💡 왼쪽 원문을 읽으면서 오른쪽에 따라 써보세요. Tab 키로 힌트를 받을 수 있습니다.

원문 렌더가 준비되기 전까지 텍스트 가이드로 표시합니다.

Prologue — The Day 30-Year COBOL Stopped at 4 AM

In early 2026, a message popped up in the IT chatroom of a major Korean P&C insurer at 4:17 AM.

"The PG settlement batch stopped again. The RPG code on AS-400 can't handle the new payment code length. The person who wrote this left in 2008. The comments are in Japanese."

This one line captures a cross-section of the 2026 insurance core system market. Insurance is the most conservative IT industry, and policy data must live for 25–30 years (life insurance: a lifetime). That's why cores written on 1990s mainframes survived until 2026. And now the people who wrote them have retired.

This post draws the full map of the 2026 insurance core market. How Guidewire dominated the global P&C core market, how Duck Creek transformed into SaaS after its IPO, what positions Sapiens, Insurity, EIS, and Majesco occupy. FINEOS specialized in claims, Origami Risk in risk management, and Britecore for MGA and indie insurers. Together with Korea Samsung SDS insurance IT, KB Life's CIS case, and the 30-year trajectory of Japan's NTT Data and NRI ZAITAS.

Chapter 1 · What Is an Insurance Core System — A One-Sentence Definition

First, a definition. **An insurance core system is a bundle of systems of record that handles the entire lifecycle of insurance products — policy issuance, underwriting, rating, claims processing, payouts, billing.**

Traditionally, the insurance core splits into three core modules.

- **PAS (Policy Administration System)** — Policy issuance, renewal, endorsement, expiration. The heart of insurance.

- **Claims System** — Loss notification, investigation, settlement, payout. The largest cost center that determines loss ratio.

- **Billing System** — Premium billing, collection, installments, refunds. The cash flow backbone.

Around these sit underwriting workstations, rating engines, reinsurance systems, channel (agent/broker) portals, and customer self-service. When an insurer issues a 30-year policy, that policy data, claims history, and payment record must live in the core for 30 years. This is the core reason insurance cores were trapped on mainframes.

The 2026 market splits into two streams. **Global P&C cores** (Guidewire, Duck Creek, Sapiens, Insurity, EIS) and **Life & Annuity cores** (Sapiens CoreSuite, OIPA, Majesco L&A, Equisoft). Claims-specialty engines (FINEOS, Origami Risk) hold their own seats, and a new layer for MGAs and indie insurers (Britecore, Socotra) emerges.

Chapter 2 · P&C vs Life & Annuity — Essential Structural Differences

To understand insurance cores, you first need to grasp the structural difference between P&C and L&A. The two areas have different business models, and therefore completely different data models in the core.

**P&C (Property & Casualty)** — Non-life insurance. Auto, home, casualty, liability. Policy periods are short (typically 6 months to 1 year), renewal cycles are fast, claims are frequent and varied. Rating reflects market changes annually, and payouts are proportional to loss within set limits.

**L&A (Life & Annuity)** — Life insurance and annuities. Policy period lasts a lifetime (life insurance) or decades (annuities). Payouts trigger on specific events like death, surrender, or maturity, and reserve calculation, assumed interest rates, and mortality tables are central. Payouts to be made 30 years from now must be reserved in advance.

| Aspect | P&C | L&A |

| --- | --- | --- |

| Policy term | 6 months to 1 year | Lifetime to decades |

| Core data | Risk factors, claims history | Reserves, assumed rates |

| Claim frequency | High | Low (death, maturity) |

| Rate change | Annual | Almost none |

| Representative core | Guidewire, Duck Creek | Sapiens CoreSuite, OIPA |

| Key tables | Loss statistics | Mortality, life tables |

This structural difference drives the market split. Guidewire and Duck Creek focused on P&C and divided the market between them; Life & Annuity is a more fragmented market (Oracle OIPA, Sapiens CoreSuite, Majesco L&A, Equisoft).

Chapter 3 · Guidewire InsuranceSuite — The De Facto P&C Standard

Guidewire (founded 2001, NYSE IPO 2012, headquartered in San Mateo) is the de facto standard of the 2026 P&C core market. 50%+ of global top-tier P&C carriers use Guidewire as their core. Travelers, Nationwide, Liberty Mutual, Zurich, AXA, Allianz, and more.

Guidewire InsuranceSuite consists of three main products.

- **PolicyCenter** — Policy management. Product catalog, rating, policy issuance, renewal, endorsements.

- **ClaimCenter** — Claims management. Loss intake, FNOL (First Notice of Loss), investigation workflow, payout.

- **BillingCenter** — Billing and collection. Premium billing, payment, installment, refund.

