RIMAC: How a Peruvian Insurance Company is Scaling AI Custom Case Solution & Analysis

Evidence Brief

1. Financial Metrics

• Market Position: RIMAC holds approximately 30 percent market share in the Peruvian insurance industry (Paragraph 2).
• Market Context: Insurance penetration in Peru remains low at roughly 2 percent of Gross Domestic Product, compared to higher averages in neighboring countries (Paragraph 5).
• Investment Scale: Significant capital allocated to the AI Factory and cloud migration initiatives since 2018 (Exhibit 1).
• Efficiency Targets: The CEO targets a reduction in the combined ratio through automated claims processing and improved fraud detection (Paragraph 14).

2. Operational Facts

• Organizational Structure: Established a centralized AI Factory in 2021 to consolidate data science and engineering talent (Paragraph 8).
• Human Capital: The data organization grew to over 100 specialized professionals including data scientists, engineers, and translators (Paragraph 12).
• Technology Stack: Transitioned from on-premise legacy systems to a multi-cloud environment using Amazon Web Services and Microsoft Azure (Paragraph 10).
• Process Maturity: Developed a standardized pipeline for AI model development, moving from ad-hoc experiments to a structured factory approach (Exhibit 3).

3. Stakeholder Positions

• Fernando Rios (CEO): Views AI as the primary driver for long-term competitiveness and operational efficiency in a low-penetration market (Paragraph 4).
• Leandro Rosi (CTO/CDO): Advocates for technical excellence and the importance of data governance as a prerequisite for scaling (Paragraph 9).
• Business Unit Leaders: Display varying levels of adoption; some view the AI Factory as a distant service provider rather than a core partner (Paragraph 15).
• IT Department: Focused on managing technical debt while supporting the rapid deployment of new data tools (Paragraph 11).

4. Information Gaps

• Specific ROI: The case does not provide the exact Return on Investment for the first ten AI models deployed.
• Talent Retention: Data on the turnover rate of specialized AI talent within the Peruvian market is absent.
• Competitor Benchmarking: Specific AI maturity levels of direct competitors like Pacifico Seguros are not detailed.

Strategic Analysis

1. Core Strategic Question

• How can RIMAC transition from a centralized AI Factory model to an integrated enterprise-wide capability without losing technical rigor or business relevance?
• How to overcome the cultural gap between technical squads and traditional insurance underwriters?

2. Structural Analysis

Using the Value Chain lens, RIMAC has successfully optimized support activities (Technology Development) but faces friction in primary activities (Marketing, Sales, and Service). The centralized nature of the AI Factory creates a bottleneck where technical output exceeds the business capacity for absorption. Supplier power in this context is high regarding specialized talent, as the local Peruvian market has a limited pool of data scientists.

3. Strategic Options

4. Preliminary Recommendation

RIMAC should adopt the Hybrid Hub-and-Spoke model. This allows the company to maintain the technical excellence of the AI Factory while ensuring that AI solutions are co-created with business owners. This model addresses the adoption gap by making business leaders accountable for the success of AI initiatives within their domains.

Implementation Roadmap

1. Critical Path

• Month 1-3: Redefine the role of Business Translators. These individuals must move from project managers to product owners within the business units.
• Month 3-6: Establish a federated data governance framework. Business units take ownership of data quality at the source.
• Month 6-12: Transition 60 percent of AI Factory staff into permanent spokes within the Life, Health, and General Insurance divisions.

2. Key Constraints

• Technical Debt: Integration with legacy mainframe systems remains the primary speed inhibitor for real-time AI applications.
• Culture: Underwriters may perceive automated risk assessment as a threat to their professional judgment.

3. Risk-Adjusted Implementation Strategy

To mitigate execution friction, RIMAC must implement a shadow-accounting mechanism. AI-driven gains should be credited to the business unit P and L to incentivize cooperation. Furthermore, the company must invest in a low-code platform to allow non-technical staff to build simple data visualizations, reducing the burden on the central factory.

Executive Review and BLUF

1. BLUF

RIMAC must pivot from an AI Factory model to an AI-Integrated organization. While the centralized factory successfully built technical infrastructure, it has hit a ceiling of business adoption. To achieve the 30 percent efficiency gains targeted by leadership, AI must move from a centralized service to a decentralized utility. The recommendation is to transition to a Hub-and-Spoke model within 12 months, shifting accountability for AI performance to business unit heads while the center maintains technical standards and common infrastructure.

2. Dangerous Assumption

The single most consequential premise is that technical scalability automatically leads to business value. The current analysis assumes that if the models are accurate, the business will use them. However, if underwriters and sales agents do not trust or understand the outputs, the technical precision is irrelevant.

3. Unaddressed Risks

• Adverse Selection Risk: If AI models prioritize speed over accuracy in a low-penetration market, RIMAC may inadvertently attract high-risk profiles that legacy systems previously filtered. (Probability: Medium; Consequence: High).
• Talent Poaching: As RIMAC trains the best data scientists in Peru, global tech firms or regional competitors may target these individuals, creating a recurring knowledge vacuum. (Probability: High; Consequence: Medium).

4. Unconsidered Alternative

The team failed to consider an Acquisition-Led Strategy. Instead of building all capabilities internally, RIMAC could acquire a mature Peruvian or regional Insurtech. This would provide an immediate injection of agile culture and pre-integrated AI products, bypassing the internal friction of the AI Factory model.

5. MECE Verdict

APPROVED FOR LEADERSHIP REVIEW