Private AI content workflow

Content Writing Agent

An in-development AI-assisted writing workflow intended to help stakeholders draft clearer enrollment and brand-aligned content while preserving human review and policy constraints.

Internal Systems Shipped 2024 - Present

Role

AI Workflow Builder / Technical Lead

University of Arizona

Tags

Applied AI WorkflowContent GovernanceHuman Review

Problem

Content writers needed help producing branded, audience-specific copy because generic AI output sounded too obviously AI-generated and did not reliably follow university or department guidance.

Users

Content writers, content team leadership, and division teams that need enrollment- and brand-aligned copy.

Ownership

  • Owned project inception, product design, AI workflow design, implementation, coordination, prototype, and production version.
  • Worked from the content team manager's pain point into an internal workflow that could support broader division use.
  • Designed the interface and workflow around how writers actually create and improve copy.

Hard Parts

  • Designing a UI that made sense for channel-specific briefs, voice, style, audience, and writing goals.
  • Supporting saved examples of strong outputs so the system could improve with more frequent use.
  • Adding research mode over approved university-specific sources while cross-referencing university and department writing guidelines.

Leadership

  • Turned a stakeholder complaint about AI writing quality into a product workflow.
  • Coordinated from concept through design and implementation.
  • Kept human review and policy constraints central instead of treating the tool as unsupervised generation.

Shipped

  • Prototype and production workflow.
  • Channel-specific briefs for voice, style, and audience.
  • Saved examples of good outputs for future improvement.
  • Research mode using approved university-specific sources.
  • Guideline cross-reference against university and department writing standards.

Impact

  • Improved content quality and made the drafting workflow easier for writers.
  • Reduced complaints about AI-sounding copy.
  • Created a safer AI writing process tied to brand and audience context.

Highlights

  • Framed the work around stakeholder content quality, enrollment messaging, brand alignment, and safer AI-assisted drafting.
  • Focused on human review and policy constraints rather than unsupervised publishing.
  • Connected the workflow to broader content editing, guidance, and review needs for non-technical teams.
  • Treated AI drafting as a stakeholder enablement problem, not just a generation feature.

Tools

LLM workflowsPrompt/workflow designStructured outputsHuman reviewBrand guidanceEnrollment content workflowsStakeholder enablement