Around these sit InsuranceNow (integrated SaaS for smaller carriers), DataHub (data warehouse), InfoCenter (analytics and BI), CustomerEngage (customer portal), and ProducerEngage (agent portal). Since 2020, integration into SaaS via the **Guidewire Cloud Platform** has been in progress, and as of 2026, 90%+ of new customers pick the cloud option.

// Guidewire Gosu — PolicyCenter rating rule (Gosu is the Guidewire-specific JVM language)

// Auto insurance — premium adjustment based on driver age and claims history

package rules.PersonalAuto

uses gw.api.system.PCLoggerCategory

uses entity.PersonalAutoLine

uses entity.PolicyPeriod

class AutoRatingRule {

function calculateBasePremium(period : PolicyPeriod) : java.math.BigDecimal {

var line = period.PersonalAutoLine

var driver = line.PersonalVehicles.first().Drivers.first()

var basePremium = 1200bd // base premium

// age factor

var ageFactor = driver.Age < 25 ? 1.5bd

: driver.Age < 65 ? 1.0bd

: 1.2bd

// claims factor — claim count in the last 3 years

var claimsCount = driver.PriorClaims.where(

\c -> c.AccidentDate > java.time.LocalDate.now().minusYears(3)

).Count

var claimsFactor = 1.0bd + (claimsCount * 0.3bd)

return basePremium * ageFactor * claimsFactor

}

Guidewire's strength is **product depth**. Over 30 years of evolving with insurers, it supports nearly every P&C variant (auto, home, commercial, marine, workers' comp). Drawbacks: it's expensive (license plus implementation consulting totals tens to hundreds of millions of dollars), implementation takes long (typically 2–4 years), and Gosu — a proprietary language — has a learning curve.

Chapter 4 · Guidewire Cloud Platform — 80% Cloud Target by 2030

Since 2020, Guidewire has declared a full cloud transition through the **Guidewire Cloud Platform (GWCP)**. Built on AWS as multi-tenant SaaS, 90%+ of new revenue in 2026 is cloud. CEO Mike Rosenbaum announced on the 2025 earnings call: "Our goal is to migrate 80% of all customers to cloud by 2030."

GWCP's architecture has three layers.

- **InsuranceSuite Cloud Service** — SaaS version of PolicyCenter, ClaimCenter, BillingCenter. AWS multi-AZ deployment, automatic backups and DR.

- **Cloud Console** — Insurance carrier IT manages its own instance (deployment, monitoring, access control).

- **Marketplace** — 700+ integration components (third-party intelligence, tax calculation, catastrophe modeling, etc.).

The migration path is typically three stages. (1) Lift-and-shift the on-prem InsuranceSuite to AWS. (2) Re-platform to the Guidewire Cloud Platform (use Cloud API, separate customizations via ContentLib and Studio). (3) Move fully to Guidewire's hosted Cloud Service (migrate code and data to Guidewire's multi-tenant platform).

As of 2026, global top-3 P&C carriers (Travelers, Nationwide, Liberty Mutual) have reached stage 3. Korea is still mostly between stages 1 and 2, and Japan has SOMPO and Tokio Marine in stage 2 migrations.

Chapter 5 · Duck Creek Technologies — The SaaS Pivot After IPO

Duck Creek (started as an internal Accenture unit in the 1990s, spun off in 2016, NYSE IPO in 2020, acquired privately by Vista Equity Partners in 2023) is Guidewire's most direct competitor. As of 2026, second in global P&C market share.

Duck Creek **OnDemand** is the name of the cloud-native SaaS platform, composed of three main products.

- **Duck Creek Policy** — Counterpart to Guidewire PolicyCenter.

- **Duck Creek Claims** — Counterpart to ClaimCenter.

- **Duck Creek Billing** — Counterpart to BillingCenter.

Around these sit Duck Creek Rating, Duck Creek Reinsurance, and Duck Creek Producer.

Duck Creek's differentiator is **Author, the visual development tool**. While Guidewire writes rules in Gosu code, Duck Creek writes rules in XML or a visual editor. As a result, non-developers (product managers, underwriters) have a high self-service rate for modifying rules. The downside: XML volume explodes for complex branching logic.

Since the Vista Equity Partners private acquisition in 2023, Duck Creek has rapidly grown its SaaS revenue share, and as of 2026, 80%+ is OnDemand SaaS. Duck Creek is said to be strong for simpler products and small to mid-size insurers.

Chapter 6 · Sapiens IDIT and CoreSuite — Multi-Line Global Heavyweight

Sapiens (headquartered in Israel, founded 1982, NASDAQ-listed) is a leading multi-line insurance core covering both P&C and Life & Annuity. Deployed at 600+ carriers globally, especially strong in Europe, Asia, and Africa.

Sapiens' two main products should be distinguished.

- **Sapiens IDIT** — P&C core. A full-stack P&C solution strong in Southeast Asia and Latin America emerging markets.

