, manage RAG pipelines, and so on. In addition, Graphene provides pre-integrated third-party and Graphene-native AI capabilities such as fraud, waste and abuse detection, intelligent document processing, and intelligent chatbots.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

ILLUSTRATION 3: Simplified overview of Graphene by Peak3

Purpose-built from the outset on microservices architecture, Graphene enables flexible modernization strategies and can be deployed flexibly to accommodate your specific circumstances. It can act as a middle-office, replace individual vertical functional modules (e.g., claims), or fully replace and unify your legacy systems horizontally across product lines.

Graphene supports multi-tenant and multi-country deployments. This allows you to build efficient and scalable business models across borders. Graphene is delivered as software-as-a-service (SaaS). This way, you can get started quickly with limited upfront commitment and investments. Furthermore, through regular upgrades, your core stays secure and evergreen.

Download Part 1 of the Series

Download Part 1

Download Part 1 of the Series

Ready to build your AI-first operations and business models or otherwise modernize your legacy core architecture? Please reach out to us here or at hello@peak3.com.

________________________________________________________________________

Interested in embarking on an innovative partnership journey? Get in touch with us here.

________________________________________________________________________

About FPT

FPT IS is a leading provider of digital transformation products, solutions, and services in Vietnam and in the region.

Ready to build your AI-first operations and business models or otherwise modernize your legacy core architecture? Please reach out to us here or at hello@peak3.com.

_______________________________________________________

References

¹ Gartner, Market Guide for Non-Life Insurance Core Platforms, Europe by Sham Gill and James Ingham, 4 March 2025 (for Gartner subscribers only). GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.

² A significant part of value creation can come from the decommissioning of legacy systems. While some parallel operations of legacy and new systems are unavoidable, a core modernization strategy should put similar focus on the decommissioning of the legacy as on the launch of the new platform.

³ Gartner, Market Guide for Life and P&C Insurance Core Systems, APAC by Richard Natale and Sham Gill, 11 February 2025 (for Gartner subscribers only)

________________________________________________________________________

About Peak3

Founded by insurance, digital and technology experts, Peak3 powers the digital operating system of the global insurance industry. We combine insurance core, distribution, and AI solutions to deliver a step change in performance for insurers, MGAs, and insurance intermediaries.

From greenfield embedded insurance ventures to digital-first, multi-country core modernization programs, our cloud-native SaaS solutions power top customers across life, health, and P&C insurance.

However, they retain some level of scepticism of the potential of digital or the need for digital transformation. For this reason, they have not allocated substantial resources towards these efforts.

Ready to build your AI-first operations and business models or otherwise modernize your legacy core architecture? Please reach out to us here or at hello@peak3.com.

_________________________