🇯🇵🇪🇺 I analyze how GenAI and digital finance are transforming the insurance sector in Japan and EU.

🔥 Current focus
• GenAI adoption in EU and Japanese insurers
• AI governance and model-risk management
• Digital finance and cross-border regulatory differences
• AI Act and DORA implications
• Operational resilience and risk themes
• Emerging use cases in underwriting, claims, sales

🌍 Background
• 14+ years at Nomura Research Institute (NRI)
• Secondment at Nippon Life X
• Overseas assignment at Tokio Marine (Shanghai)
• Now based in Frankfurt to follow EU insurers and GenAI trends
• Speaker / researcher on GenAI and insurance

✏️ I publish analytical pieces and deliver expert talks on GenAI, digital finance, and insurance innovation from a JP–EU perspective.

🇬🇧 What the ITC Europe 2026 themes revealed about where insurance is heading

With ITC Europe now closed, the 2026 themes reveal a shift that is no longer subtle.

Until recently, industry focus centred on capability 🚀
What AI could automate.
How fast processes could move.
Which technologies unlocked efficiency.

In 2026, that framing clearly changed.

⸻

🔹 Tech & AI
Before: AI as experimentation and tooling 🤖
Now: AI as operating model change.
Legacy systems, siloed data, and cultural inertia were no longer peripheral issues. They surfaced as the primary execution constraint.

🔹 Risk & Coverage
Before: Climate, cyber, and emerging risks discussed in isolation
Now: Risk discussed as an interconnected portfolio 🌍
AI liability, geopolitics, cyber, and climate appeared increasingly bound together by a single question: underwriting responsibility.

🔹 Growth & Investment
Before: Independent insurtech scaling and funding narratives
Now: Growth framed around integration and collaboration 🤝
Working with incumbents was no longer a strategic option. It emerged as a structural condition.

🔹 Distribution & Business Models
Before: Customer experience and digital channels
Now: How risk actually flows 📊
Embedded insurance, delegated authority, and data led targeting reshaped the conversation from channels to control.

⸻

💡 Takeaway

What became visible through the 2026 themes was not faster innovation.

The shift is not about technology.
It’s about who controls risk flow.


#ITCEurope
#Insurance
#GenerativeAI

Hiroko Washiyama

Insurance, GenAI & Digital Finance | JP–EU Research | Ex-NRI (14+ yrs)

Insurance, GenAI & Digital Finance | JP–EU Research | Ex–Nomura Research Institute (14+ yrs)

💫2026: How GenAI’s Role Changes in Insurance

Simple efficiency gains are done.
The basic productivity phase of GenAI is largely complete.

What comes next is not smarter AI.
It is exposed operations.

⸻

🔍 What GenAI really surfaces in 2026

As GenAI moves beyond basic automation, it begins to reveal operational gaps insurers have lived with for years:
　⚠️  processes that work only because people fill in the blanks
　⚠️  decisions based on tacit understanding rather than explicit rules
　⚠️  operations that can be explained, but not consistently repeated

This is not a technology issue.
It is an operational reality check.

⸻

⚙️ Why this matters now

In 2026, GenAI removes the human buffer that used to absorb ambiguity.

What was once:
	•	“handled case by case”
	•	“managed by experienced staff”
	•	“good enough in practice”

becomes visible, inconsistent, and hard to scale.

⸻

🎯 Executive takeaway

GenAI does not change insurance operations.
It reveals which operations were never fully defined in the first place.

GenAI is no longer an efficiency tool.
It is an operational stress test.


#GenAI
#Insurance

graphical user interface, application, Word

🇪🇺 EU Inc and its real impact on insurance

This is not about regulation.
It is about the unit of design.

⸻

1. What is EU Inc (only the essentials)

EU Inc is a proposal that allows companies to operate across the EU based on a single legal entity.

This is not about simplifying legal paperwork.

The core idea is straightforward:
👉 capital, governance, and business scale can be designed at EU level.

This directly reflects the concern repeatedly raised by the European Commission:
as long as European companies remain fragmented by national structures,
they cannot compete with US firms on scale.

⸻

2. Why this matters for insurance

Three layers of impact


① Legal and capital layer 🏦

EU Inc challenges the long-standing insurance group model of holding company plus country-level subsidiaries.

Key questions resurface:
• Where should capital actually sit
• At what level should risk be pooled
• How far can internal models be standardised across the EU

All of these are design questions under Solvency II,
where supervisory interpretation matters.