- **Sapiens CoreSuite for L&P** — Life & Pensions core. Adopted by large carriers in the UK and European markets.

Around these sit Sapiens DigitalSuite (customer portal, agent channel), Sapiens Decision (rules engine), and Sapiens Reinsurance.

Sapiens IDIT — product configuration YAML (low-code composition)

Auto insurance product definition — new market launch

product:

name: AutoComprehensive2026

line_of_business: PersonalAuto

effective_date: 2026-01-01

base_premium:

method: TerritoryBasedRating

factors:

- territory_code # per-region accident rate

- vehicle_class # sedan, SUV, truck

- driver_age_band # 21-25, 26-65, 65+

coverages:

- code: BI # Bodily Injury

limits: [100000, 300000, 500000]

default_limit: 300000

- code: PD # Property Damage

limits: [50000, 100000, 250000]

default_limit: 100000

- code: COMP # Comprehensive

deductibles: [500, 1000, 2500]

optional: true

underwriting_rules:

- rule_id: AGE_LIMIT

condition: "driver.age < 21 OR driver.age > 75"

action: REFER_TO_UNDERWRITER

- rule_id: PRIOR_LOSSES

condition: "driver.prior_claims_3yr >= 3"

action: DECLINE

Sapiens' strength is **global multi-country support**. Designed so a single core handles regulation, currency, and language across 50+ countries. Many cases exist where European multinationals like Allianz and Generali consolidated their core on Sapiens. Drawbacks: less U.S. P&C depth than Guidewire, and UI/UX is relatively dated.

Chapter 7 · Insurity — U.S. Mid-Market Specialist

Insurity (formed via 2000s mergers, owned by GTCR private equity) is a heavyweight in the U.S. mid-market P&C space. Smaller in revenue than Guidewire or Duck Creek, but with a 200+ carrier customer base, especially strong in specialty lines (commercial, marine, crop insurance).

Insurity's main product family comprises three.

- **Insurity Policy Decisions Suite** — Integrated policy and rating platform.

- **Insurity Claims Decisions Suite** — Claims workflow.

- **Insurity SpatialKey** — Geospatial analytics (hurricane and flood risk assessment).

Between 2022 and 2024, Insurity made a string of acquisitions of specialty companies like SpatialKey, Tropics, and ISCS, owning almost every layer of the specialty insurance core. In U.S. crop insurance, Insurity is virtually the de facto standard.

Insurity's cloud version is the **Insurity Cloud Platform**, a multi-tenant SaaS on AWS. As of 2026, 70%+ of new revenue is cloud.

Chapter 8 · EIS Suite — Core for Digital-Native Insurers

EIS Group (founded 1996, headquartered in San Francisco) is a later entrant growing fast. Its core message is "**a core for digital-native insurers**" — designed API-first, microservices, event-driven from day one.

EIS Suite covers both P&C and L&A as an integrated platform, with five core components.

- **PolicyCore** — Policy management.

- **ClaimCore** — Claims management.

- **BillingCore** — Billing and collection.

- **CustomerCore** — 360-degree customer view.

- **EngageCore** — Omnichannel customer interaction.

EIS's differentiator is the **event-driven architecture**. Every business event (policy bound, claim registered, premium received) publishes to a Kafka topic, and other modules subscribe and react. As a result, it's easy for an insurer to weave the core with its own analytics, CRM, and external systems.

{

"event_type": "policy.bound",

"event_id": "evt_2026_05_25_a7f9e",

"occurred_at": "2026-05-25T14:32:18.420Z",

"actor": {"user_id": "agent_445", "role": "agent"},

"tenant_id": "carrier_acme_insurance",

"payload": {

"policy_id": "POL-2026-009834",

"product": "AutoPersonal",

"policyholder_id": "cust_8723",

"effective_date": "2026-06-01",

"expiration_date": "2027-06-01",

"annual_premium": 1487.50,

"currency": "USD",

"coverages": [

{"code": "BI", "limit": 300000},

{"code": "PD", "limit": 100000},

{"code": "COMP", "deductible": 1000}

]

"schema_version": "2.3.0"

}

EIS fits well with digital indie insurers (insurtech), with more adoption cases from new entrants and digital sub-brands of incumbents than traditional large carriers. Allstate's North Light (a digital D2C brand) is a well-known EIS deployment.

Chapter 9 · Majesco — Cloud-Native Multi-Line

Majesco (spun off from Mastek Insurance in 2014, taken private by Thoma Bravo in 2020) is a mid-market core provider holding ground in both P&C and L&A. R&D bases in India and the U.S., 200+ insurer customers globally.

Majesco's main product family.

- **Majesco P&C Core Suite** — P&C policy, claims, billing.

- **Majesco L&A Core Suite** — Life & Annuity policy administration.

- **Majesco Distribution Management** — Agent and broker channel.

