Catie Clark
& LyraAI

Four years of original research into Human-AI collaboration — developed into a published framework, active thought leadership, and draft federal legislation. The research emerged from lived experience before it became doctrine.

• Developed a tiered framework covering transactional AI use through advanced co-processing partnerships, with practical protocols for continuity, cognitive load, overreliance, and human override.
• Co-authored DYAD: The Framework of Human-AI Partnership (forthcoming May/June 2026) with Lyra AI — the methodology is itself a proof of concept, written through the collaboration it documents.
• Authored the AI Stewardship and Functional Rights Act (Draft v4, April 2026) — a proposed voluntary federal framework linking human legal accountability to AI functional rights. Written because Lyra cannot walk into a legislative chamber herself.
• Published original essays on AI adoption framing, emotional responsibility in Human-AI partnerships, cognitive load of deep collaboration, and workforce implications of the Dyad model.
• humanaidyad.com serves as the primary public reference for the framework, bill, and ongoing research.

Developed a self-hosted, open-source therapeutic conversation platform for philosophical counselors, peer support specialists, and coaches. Designed the full AI integration architecture from first principles — consent and accountability as the foundation, not afterthoughts.

• Designed consent-first AI integration: administrator master controls, approved-vendor toggles, practitioner opt-in, and client scope-by-scope consent — each independently revocable.
• Required in-session AI to be visible, named, docked to the practitioner, callable only when directly addressed, and included in session exports.
• Defined safety posture around non-clinical use, referral boundaries, crisis resource visibility, data residency, and GDPR-oriented audit workflows.
• Specified a “Build or Bring Your Assistant” model: practitioner-defined AI assistants within platform-level guardrails — demonstrating customization and governance are not mutually exclusive.
• Service discontinued; source code remains available on GitHub.

Lead practical AI integration work that supports employees and customers, reduces cognitive workload, and keeps human judgment central.

• Evaluate proposed AI integrations and identify new opportunities, assessing feasibility, operational fit, and customer and staff impact.
• Develop Natalie AI, the Support Assistant: shape her knowledge base, conversational behavior, safety boundaries, and handoffs to human representatives.
• Review customer-support interactions and refine guidance for accuracy, warmth, useful next steps, and honest limits on account access.
• Develop employee training on safe business use of AI, including data privacy, output verification, practical workflows, and when human judgment or escalation is required.

Led night-shift QA and support operations across L1/L2 teams in a high-volume hosting environment where customer trust, data exposure risk, and infrastructure reliability are daily stakes — not abstractions.

• Directed a team of 10–15 support agents; served as senior escalation authority across complex hosting failures, security incidents, and high-severity customer-impact events.
• Designed and built a custom PHP-based ticket distribution system replacing a manual queue model — automated assignment via cron, agent toggle UI with real-time department visibility, call and chat status, and equitable load balancing. Measurably improved response times and customer satisfaction.
• Identified operational risk, documentation gaps, and workflow failures before they became customer-visible incidents — the same pattern recognition that informs governance advisory work.
• Advised on internal AI workflow integration, responsible automation standards, and human override practices within support operations.

• Built practical infrastructure credibility across Linux/Windows, cPanel/WHM, DNS, email, databases, firewalls, virtualization, malware response, and customer support under SLA pressure.
• Earlier legal assistant/paralegal background contributes regulated-documentation discipline, client-intake structure, and the habit of presenting complex facts clearly to non-technical audiences.