The AI Imperative in Insurance

Why C-Suite Alignment is the Foundation
for Transformation

Part 1 of 6: Rewiring Insurance for the Age of AI

Insurance companies that have successfully embedded AI across their operations are generating 6.1 times the Total Shareholder Return of their peers over the past five years. To put this in perspective, in most industries undergoing digital transformation, the gap between leaders and laggards typically ranges from 2 to 3 times. The insurance sector's 6.1x differential suggests we're witnessing more than competitive advantage. We're seeing the emergence of a structural shift in how value is created and captured.

Understanding Pilot Purgatory

Many insurance organizations today find themselves in what I've come to recognize as "pilot purgatory." They've launched AI chatbots, deployed fraud detection models, and automated portions of their workflows. These initiatives often show promising results in isolation:

• Customer satisfaction improvements of 0.5%

• Loss reductions of 0.1%

• Efficiency gains of 0.2%

While these achievements represent genuine progress, they often don't translate into material business impact. Through my work with various organizations, I've observed five common patterns that keep companies in this state:

1. Lack of Enterprise-Wide Strategic Alignment

Many AI initiatives begin as technology projects rather than business transformations. A data science team identifies an opportunity, builds a proof of concept, demonstrates promising results, and receives approval to deploy. The initiative is measured on technical metrics such as model accuracy or processing speed rather than business outcomes like cost reduction or revenue growth.

In my experience leading transformation programmes, I've learned that without clear financial targets and executive accountability, AI projects struggle to compete for resources and attention. The business value needs to be clearly articulated from the outset.

2. Underestimation of Required Investment

Organizations frequently underestimate the full scope of investment needed for successful AI transformation. Budgets often account for data scientists and cloud infrastructure but may overlook the broader ecosystem: data engineering for pipelines and quality assurance, MLOps capabilities for deployment and monitoring, change management for user adoption, staff retraining, legacy system integration, and ongoing model maintenance.

When only the visible components are funded, even talented teams can find themselves constrained by inadequate infrastructure, data quality challenges, or insufficient organizational support.

3. Use Case Fragmentation Rather Than Domain Integration

A common pattern involves launching separate initiatives for chatbots, fraud detection, claims automation, and underwriting support, each managed by different teams with different vendors and technology stacks.

While none of these initiatives are failures individually, their isolated nature often limits their collective impact. A chatbot improving satisfaction by 0.5% or a fraud model reducing losses by 0.1% may not materially affect overall business performance.

The alternative approach focuses on domain-based transformation, where AI is deployed comprehensively across an entire functional area such as claims or underwriting. This integrated approach can improve domain-level profitability by 7 to 12 percent, creating meaningful business impact.

4. Limited Component Reusability

Organizations sometimes build bespoke AI solutions for each use case with minimal reuse. Different teams may independently develop similar capabilities for natural language processing, document processing, or data extraction.

In my work on AI evaluation frameworks, one key insight has been that building horizontal components once and reusing them across applications can significantly accelerate development timelines and enable organizations to pursue multiple AI opportunities in parallel.

5. Over-Reliance on Vendor Solutions

While vendor partnerships play an important role, over-reliance on external solutions can create limitations. Vendor products are designed to serve broad markets and incorporate general best practices, but they may not capture the specific institutional knowledge and processes that differentiate one insurer from another.

When multiple competitors deploy identical vendor solutions, performance tends to converge toward market median. True competitive advantage often comes from how organizations orchestrate and customize AI capabilities to their unique context.

Why AI Requires Organizational Rewiring

Traditional technology implementations typically follow a pattern of identifying a process, selecting appropriate technology, configuring the system, training users, and deploying. The underlying business process often remains largely unchanged.

AI transformation requires a different approach because AI's value proposition is fundamentally different. Rather than simply accelerating existing processes, AI enables entirely new ways of working. An AI-powered underwriting system, for example, can evaluate risk factors that are difficult for humans to process manually, identify patterns across millions of data points, and make real-time decisions that previously required days of analysis.

Comprehensive AI transformation typically requires evolution across three interconnected dimensions:

Workflow Redesign

Major business processes benefit from being reimagined with AI capabilities as a core assumption. In claims processing, for instance, traditional workflows were designed around sequential human review. An AI-enabled workflow might involve intelligent triage that routes straightforward claims to automated processing, complex claims to specialized adjusters with AI-generated insights, and potentially fraudulent claims to investigation teams with supporting evidence already compiled.

Operating Model Evolution

Organizational structures, roles, and ways of working often need to evolve to support AI-driven operations. This can include creating new roles such as AI product managers and MLOps engineers, establishing governance processes for model deployment and monitoring, and developing performance metrics that reflect AI-enabled capabilities.

Traditional organizational structures can sometimes create barriers to AI adoption. Siloed business units may limit necessary data sharing, while hierarchical approval processes can slow the rapid iteration that AI development often requires.

