About McCarthy Tire Service

McCarthy Tire Service is a 99-year-old commercial tire and wheel services company, ranking among the top 5 commercial tire businesses in North America. They operate 100+ retail locations and a fleet of field technicians servicing tires and performing on-site repairs for customers. 

Turgon Services Provided

• Data Ontology & Lineage
• Data Migration & Pipeline Build 
• Data Warehouse Build 
• Real-Time Analytics Integration 
• Semantic Data Interface

The Challenge: When Legacy Infrastructure Blocks Business Agility

Over 100,000 companies worldwide run their most critical operations on IBM i systems launched 35+ years ago. Major retailers, banks, and insurers rely on this foundational platform for inventory, financial, and customer records. However, the people with institutional knowledge to maintain and modernize these systems are retiring. Connecting legacy infrastructure to modern cloud analytics and AI-powered decision tools remains prohibitively expensive and time-consuming. 

McCarthy Tire Service faced exactly this IT dilemma. Their IBM i AS/400 mainframe was the backbone of their business, but it was increasingly becoming an operational blocker. 

Time-consuming, manual reporting 

McCarthy's IT team spent hours each week manually extracting sales and inventory data from their IBM i system, cleaning and aggregating it across 100+ retail locations, then compiling Excel reports for management. 

Delayed decision making

Leadership needed near-real-time insights to optimize inventory, sales, and purchasing decisions across their growing business. The manual reporting process meant leadership was always looking at days-old data. 

Undocumented system complexity

McCarthy's IBM i system contained 30+ years of critical business data across 40,000+ tables with limited schema documentation. Understanding and accurately transforming this data for cloud data warehouses would take months of discovery workshops and would likely still be incomplete. 

Turgon's Solution: Specialized AI Agents Learn Directly from the System

Turgon deployed its pod of specialized AI agents with human validation to execute each job by system. Instead of humans teaching IT consultants who then document the IBM i system, AI agents analyzed the database structure, data patterns, and relationships to create a semantic model. This became the foundation for other AI agents working on pipeline generation, data migration, and cloud analytics.

Build complete semantic model in four days

Turgon's Data Dictionary Agent built a complete semantic model of McCarthy's entire IBM i system in four days. The agent analyzed table structures and relationships across 40,000+ tables, examined actual data patterns to infer business logic that had never been formally documented, cross-referenced historical query logs to understand how tables were typically used together, and flagged inconsistencies for human review. The agent identified implicit relationships that would be invisible to human consultants: foreign key patterns inferred from data values rather than schema definitions, hierarchies inferred from naming conventions across decades of development, and lineage inferred from the sequence in which tables were queried in common workflows.

Turgon's data engineers then worked with McCarthy's team to validate the semantic model against known business processes. Where the AI agents identified multiple possible interpretations of table relationships, human experts provided the business context to select the correct mapping. The result was a complete, verified semantic model documented in a machine-readable knowledge graph.

Design extraction and transformation pipelines

With the semantic foundation in place, specialized AI agents divided the migration work across platform expertise leveraging best practices. The entire pipeline from IBM i through S3 to Snowflake was designed, coded, tested, and deployed in weeks.

A Data Integration Agent designed extraction logic to pull data from IBM i into Amazon S3 using Apache Iceberg format. Coding Agents generated transformation code implementing a bronze/silver/gold medallion architecture in Snowflake: raw data from IBM i landed in the bronze layer, transformation logic cleaned and standardized the data into silver, and business logic aggregations prepared analytics-ready datasets in gold. 

Turgon’s IBM i AI agents ensured extraction queries respected the legacy platform's performance characteristics and connection limits. Snowflake AI agents optimized warehouse configurations and clustering strategies for the expected query patterns. 

Deploy real-time analytics dashboards

Turgon built dozens of interactive dashboards in AWS QuickSight connected to the Snowflake warehouse. Leadership gained instant visibility into inventory levels across all 100+ locations, sales performance by region and product line, and purchasing patterns to optimize vendor negotiations. Dashboards refresh automatically as new data flows through the pipeline. What used to require manual Excel compilation every week now updates in near-real-time.

Ongoing data observability through semantic interface

Turgon deployed a Data Dictionary Chat interface that allows McCarthy's team to query the semantic model in natural language. Business users can ask questions about table relationships, field definitions, and data lineage without needing to understand database schemas or SQL. 

For example, when someone asks where the authoritative record for a specific customer lives, or what the current inventory level for a particular SKU is, the chat interface identifies the correct source table, retrieves the data, and returns it in plain language. This makes their IBM i system self-documenting and accessible to non-technical stakeholders, and future integrations and answering data questions no longer require rediscovery work.

The Results

Witnessing Turgon’s speed and exceptional quality of work, McCarthy Tire Service’s CIO Lee Lispi is a true believer of the power of AI-led infrastructure modernization and sees future opportunities to enhance their operations throughout his organization.

Three times faster delivery 

Turgon delivered the complete semantic model in four days. Full production pipelines and interactive dashboards went live in five weeks. Traditional systems integrators scoped the same project taking 6+ months. Turgon's AI agents worked 24/7 and handled 90% of the semantic modeling, pipeline code generation, and integration work that humans would typically perform over extended timelines. As a result, McCarthy's team was able to access real-time insights earlier and optimize inventory and purchasing decisions.

70% cost reduction 

Turgon’s agentic approach reduced costs by 70% compared to traditional system integrators that quoted $500K to $1.5M. Turgon delivered the same scope of work with 3 subject matter experts/engineers plus AI agents in 5 weeks, while a traditional SI project would require 8+ consultants/engineers using standard discovery and development methods consuming thousands of hours. 

Zero operational disruption

AI agents built extraction pipelines that read from IBM i without impacting production system performance. McCarthy's operations continued normally while the new analytics infrastructure was built and validated in parallel. 

Living infrastructure for autonomous operations

McCarthy’s IT team now has institutional knowledge embedded in its infrastructure, self-documenting and continuously maintained. The semantic foundation that took four days to build is accessible for every downstream application, from analytics dashboards to future AI-driven apps, making it possible to add new use cases and autonomous workflows, without additional costs or delays.