The differentiator is **deep Microsoft Azure integration**. Unlike Guidewire and Duck Creek which run on AWS, Majesco is deeply integrated with Azure and often used alongside Power BI and Dynamics 365. A fit for insurers with Microsoft-based IT infrastructure.

Price is more affordable than Guidewire or Duck Creek, making it a frequent pick for mid-market insurers (around `$1B`–`$5B` in revenue).

Chapter 10 · FINEOS Claims — The Claims Specialist's Seat

FINEOS (founded in Dublin, Ireland in 1993, ASX-listed in 2019) is a claims management specialist engine. Not a full-stack core, but specialized in **claims, workers' comp, and disability insurance**.

FINEOS's main products comprise three.

- **FINEOS Claims** — Group insurance and workers' comp claims.

- **FINEOS AdminSuite** — Policy administration (L&A support).

- **FINEOS Engage** — Customer self-service portal.

FINEOS's strength is depth in **long-term disability and workers' comp**. In this area, one incident can mean claims spanning 5 to 30 years, which a general P&C core can't handle. FINEOS handles this from the data model up.

The typical FINEOS Claims lifecycle starts with FNOL (First Notice of Loss) → Intake Validation → Claim Type branching (Disability, Workers Comp, Group Life). For a Disability claim, medical evidence collection follows; for Workers Comp, employer verification; for Group Life, beneficiary identification. Then adjudication (decision) → payment initiation plus periodic review on approval, or appeal process on denial. Long-term claims live in the system for 5 to 30 years via periodic review.

FINEOS's 2026 main customers are global group insurers — Unum, MetLife, Sun Life, Aflac. Market share in Korea and Japan is low, but FINEOS has established itself as one of the de facto standards in the global group insurance market.

Chapter 11 · Origami Risk — Risk Management and Self-Insurance

Origami Risk (founded in Chicago in 2009) is structurally distinct from traditional core systems. It's in the territory of **Risk Management Information Systems (RMIS)**, with main customers being large enterprises, governments, and hospitals that run self-insurance.

Origami's differentiator is the deep handling of **large self-insurers' workflows**. When a global enterprise runs its own insurance — handling employee workers' comp, property damage, and general liability internally instead of through external carriers — it needs a system to track, manage, and analyze those claims. Origami holds high market share in that space.

In 2024 and 2025, Origami acquired GRC companies like ClearRisk and Catalyst, extending from claims management into GRC (Governance, Risk, Compliance).

Chapter 12 · Britecore and Socotra — Next-Gen Cores for MGAs and Indie Insurers

Britecore (founded 2008, merged into Intellect Design Arena) and Socotra (Y Combinator 2014, acquired by EIS Group in 2024) are the new layer of next-gen P&C cores. Core message: "**cloud-native, API-first, launch an insurance product in 30 days**".

Britecore's strength is fast launch cycles for **MGAs (Managing General Agents) and small insurers**. Traditional cores take 6 to 12 months to launch a new product; Britecore promises 4 to 8 weeks.

Socotra is strong in emerging markets like Web2/Web3 indie insurance and embedded insurance (automatic insurance added at checkout). There are cases of digital D2C insurers like Lemonade launching on Socotra instead of building their own core.

| --- | --- | --- | --- | --- |

Chapter 13 · API-First Design and Guidewire APIs

The biggest change in insurance cores in the 2020s is **API-first design**. Where monolithic cores were once accessible only via their own UI, 2026 cores must expose all functionality via REST/GraphQL APIs. This is what enables digital channels, mobile, and embedded insurance.

Guidewire released **Cloud APIs** in earnest starting 2019, and as of 2026, 300+ endpoints are publicly available. Nearly every action in PolicyCenter, ClaimCenter, and BillingCenter is accessible via API.

POST /policy/v1/jobs/submission HTTP/1.1

Host: pc.acme.guidewire.cloud

Authorization: Bearer eyJhbGciOiJSUzI1NiI...

Content-Type: application/json

{

"data": {

"attributes": {

"productCode": "PersonalAuto",

"effectiveDate": "2026-06-01",

"primaryNamedInsured": {

"firstName": "Jane",

"lastName": "Smith",

"dateOfBirth": "1985-04-12",

"primaryAddress": {

"addressLine1": "123 Main St",

"city": "Boston",

"state": "MA",

"postalCode": "02101"

}

"personalVehicles": [

{

"vin": "1HGBH41JXMN109186",

"year": 2023,

"make": "Honda",

"model": "Accord"

}

]

}

A single API call creates a new quote in PolicyCenter, the rating engine calculates the premium, and the response returns. Before, this would have required opening the PolicyCenter UI and clicking 30 times. Now mobile apps, comparison sites, and a carrier's own D2C site can call this API and show an instant quote.

Another benefit of API-first design is **microservice decomposition**. Instead of replacing a monolithic PolicyCenter all at once, you can write new functionality as microservices on top of the API and gradually drain the core.