Technology Stack Modernization

Legacy infrastructure may not be optimized for AI workloads. Successful deployment typically requires modern data platforms capable of large-scale processing, cloud infrastructure providing elastic compute capacity, real-time data pipelines enabling immediate insights, and MLOps capabilities allowing rapid model deployment and monitoring.

As I've written about in my work on data migration programmes, attempting to deploy AI on legacy foundations can significantly limit what's achievable. Infrastructure modernization often needs to be part of the transformation journey.

The Six Signature Moves Framework

Based on research into successful transformations and practical experience, I've identified six signature moves that tend to distinguish AI leaders from those still working toward breakthrough results. These aren't sequential steps but rather interconnected capabilities that organizations develop in parallel:

1. Align the C-suite around a business-led roadmap focused on domain transformation

2. Build the right talent bench with 70 to 80 percent in-house expertise

3. Adopt a scalable operating model that enables distributed innovation

4. Use technology for speed through reusable AI components

5. Embed data everywhere as a foundational capability

6. Invest in adoption and change management at parity with technology development

This series will explore each move in depth. The first move, C-suite alignment, merits particular attention as it provides the foundation for the others.

Achieving C-Suite Alignment: Four Critical Dimensions

C-suite alignment extends beyond agreement that AI is important. It involves achieving consensus on four critical dimensions:

1. Strategic Priorities: From Use Cases to Domain Transformation

The first alignment challenge involves shifting from evaluating individual use cases to considering domain transformation opportunities.

In a use case approach, AI opportunities are evaluated individually based on standalone ROI. Each initiative is approved or rejected based on its individual business case.

While this seems logical, it can lead to fragmentation, may undervalue initiatives that create platform capabilities, and often results in a portfolio of incremental improvements that don't accumulate to material impact.

A domain transformation approach involves deploying AI comprehensively across an entire functional area with the goal of fundamentally improving domain-level performance. Rather than asking about the ROI of a specific chatbot, leadership considers how to transform an entire operation through comprehensive AI deployment.

Case Example: Aviva's Claims Transformation

Aviva provides an instructive example. Rather than deploying AI selectively for specific use cases, they made a strategic decision to transform their entire claims domain. They deployed over 80 AI models across claims operations, touching virtually every aspect of the claims journey.

The results included savings of over £60 million ($82 million) in 2024, a 23-day reduction in liability assessment time, and a 65 percent decrease in customer complaints. These outcomes represent fundamental business performance improvements that materially affect profitability and competitive position.

2. Investment Levels: Understanding the Full Scope

The second dimension involves understanding and committing to the full scope of investment required. This extends beyond data scientists and cloud infrastructure.

Comprehensive transformation typically requires investment across four integrated layers:

Infrastructure Layer: Cloud platforms, security and governance frameworks, MLOps capabilities, and monitoring systems. For a mid-sized insurer, this typically represents $3 to 5 million in the first year, with ongoing annual costs of $2 to 3 million.

Data Platform Layer: Unified data architecture, real-time processing capabilities, automated data quality systems, and comprehensive governance. Investment typically ranges from $5 to 8 million initially, with annual operating costs of $3 to 4 million.

AI and ML Capabilities Layer: Foundation models, custom models trained on proprietary data, agent orchestration systems, and retrieval-augmented generation systems. This varies based on scope but typically ranges from $4 to 10 million annually for a comprehensive programme.

Engagement Layer: Multimodal conversational experiences, omnichannel consistency, personalization engines, and feedback loops. This typically represents $2 to 4 million initially with ongoing enhancement costs.

For a mid-sized insurer, total investment for comprehensive transformation typically ranges from $15 to 25 million in the first year, with ongoing annual investments of $10 to 15 million.

The business case becomes compelling when evaluated at the domain level. A claims transformation that improves profitability by 10 percent in a $500 million claims operation generates $50 million in annual value, significantly exceeding the investment when AI is deployed at scale.

3. Success Metrics: Focusing on Business Outcomes

The third dimension involves agreement on how success will be measured. The distinction between technology metrics and business outcomes is important.

Technology metrics such as model accuracy, processing speed, and system uptime are valuable for operational management but may be insufficient for evaluating business impact.

Business outcome metrics such as claims processing cost reduction, premium growth rates, loss ratio improvement, and customer retention directly tie AI initiatives to financial performance.

Establishing clear baselines, setting specific targets, and creating accountability for achieving them helps ensure AI investments deliver expected returns.

For a claims transformation, appropriate metrics might include:

• Reduce average claims processing cost by 15 percent within 18 months

• Decrease claims cycle time by 25 percent within 12 months

• Improve customer satisfaction scores by 20 points within 24 months

• Reduce claims leakage by 5 percent within 18 months

4. Governance Structures: Enabling Sustained Execution

The fourth dimension involves establishing governance structures that enable sustained execution.

The Ownership Question: Successful AI transformations typically involve collaboration across multiple functions. The CIO controls technology resources, the Chief Data Officer manages data assets, the Chief Analytics Officer leads data science teams, and business unit leaders are accountable for business outcomes.