In practice, this puts EIOPA squarely at the centre.

⸻

② Operations and AI layer 🤖

Today, many European insurers still design:
• IT platforms
• data governance
• AI usage rules

on a country-by-country basis.

If EU Inc becomes the operating assumption,
AI can no longer be a collection of local PoCs.

It must be designed as an EU-wide, governed operational layer.

This is not an IT integration issue.
It is a question of AI governance and operating-model design.

⸻

③ Competition layer 🚀

This is where EU Inc becomes truly disruptive.

With:
• one legal entity
• one product architecture
• one AI foundation

it becomes possible to target the entire EU market from day one.

This environment tends to favour players with high design flexibility,
often more than insurers that expanded country by country.

⸻

Takeaway 🎯

EU Inc is not deregulation.

It is a proposal that raises the unit of insurance business design
from national level to EU level.

The real questions it forces are simple:
• At what level do you design capital and risk
• At what level do you govern AI and operations
• At what level do you compete

EU Inc pushes insurers to answer these questions at EU scale.


#EUInc #Insurance

graphical user interface, text, application

🇩🇪Why Allianz’s partnership with Anthropic matters 🔍


In January 2026, Allianz announced a global partnership with Anthropic.

This is not an exclusive AI strategy, nor simply one model among many.
One fact stands out 👇

Claude becomes part of Allianz’s internal AI platform, available to all 156,000 employees globally.

This marks a shift from experimental use toward
standardized, enterprise-wide adoption of GenAI
under strong governance, safety, and explainability requirements 🛡️

From that perspective, three points are worth noting.

⸻

1️⃣ From pilots to standardized enterprise deployment

The partnership is framed around production use, not isolated PoCs.

Making Claude broadly available across the organization signals that GenAI is being treated as a
standard productivity and decision-support tool,
rather than a niche capability for select teams.

This alone sets the announcement apart from many prior AI partnerships.

⸻

2️⃣ Governance and explainability as prerequisites

Anthropic’s public positioning emphasizes safety by design, explainability, and controllability.

These characteristics align with the requirements faced by large, regulated insurers operating under supervisory scrutiny 🏛️

The partnership reinforces a clear message:
model capability without governance readiness is insufficient for enterprise-scale deployment.

⸻

3️⃣ Attention shifts toward agentic workflows ⚙️

Beyond chat-based use, the partnership is widely discussed in the context of
agentic workflows — AI systems that support multi-step processes rather than single interactions.

While concrete use cases are still evolving, this framing reflects a broader 2026 trend:
GenAI is increasingly evaluated by how it fits into end-to-end operational workflows,
not just how well it responds to prompts.

⸻

Takeaway 🎯

This partnership does not represent a sudden technological breakthrough.

It reflects a maturation in how insurers approach GenAI:
from experimentation,
to standardization,
and toward workflow-level integration under governance constraints.


#Allianz
#Anthropic
#Insurance
#GenAI

🇨🇭🌍 Davos AI narratives, seen from insurance

AI at Davos is largely discussed in terms of speed, autonomy, and scale 🚀
Self-improving systems, agentic workflows, and organizational redesign dominate the conversation.

From an insurance perspective, this matters —
but not in the way it is often interpreted.

Davos is not where AI meets operational reality.
It is where direction is negotiated before constraints appear.

Insurance operates under a different logic.

AI adoption in insurance is shaped less by technical capability
and more by structural responsibility 🧩:
	•	decision ownership
	•	auditability
	•	liability allocation
	•	regulatory accountability

Because of this, the relevant question is rarely
“what can AI do?”
but rather
“where is AI allowed to sit inside the decision architecture?” 📐

This is why, in practice:
	•	autonomy is layered gradually
	•	human oversight remains explicit 👤
	•	deployment speed differs sharply by function
(pricing, underwriting, claims)

The risk for insurers is not ignoring Davos narratives —
it is importing them without translation.

Davos helps clarify direction 🧭
Insurance must define conditions.

AI strategy becomes actionable only when ambition is filtered
through accountability, regulation, and liability 
—
the realities that Davos discussions can afford to abstract away.


#Insurance #GenAI #WEF #Davos

🇯🇵SOMPO’s AI Design: Change the Apps, Fix the Foundation💫

Choosing not to mix Generative AI and Decision Architecture

Many insurers have adopted generative AI 🤖
Yet in reality, very few have changed how decisions are actually made.