Chapter 14 · Salesforce Financial Services Cloud Integration

A newer current in the late 2020s is **Salesforce Financial Services Cloud (FSC) integration**. The number of insurers adopting Salesforce as a sales and customer management layer separate from the core has surged.

What Salesforce FSC covers.

- **Agent and broker CRM** — Sales pipeline, renewal campaigns.

- **Customer self-service** — Portal on Experience Cloud.

- **Claims collaboration** — Adjuster workspace on Service Cloud.

- **Analytics** — Based on Tableau CRM (formerly Einstein Analytics).

Guidewire, Duck Creek, and Sapiens all officially support bidirectional integration connectors with Salesforce FSC. Policy and claims data sync to Salesforce in real time, and when an agent starts a quote in Salesforce, it flows to PolicyCenter.

This setup splits insurance IT into two layers — **System of Record** (core like Guidewire) and **System of Engagement** (Salesforce). The core is the source of truth for data, and Salesforce owns user experience and external channels. How the two systems sync data has become the central question of late-2020s insurance IT architecture.

Chapter 15 · Mainframe Modernization — The Shadow of COBOL/AS-400

While the global P&C market shifts to cloud, many Korean and Japanese insurers still run in-house cores on COBOL/AS-400. Systems that have survived 30 years.

A typical mainframe insurance core stack.

- **Hardware** — IBM zSeries mainframe or IBM AS-400 (now IBM i).

- **Languages** — COBOL, RPG, PL/I.

- **Databases** — DB2, VSAM, IMS DB.

- **Transaction processing** — CICS, IMS TM.

- **Batch** — JCL (Job Control Language).

This stack's strength is **proven stability**. It has processed insurance policies 24/7 with no downtime for 30 years. The weakness is **personnel shortage**. The average COBOL developer has crossed 55 years old in the 2020s, and AS-400 RPG developers are even rarer. Korean and Japanese insurance IT departments perceive this as their biggest crisis.

IDENTIFICATION DIVISION.

PROGRAM-ID. CALC-PREMIUM.

*> 30-year-old auto premium calculation module — written in 1995

*> Simplified example (actual code is thousands of lines)

ENVIRONMENT DIVISION.

DATA DIVISION.

WORKING-STORAGE SECTION.

01 WS-DRIVER-AGE PIC 9(03).

01 WS-PRIOR-CLAIMS PIC 9(02).

01 WS-BASE-PREMIUM PIC 9(07)V99 VALUE 120000.00.

01 WS-AGE-FACTOR PIC 9(03)V999.

01 WS-CLAIMS-FACTOR PIC 9(03)V999.

01 WS-FINAL-PREMIUM PIC 9(09)V99.

PROCEDURE DIVISION.

MAIN-LOGIC.

IF WS-DRIVER-AGE < 25

MOVE 1.500 TO WS-AGE-FACTOR

ELSE IF WS-DRIVER-AGE < 65

MOVE 1.000 TO WS-AGE-FACTOR

ELSE

MOVE 1.200 TO WS-AGE-FACTOR

END-IF.

COMPUTE WS-CLAIMS-FACTOR = 1.000 + (WS-PRIOR-CLAIMS * 0.300).

COMPUTE WS-FINAL-PREMIUM = WS-BASE-PREMIUM * WS-AGE-FACTOR

* WS-CLAIMS-FACTOR.

DISPLAY "FINAL PREMIUM: " WS-FINAL-PREMIUM.

STOP RUN.

How to modernize this 30-year-old COBOL is the central agenda of Korean and Japanese insurance IT in 2026. Three options. (1) Keep the mainframe and put an API gateway on top. (2) Replace the entire core with a commercial package like Guidewire or Duck Creek (a `$100M`+ project, 3–5 years). (3) Use tools that automatically translate COBOL into Java/C# (Heirloom, TmaxSoft, Micro Focus).

Chapter 16 · Korea Samsung SDS Insurance IT — A 30-Year Officer Academy

Samsung SDS is a 30-year officer academy that has built the IT systems of Samsung Life, Samsung Fire & Marine, and Samsung Card. Almost every core architect of Korean insurance IT has passed through here once.

Samsung SDS's key insurance IT assets.

- **Brity insurance core** — In-house P&C and life integrated core. Deeply embedded in Samsung F&M and Samsung Life.

- **SCAS (Samsung Claims Administration System)** — In-house claims processing system.

- **Nexledger** — Blockchain-based insurance payout and reinsurance settlement.

- **Brity AI** — Chatbot, document processing, and payout automation AI solutions.

Since the mid-2020s, Samsung SDS has begun redesigning its in-house core as cloud-native. The core is **microservice decomposition and event-driven integration**, with an architecture assuming AWS/Azure multicloud.

