AI Adoption Discovery
For teams with a promising AI workflow, real materials, and unclear adoption risks.
Prototype the workflow, define eval questions, and decide what is worth building
Enablement Engineering
Reviewable AI workflows for accessibility, education, and high-trust content teams.
We build Evals Agent tools Skills Adoption loops
Start with real materials; leave with representative tests, review gates, and handoff routines.
AI Capability Is Not Adoption
We build the evals, agent tools, demos, training assets, and feedback loops that move LLM workflows from prototype to durable practice in accessibility, education, and governed content operations.
Production AI Workflows
For teams ready to turn a proven workflow into agents, tools, review gates, and operational infrastructure.
Build the deployment surface, eval harness, and acceptance criteria
Enablement & Handoff
For teams that need training, use-case libraries, demos, and feedback loops after the first workflow ships.
Teach the team, watch drift, and keep practice improving after launch
Best fit
For teams with real content, review pressure, and AI output that must be checked.
The strongest projects start with a workflow people already care about: course materials, accessibility review, content transformation, research workflows, or operational processes where mistakes have consequences.
Real materials
The work uses actual documents, examples, edge cases, constraints, and stakeholder expectations instead of idealized demo inputs.
Visible review
Outputs need acceptance criteria, review gates, correction paths, and evidence people can challenge before they trust the system.
Practice transfer
The team should leave with procedures, training surfaces, and feedback loops they can operate without treating the AI as a black box.
Featured Work
University of Illinois Chicago — Digital Accessibility Services
Equalify Reflow
A production AI pipeline that turns inaccessible PDFs into semantic, accessibility-first markdown. Built for UIC Digital Accessibility Services and released under AGPL-3.0.
Featured Writing
May 6, 2026essay
POUR for Agents
AI agents are now participating in accessibility work. The question is whether we've made our practice available to them. POUR — perceivable, operable, understandable, robust — wasn't just for users.
- accessibility
- ai-agents
- context-engineering