What SOMPO shows is not about what AI to use,
but about deciding where AI should sit inside the organisation 📍.

⸻

The conclusion is simple ✅
	•	⚙️ Generative AI is primarily built in house
Flexible. Replaceable.
	•	🏗️ Decision and data architecture is centralized with a single provider
Fixed. Governed.

These two layers are intentionally kept separate.

⸻

Generative AI: in-house by design 🛠️

Generative AI is used as a practical business tool:
	•	internal Q&A
	•	document drafting
	•	daily operational support

External partners may be used selectively,
but ownership remains internal, and models or vendors can be swapped 🔄.

👉 Generative AI is treated as a convenient tool, nothing more.

⸻

The foundation: single-vendor concentration 🏛️

By contrast, the core of decision-making and data integration
is centralized on Palantir.

The focus here is:
	•	consistency of decisions
	•	traceability
	•	accountability

This layer exists to preserve how decisions were made,
and it is not designed to change easily 🔒.

⸻

🇪🇺 Takeaway for Europe

🟢 Use generative AI flexibly.
🔴 But do not let it lead your decision and governance structures.

Before choosing models or vendors, decide this first:
what can change — and what must remain fixed.


#SOMPO
#Insurance
#AI

graphical user interface, text, application, Teams

💫What CES Clearly Teaches Insurers About Where to Start with GenAI


CES makes one thing clear for insurers:
👉 it highlights where to start, not where to showcase.

CES is not just a tech exhibition.
For insurers, it functions as a reality filter.

Based on an analysis of GenAI-related exhibits and keynote themes at CES, two conclusions consistently emerge.

⸻

🔹 1. AI performance is not the primary differentiator

👉 Preconditions matter more.

GenAI outcomes depend less on model capability and more on how clearly operations are defined.

Across CES, GenAI solutions that gain attention typically operate in environments where decision scope, input structure, and success criteria are already explicit.
This clarity makes live, repeatable demonstrations possible.

📍 Insurance takeaway
GenAI struggles not because models are weak,
but because decision boundaries, accountability, and responsibility splits are often left implicit.

This is why scaling PoCs without fixing operational definitions rarely leads to production impact.

⸻

🔹 2. Agentic AI does not imply full automation

At CES, Agentic AI is presented as controlled delegation, not end-to-end autonomy.

What is emphasized instead is a consistent design pattern:
	•	narrowly scoped objectives
	•	built-in human escalation
	•	clearly defined recovery paths

📍 Insurance takeaway
The realistic path forward is to:
	•	decompose judgment
	•	delegate only bounded decisions
	•	design human re-entry points upfront

This approach aligns naturally with EU regulatory expectations,
because control and accountability are designed in, not added later.

⸻

🎯 Executive takeaway

CES does not reward smarter AI.
It rewards better-defined decisions.

These conclusions are not opinions.
They are structural lessons drawn from what CES makes possible on stage.


#Insurance #GenAI #CES

graphical user interface, text, application, email

💫Why ICE Barcelona also matters for insurance and financial institutions 🤔


An event about AI, but really about operating assumptions

ICE Barcelona is known as a gaming industry event.
But its real value lies elsewhere.

The event consistently shows how AI is actually operated in highly regulated environments.

What stands out is not model sophistication.
It is how operational constraints are defined before AI goes live.

⸻

1. Fraud and AML solution demos 👀

Across fraud and AML vendor booths, the explanation order is strikingly consistent.
	•	🤖 AI scores all transactions in real time
	•	📥 Only the top X percent are routed to a human review queue
	•	⏱️ Queue size is designed around a fixed assumption such as cases per reviewer per day

The first fixed variable is human processing capacity.
Model accuracy comes later.

⸻

2. Responsible Gaming sessions and regulatory panels ⚖️

These sessions are not just about addiction prevention.

What insurance and financial professionals observe is how regulators frame their questions.

The discussion consistently focuses on:
	•	🚨 Which signals trigger human intervention
	•	👤 Where decision authority sits within the organization
	•	📊 Whether intervention volumes align with operational capacity

The core issue is simple.
Which decisions are automated, and where human responsibility begins.

What matters is not vision, but
🔁 production ready operational flows.

⸻

In short:
ICE Barcelona uses AI as the lens, but exposes the operating assumptions management must decide.


#ICEBarcelona
#Insurance
#GenAI

taplio

Hiroko Washiyama's Linkedin Analytics