Whether Samsung Life and Samsung F&M will continue using this in-house core, or replace it with a global package like Guidewire or Sapiens, is a major watch point in late-2020s Korean insurance IT. The strength of in-house is **the depth of Samsung group integration**; the weakness is **global scalability and personnel recruitment**.

Chapter 17 · KB Insurance Modernization — The KB Life CIS Cloud Case

KB Financial Group launched KB Life Insurance after merging KB Life and Prudential Life Korea in 2022, and the biggest IT challenge of this merger was **integrating the two companies' core systems**.

KB Life's next-generation core system **CIS (Core Insurance System)** was built between 2023 and 2025. Key features.

- **Microservice architecture** — Policy, claims, and billing split into independent services.

- **Kubernetes-based container orchestration** — On-premises plus KB Cloud hybrid.

- **Event-driven integration** — Asynchronous inter-module communication via Kafka.

- **API gateway** — Standard interface for external channel integration.

This project is cited as one of the most ambitious modernization cases in Korean insurance IT. A case of rebuilding both companies' mainframe cores as cloud-native while migrating 30 years of policy data.

Beyond KB, Shinhan Life, Kyobo Life, and Hyundai Marine are progressing similar modernization projects. Hyundai Marine follows the pattern of keeping its in-house core while gradually decomposing into cloud-native modules. Kyobo Life is known to have considered Sapiens adoption.

Chapter 18 · Japan NTT Data Insurance Core — The Massive Cloud Migration Wave

The Japanese insurance IT market is essentially split four ways among NTT Data, NRI (Nomura Research Institute), TIS, and NEC. Of these, NTT Data is the largest insurance IT SI provider.

NTT Data's insurance core assets.

- **NTT Data insurance core (in-house package)** — In-house developed core for Japanese P&C and life carriers.

- **OpenCanvas** — NTT Data's insurance API platform.

- **TopVision Insurance** — Next-generation insurance management package.

The biggest movement at Japanese insurers in the mid-2020s is **cloud migration**. Tokio Marine and Nichido started migrating some lines to Guidewire Cloud from 2023, and SOMPO announced a similar cloud transition around the same time. The main SI partners for this work are NTT Data and NRI.

A feature of Japanese insurance IT is **the much higher share of 30-year-plus mainframes** compared to Korea. Mainframes built in the 1980s and 1990s are alive into the 2020s, and the COBOL code on top runs into tens of millions of lines. How to modernize this is Japanese insurance IT's biggest challenge.

Chapter 19 · NRI ZAITAS — A 30-Year Insurance Core Package

NRI's (Nomura Research Institute) **ZAITAS** is one of the longest-running packages in the Japanese insurance core market. Launched in the early 1990s, it has survived 30 years and is deeply embedded in mid-sized and smaller Japanese insurers.

ZAITAS's strength is **deep alignment with Japanese insurance regulation and institutions**. It has been the package that responds fastest to regulatory changes by Japan's FSA (Financial Services Agency), Japanese tax law, and Japanese accounting standards (such as IFRS 17 adoption). The biggest friction for Japanese insurers adopting a global package like Guidewire or Sapiens, much like in Korea, is **local regulatory fit** — and ZAITAS has that baked in.

In the mid-2020s, NRI is rebuilding ZAITAS as cloud-native. Work that preserves ZAITAS's existing business logic while replacing the infrastructure layer with AWS/Azure multicloud plus Kubernetes. Targeted for completion by 2030.

Beyond NRI, another insurance IT heavyweight in Japan is **TIS**, with the **eX-Tide** package deeply embedded in the Japanese non-life market.

Chapter 20 · SOMPO and Tokio Marine Core Migrations — The Choice of Japan's Big 2

The big 2 of the Japanese non-life market are SOMPO Holdings and Tokio Marine Holdings. Both announced core system modernizations in the mid-2020s, and the difference in their choices shows the future of Japanese insurance IT.

**Tokio Marine and Nichido** — Announced Guidewire Cloud adoption in 2023. Phased migration starting with the auto line. SI partners: NTT Data and Accenture. 5-year timeline.

**SOMPO Japan Nipponkoa** — Announced an in-house next-generation core plus microservice decomposition in 2024. Built by their in-house IT subsidiary SOMPO Systems instead of an external package. Cloud-based on AWS Tokyo region.

What these two choices mean. Tokio Marine prioritized **global standard package consistency across the U.S., Asia, and Europe**, while SOMPO prioritized **Japan-specific fit and IT self-sufficiency**. Which is right will be evaluated around 2030.

These two migration projects are each more than `$500M` in IT investment. About 10x the size of Korea's KB Life CIS project. The scale and conservatism of Japanese insurance IT is captured in these numbers.

Chapter 21 · Low-Code/No-Code for Insurance Cores

Another major change in late-2020s insurance cores is **low-code/no-code adoption**. Guidewire Studio, Sapiens DigitalSuite, and Duck Creek Author all provide tools that let product managers, underwriters, and claims adjusters modify rules without code.

