AI-Ready Content Taxonomy

The Challenge

PVS's enrollment knowledge base housed over 800 articles in Salesforce—but the content ecosystem lacked the structure needed to support intelligent search, AI integration, or scalable content governance. As enrollment processes evolved and new tools like an AI-powered chatbot and a centralized Enrollment Portal neared rollout, the system’s underlying architecture could no longer keep up.

Key challenges included:

• • Content Fragmentation: Articles were inconsistently categorized, with duplicative or conflicting entries across teams and SharePoint sites.

  • Poor Discoverability: Inadequate tagging and non-standard metadata made it difficult for staff and families to find timely, relevant information.

  • Inflexible Structure: The existing categorization was too role- or team-specific, preventing the knowledge base from scaling or adapting to changing workflows.

  • Lack of Readiness for Automation: Without a standardized taxonomy, content could not be reliably surfaced by AI tools or repurposed across platforms.

  • Decreased User Confidence: Both staff and customers reported frustration due to content inconsistency, search failures, and reliance on outdated resources.

The challenge wasn’t just about cleaning up content—it was about building a system that could support automation, improve service delivery, and scale with the business.

My Role

I served as the strategic lead and architect of the taxonomy initiative, responsible for defining its structure, alignment, and downstream utility across platforms and users. This work required balancing business logic, user needs, and system capabilities to ensure the taxonomy could support both human and AI-driven retrieval.

Key responsibilities and decisions included:

• • Conducting stakeholder interviews across enrollment, support, and product teams to uncover core findability and structure challenges within the knowledge base

  • Auditing and inventorying 800+ articles to assess redundancy, outdated content, and alignment with enrollment workflows

  • Designing a hierarchical taxonomy model grounded in PVS's business structure and customer journey logic to enable intuitive navigation and intelligent retrieval

  • Creating a metadata and labeling system to support machine-readable content and optimize search performance

  • Building a content classification guide and workflow to ensure long-term consistency and governance across distributed teams

The Solution

To address fragmented content, poor searchability, and a lack of AI readiness in PVS's knowledge base, I developed a scalable taxonomy and metadata framework that restructured how enrollment content was organized, labeled, and surfaced in Salesforce Knowledge.

Key components of the solution included:

• • A five-level taxonomy model that mapped content to PVS's business structure, enrollment workflows, and user journeys. This hierarchy replaced inconsistent, team-specific categorization with a unified framework that could scale across business units and adapt to future needs.

  • A metadata-driven labeling system that embedded consistent attributes—such as business function, visibility, and content format—into each article. This laid the foundation for smarter search filtering and machine-readable classification, which would later support chatbot and portal delivery.

  • A systematic content audit that eliminated over 50% of outdated, duplicate, or irrelevant articles. This streamlined the knowledge base, reduced user confusion, and ensured only accurate, actionable content remained.

  • A standardization model for article structure to improve readability, ensure consistency, and enable AI tools to extract key information more effectively.

  • Documentation and governance guidelines to support ongoing maintenance, future tagging, and scalability across new teams or initiatives.

The solution not only improved internal access to knowledge—it created the structural foundation that enabled AI-powered tools and the Enrollment Portal to deliver accurate content with speed and consistency. By making the knowledge base both human- and machine-readable, this work positioned PVS to scale support across platforms without sacrificing quality or trust.

Results and Impact

The taxonomy initiative transformed PVS's enrollment knowledge base into a streamlined, intelligent system that improved day-to-day usability while enabling long-term digital scalability.

Key outcomes included:

• • Improved content discoverability through consistent classification, metadata, and labeling

  • Enabled AI-powered chatbot and Enrollment Portal rollout by making content machine-readable and structurally aligned

  • Reduced article volume by 50%, removing outdated and duplicative content while increasing trust and clarity

  • Established a governance model that supports sustainable content lifecycle management across teams

  • Improved the enrollment experience for staff and families by enabling faster, more accurate access to information

  • Positioned PVS for future scalability by creating a taxonomy infrastructure adaptable to evolving systems, roles, and user needs

This work became a foundation for PVS's digital transformation—connecting strategy, structure, and user experience through intelligent content design.

Reflection

This project reinforced that content architecture is not just about organization—it’s about creating systems that empower both people and technology. Designing a taxonomy for AI-readiness required seeing beyond individual articles and understanding how structure influences behavior, automation, and access at scale.

What began as a cleanup effort became a foundation for clarity, consistency, and intelligent delivery. It challenged me to think systemically, communicate cross-functionally, and build a solution that could evolve with the business—not just support it temporarily.