Many successful transformations establish a dedicated AI transformation office that reports to the CEO and includes representation from all key stakeholders.

The AI Transformation Office typically serves several functions:

• Strategic Planning: Developing the enterprise AI roadmap, prioritizing domain transformations, and allocating resources

• Standards and Architecture: Establishing technical standards, defining reusable components, and ensuring architectural coherence

• Talent Development: Building AI capabilities through hiring, training, and knowledge sharing

• Vendor Management: Evaluating and managing relationships with technology vendors

• Performance Tracking: Monitoring business outcomes, identifying issues, and ensuring accountability

The transformation office enables domain teams to build solutions effectively by providing strategy, standards, platforms, and support, while accountability for business outcomes rests with domain leaders.

The Build Versus Buy Strategic Decision

One consequential decision requiring C-suite alignment is the organization's approach to building AI capabilities in-house versus purchasing vendor solutions.

Building Core Capabilities

For capabilities central to competitive differentiation, in-house development often makes strategic sense:

Underwriting Intelligence: Risk models and pricing algorithms represent core insurance value creation. These capabilities incorporate institutional knowledge, proprietary data, and market-specific insights.

Claims Adjudication: The judgment that distinguishes excellent claims organizations requires deep institutional knowledge about when to settle versus litigate and how to balance customer satisfaction with cost control.

Customer Intelligence: Understanding customer behavior, preferences, and lifetime value is closely tied to specific customer bases, distribution channels, and product portfolios.

Distribution Optimization: How organizations identify prospects, engage them, and convert them reflects unique go-to-market strategies.

Building these capabilities requires significant investment but can create sustainable competitive advantage.

Purchasing Commoditized Capabilities

For capabilities that are important but not differentiating, vendor solutions can offer speed and proven functionality:

Finance and Accounting: AI-powered financial processes are important for efficiency but typically don't differentiate insurers in the market.

Human Resources: AI applications in HR benefit operations but aren't core to insurance business models.

IT Operations: Infrastructure monitoring and security are critical but generally commoditized across industries.

Standard Data Services: Weather data, geospatial intelligence, and credit scores are best purchased from specialized providers.

The Orchestration Capability

Even when purchasing external solutions, organizations benefit from developing internal orchestration capabilities for integrating disparate systems, customizing vendor platforms, managing vendor performance, and maintaining strategic flexibility.

This orchestration capability can itself become a competitive advantage.

The Path Forward: From Alignment to Action

Achieving C-suite alignment is an ongoing process rather than a one-time event. As AI capabilities evolve and competitive dynamics shift, strategic roadmaps benefit from continuous refinement.

The First 90 Days

For organizations beginning their transformation journey, the first 90 days might focus on:

Weeks 1 to 4: Assessment and Baseline

• Comprehensive assessment of current AI initiatives and capabilities

• Establishing baseline metrics for priority domains

• Identifying opportunities to demonstrate value and build momentum

• Assessing organizational readiness and identifying capability gaps

Weeks 5 to 8: Strategy Development

• Developing domain transformation roadmaps for priority areas

• Defining success metrics and accountability structures

• Establishing governance framework and transformation office

• Creating business cases and securing budget commitments

Weeks 9 to 12: Foundation Building

• Recruiting key leadership for transformation office and domain teams

• Initiating infrastructure and data platform investments

• Launching first domain transformation with clear milestones

• Establishing communication and change management processes

First Year Milestones

By year end, successful transformations often achieve:

• Organizational Alignment: C-suite consensus on strategy, investment, and metrics with established governance

• Platform Foundation: Core infrastructure and data platform capabilities operational

• First Domain Success: At least one domain transformation showing measurable business impact

• Talent and Capability: Core team in place with appropriate mix of expertise

• Momentum and Learning: Visible progress creating organizational confidence

Sustaining Transformation

AI transformation is a multi-year journey that benefits from:

• Continuous Communication: Regular updates on progress, challenges, and achievements

• Iterative Refinement: Learning from early initiatives to inform continuous improvement

• Expanding Scope: Building on initial successes to expand to additional domains

• Capability Building: Ongoing investment in talent development and knowledge sharing

• Cultural Evolution: Gradual shift toward embedding AI as core to operations

Conclusion

The insurance industry is experiencing a significant shift driven by AI capabilities. The performance gap between organizations that have successfully embedded AI and those still in early stages continues to widen as AI capabilities compound and customer expectations evolve.

Organizations face an important strategic choice: commit to comprehensive transformation or risk competitive disadvantage. C-suite alignment on strategic priorities, investment levels, success metrics, and governance structures provides the foundation for pursuing domain transformations that can deliver material business impact.

In the next article in this series, I'll explore the second signature move: building the right talent bench with 70 to 80 percent in-house expertise. Once the C-suite is aligned on strategic direction, having the right team becomes essential for execution.

Next in Series: Part 2 - Building Your AI Dream Team: The 70-80% In-House Rule