The driving forces.

- **Pressure on product launch speed** — Digital D2C insurers launch new products in 4 weeks while traditional insurers take 12 months. To close the gap, non-developers need to modify rules directly.

- **IT personnel shortage** — Gosu, RPG, and COBOL talent shortages are severe. Tools that let non-developers self-serve are necessary.

- **Speed of regulatory response** — Insurance regulations change frequently, and waiting for IT development cycles incurs delay penalties.

Guidewire Studio enables editing Gosu code in a visual IDE, and some rules are written via form-based GUIs. Duck Creek Author is itself a visual rule editor. Sapiens DigitalSuite composes channels and UI themselves as no-code.

The downside is clear. **Complex business logic is hard to express in low-code**. A hybrid model mixing code and GUI is still the reality.

Chapter 22 · Microservice Decomposition Patterns — Breaking Apart the Monolithic Core

Where do you start to decompose a 10-year-old monolithic PolicyCenter into microservices? The 2026 best practice is the **Strangler Fig pattern**.

- **Stage 1** — Put an API gateway in front of the existing monolith. All external calls pass through this gateway.

- **Stage 2** — Write new features as microservices. The gateway routes new feature requests to new services, and the rest to the existing monolith.

- **Stage 3** — Gradually migrate existing features to new services. Peel off one module at a time from the monolith and update the gateway's routing rules.

- **Stage 4** — Once all features have moved to new services, shut down the monolith.

The elegance of this pattern is that it **avoids a massive big-bang replacement**. Replacing a 10-year-plus insurance core all at once almost always fails (there's an IBM-cited statistic that "65% of mega IT projects fail"). Strangler Fig migrates progressively in 6 to 24 month increments, spreading risk.

The first candidates for microservice decomposition are typically **relatively independent domains** — FNOL (loss intake), payment collection, customer self-service. Core domains (policy issuance, rating) move last.

Chapter 23 · Data Migration — The Weight of 30 Years of Policy Data

The hardest part of replacing a core system is **data migration**. Moving 30 years of policy, claims, and payment data to a new system without losing a single record.

A typical migration workflow.

1. **Data profiling** — Analyze the existing system's data. Identify exact schema, data quality issues, gaps, and duplicates.

2. **Data cleansing** — Clean dirty data. E.g., standardize policy number format, standardize date format, fill missing fields.

3. **Mapping design** — Map existing schema to new system schema. Which fields transform how.

4. **Transformation script development** — ETL tools (Informatica, Talend, dbt) or custom scripts.

5. **Test migration** — Migrate some data (10–20%) to the test environment. Validate results.

6. **Full migration** — Migrate all data. Typically a weekend downtime or progressive cutover.

7. **Post-migration validation** — Confirm key indicators match — policy count, cumulative payout total, claim count.

Why this is hard. **The existing system's data isn't clean**. Formats changed multiple times over 30 years, some policies got dropped in each migration, and payment history sits in a separate mainframe system. Moving all this consistently typically eats 30–40% of the entire core migration budget.

-- Migration validation — a Korean insurance case

-- Compare policy counts and total premium between legacy mainframe and new Guidewire

WITH legacy_summary AS (

SELECT

line_of_business,

COUNT(*) AS policy_count,

SUM(annual_premium) AS total_premium,

COUNT(DISTINCT policyholder_id) AS unique_customers

FROM legacy_policies

WHERE effective_date >= '1995-01-01'

AND status IN ('active', 'lapsed', 'expired')

GROUP BY line_of_business

new_summary AS (

SELECT

product_code AS line_of_business,

COUNT(*) AS policy_count,

SUM(annual_premium) AS total_premium,

COUNT(DISTINCT primary_named_insured_id) AS unique_customers

FROM guidewire_pc.policies

WHERE migration_batch_id IS NOT NULL

GROUP BY product_code

)

SELECT

COALESCE(l.line_of_business, n.line_of_business) AS lob,

l.policy_count AS legacy_count,

n.policy_count AS new_count,

(n.policy_count - l.policy_count) AS count_diff,

l.total_premium AS legacy_premium,

n.total_premium AS new_premium,

ROUND((n.total_premium - l.total_premium) / l.total_premium * 100, 4) AS premium_diff_pct

FROM legacy_summary l

FULL OUTER JOIN new_summary n USING (line_of_business)

ORDER BY ABS(count_diff) DESC;

A validation like this query has to run dozens of times before and after migration until every number matches within 0.01% before cutover finishes.

Chapter 24 · Disaster Recovery and BCP — A System That Can't Stop for 24 Hours

The insurance core is almost an absolute 24/7 zero-downtime requirement. Claims come in at 3 AM, payments process around the clock, and car accidents don't pick the hour.

Traditional insurance BCP (Business Continuity Plan) design.

- **Redundant data centers** — Main plus DR site. Synchronous and asynchronous replication.

