Human-AI Workflow Architecture
How teams structure work when AI systems perform meaningful portions of analytic and operational tasks — division of labor, handoffs, and where human judgment is preserved by design.
Developing Research Area
Graduate Research
This page describes a developing research area, not a finalized thesis. The intent is to make the working direction transparent and to share progress as it accumulates.
Focus Areas
AI Oversight
What useful oversight looks like in practice when model outputs are produced faster than reviewers can read them, and when verification itself is partially automated.
Parallel Review & Recursive Iteration
Processes which add dimension to AI workflow Architectures. Lorem ipsum dipsum text
Enterprise AI Governance
Policy, procurement, and internal controls that determine whether enterprise AI deployments produce trustworthy outputs and clear accountability lines.
Working Research Questions
- Q01
How should organizations design review pipelines when first-pass output is produced by AI and subsequent review is partially automated?
- Q02
Where do oversight mechanisms degrade silently as throughput increases, and what early indicators are observable?
- Q03
What does meaningful human-in-the-loop look like once humans review only a sampled fraction of decisions?
- Q04
How do procurement and vendor structures shape the de facto governance regime inside firms?
- Q05
What workflow features predict whether AI deployment improves or degrades decision quality over time?
These questions are exploratory and likely to be refined or replaced as the research matures.
Future Papers & Thesis Updates
- In progress
Working paper: A taxonomy of human-AI review architectures
Draft synthesis of observed review patterns across enterprise and research deployments.
- Planned
Case study series on oversight degradation under throughput pressure
Qualitative interviews and structured observation of mixed-autonomy workflows.
- Planned
Thesis proposal — Governance of multi-step AI systems
Framing the research question, methods, and contribution. Target submission forthcoming.
This work is preliminary. Findings, framings, and questions on this page should be read as work in progress rather than published conclusions.