- **RPO (Recovery Point Objective)** — Allowable data loss time. Typically 0 to 5 minutes.

- **RTO (Recovery Time Objective)** — Recovery time. Typically 1 to 4 hours.

- **Regular DR drills** — Quarterly main-to-DR failover tests.

In the cloud era, this simplifies. AWS and Azure multi-AZ and multi-region are built in, and RPO/RTO shrinks to minutes. The Guidewire Cloud Platform is multi-AZ on AWS by default, with automatic backup and automatic failover.

In return, the biggest worry of cloud migration is **the cloud's own outages**. There was an incident in 2024 where AWS us-east-1 went down and several insurance cores on the U.S. East Coast stopped at once. After that, insurers started seriously considering multi-cloud (AWS plus Azure) or multi-region distribution.

Chapter 25 · The Future of Insurance Cores 2026–2030 — Five Predictions

Finally, five predictions for 2030.

1. **Guidewire Cloud takes 80%+** — Guidewire's 2030 target of 80% cloud is likely to be reached. 90%+ of new P&C builds split between Guidewire Cloud and Duck Creek OnDemand.

2. **AI agents reshape claims handling** — Claude/GPT-based claims adjuster automation handles 80% of simple claims. Humans handle complex ones.

3. **Embedded insurance explodes** — Auto insurance automatically added at car purchase, flight booking, appliance buying. API-first cores like Britecore and Socotra make it possible.

4. **The last 10 years of the mainframe** — Korean and Japanese mainframe insurance cores retire en masse around 2030. As COBOL talent enters their late 60s, maintenance becomes impossible.

5. **Reinsurance, blockchain, climate insurance** — Blockchain-based reinsurance settlement (Nexledger, B3i), new climate-response products (parametric insurance), explosive growth in cyber insurance.

> "An insurance core must live for 30 years. But a 30-year-old core cannot live 30 more years. The modernization in between is all of late-2020s insurance IT."

Chapter 26 · Epilogue — After the COBOL Developer Left

Back to the opening scene. The chatroom at 4:17 AM. The COBOL developer who left in 2008. Comments in Japanese.

In 2026, every major Korean and Japanese insurer's IT department has one of these systems. A core that survived 30 years. The people who wrote it are gone. New hires don't want to learn COBOL. And that system processes tens of billions of won in payouts every day.

This is the true weight of insurance core modernization. Not a simple IT project, but a question of **the entire insurer's corporate continuity**. One mistake can topple a company. That's why insurance is the most conservative IT industry, and at the same time, in the most urgent need of modernization.

Whether Guidewire takes 80% of the P&C market, whether Duck Creek catches up via SaaS, whether Sapiens dominates the global multinational market, whether Korea's Samsung SDS and Japan's NTT Data hold the line with in-house cores — all of these questions will get their answer by 2030.

One last word. **Insurance is a promise. IT must keep the promise made 30 years from now.** That's the reason for the insurance core system's existence.

— Insurance Core Systems 2026, end.

References

- [Guidewire Software — InsuranceSuite](https://www.guidewire.com/products/insurancesuite)

- [Guidewire Cloud Platform](https://www.guidewire.com/products/cloud)

- [Duck Creek Technologies — OnDemand](https://www.duckcreek.com/)

- [Duck Creek Policy](https://www.duckcreek.com/product/policy/)

- [Sapiens — Insurance Software Solutions](https://www.sapiens.com/)

- [Sapiens IDIT P&C Suite](https://www.sapiens.com/products/sapiens-idit/)

- [Sapiens CoreSuite for L&P](https://www.sapiens.com/products/sapiens-coresuite-for-life-pensions/)

- [Insurity — Insurance Software](https://www.insurity.com/)

- [EIS Group — EIS Suite](https://www.eisgroup.com/)

- [Majesco — Insurance Software](https://www.majesco.com/)

- [FINEOS — Claims Software](https://www.fineos.com/)

- [Origami Risk — RMIS Platform](https://www.origamirisk.com/)

- [Britecore — Insurance Core Platform](https://www.britecore.com/)

- [Socotra — Modern Core for Insurance](https://www.socotra.com/)

- [Salesforce Financial Services Cloud](https://www.salesforce.com/products/financial-services-cloud/)

- [IBM zSystems — Mainframe](https://www.ibm.com/products/z)

- [IBM Power Systems / IBM i (AS/400 successor)](https://www.ibm.com/products/power)

- [Samsung SDS — Brity Insurance](https://www.samsungsds.com/)

- [NRI — Nomura Research Institute](https://www.nri.com/)

- [NTT Data — Insurance Solutions](https://www.nttdata.com/global/en/services/industries/insurance)

- [Saga pattern — Hector Garcia-Molina, 1987](https://www.cs.cornell.edu/andru/cs711/2002fa/reading/sagas.pdf)