proof
Every engagement ends. The systems don't.
Five case studies, told the same way I run engagements: what was breaking, what was actually wrong, what got redesigned, what got built, who owns it now — and the only metric that matters in the end: what's still running.
case studies
airbnb · ai workflow transformation
The Friction
The Help Center's manual content operation was straining, and AI was arriving whether the workflow was ready or not — with no evaluation criteria, no unified standards, and seven style guides that disagreed with each other.
The Diagnosis
This wasn't a writing problem or a tooling problem. It was an operations problem: the workflow couldn't absorb AI because nothing existed to govern what “good” meant — and no process connected the teams who'd have to agree on it.
The Redesign
I audited the manual workflow, redesigned the operational process, and standardized the team's SOPs. Then I proposed a new AI-assisted content workflow — escalated to the CEO — and designed the operating model to run it alongside the manual practice, aligning UX Writing, AI, Chatbot, and Machine Learning teams around a way of working none of them individually owned.
The System
The governance and evaluation layer AI-generated support content needed to be trusted: the first curated training dataset, seven style guides consolidated into unified standards, voice-and-tone evaluation criteria for judging model outputs, and human-preference evaluation run on hundreds of paired outputs — ahead of a global rollout localized into 61 languages.
The Hand-off
The framework launched a second team, staffed and operating — alongside the manual team, not replacing it.
What's Still Running
Both teams, today. The support flows I shipped — ID verification, on-trip issues, refunds — are still live in the app. And the 8-week content test I led became Airbnb's first self-serve refund resolution experience: measurable ticket deflection, an estimated $3M saved.
coursera · capability from zero
The Friction
Four product organizations — learner, enterprise, institution, government — with no shared content practice, no function to build one, and 1,500 people writing at cross purposes.
The Diagnosis
Not a staffing gap — a capability gap. Hiring writers into a system that didn't exist would have scaled the inconsistency, not solved it.
The Redesign
I established the content design function from the ground up: the operating model, the cross-functional workflows, and design-operations practices — before a formal Design Ops team existed to ask for them.
The System
A shared content design system and narrative architecture spanning all four product organizations, with accessibility standards (WCAG, ARIA) embedded in Storybook and decisions validated through user research and multivariate testing.
The Hand-off
A hired, mentored team of three — and ~200 designers, product managers, and engineers trained through workshops, office hours, and mentorship to write to the system without anyone standing over it. This engagement is where the method's core belief was earned: adoption is a teaching problem, not a documentation problem.
What's Still Running
The system was adopted company-wide and remains in use at scale — years after I left.
navy federal credit union · standards in a regulated environment
The Friction
A support chatbot serving three distinct audiences, authored by teams working from departmental style guides that didn't agree — in a regulated financial environment where inconsistency isn't a quirk, it's a risk.
The Diagnosis
The chatbot wasn't the problem; the authoring system behind it was. Every department writing by its own rules guaranteed the chatbot couldn't speak with one voice.
The Redesign
An audit of the chatbot across all three audiences, and a consolidation path from many standards to one.
The System
A single content standard inside the client's Microsoft environment, plus a content-operations framework and roadmap giving the team a repeatable way to author, review, and govern chatbot content.
The Hand-off
I trained the chatbot lead and three stakeholders to run the system — four named owners, not a binder on a shelf.
What's Still Running
The four-person team maintains the system independently. The engagement was four weeks; the capability is theirs permanently.
luxe incentives · operations rebuilt
The Friction
A growing incentive-travel business running on ad-hoc processes — every proposal, campaign, and follow-up handmade from scratch, with the owner as the bottleneck.
The Diagnosis
Not a marketing problem — an operating model problem. Nothing repeatable existed for the AI tools to accelerate, and automating chaos just makes faster chaos.
The Redesign
I restructured the team and redesigned the sales-and-marketing workflow into something repeatable before connecting any technology to it.
The System
A connected stack — HubSpot, Airtable, custom GPTs — implementing the redesigned workflow, plus branded proposal infrastructure and a three-year growth-and-exit path toward a scalable, AI-enabled platform.
The Hand-off
The owner and team run the stack; the workflow no longer depends on any single person's memory.
What's Still Running
The system produced the proposals that won Abbott, Rocket Money, and Celebrity Cruises — and the operating model continues to carry the business toward the three-year plan.
ux content collective · capability at industry scale
The Friction
An entire profession being reshaped under its practitioners' feet — content leaders asked to govern AI-mediated work with no method, no shared vocabulary, and no curriculum that treated the organization, rather than the tool, as the unit of change.
The Diagnosis
The industry didn't need another prompt course. It needed an operating discipline: a way to assess organizational design, business functions, governance, and knowledge ecosystems as the foundation for AI adoption.
The Redesign & The System
Two original eight-part frameworks, built and refined across a four-year engagement: a Content Operations framework (operating models, governance, workflows, decision rights, collaboration practices) and an AI Enablement framework, delivered as the AI Content Strategy workshop.
The Hand-off
Hundreds of practitioners trained globally — including teams supporting Fortune 100/500 organizations — carrying the method into their own companies. Alongside the teaching, I advised the CEO across the full engagement on product strategy and the company's expansion into AI education.
What's Still Running
The frameworks live on in advanced curriculum — including two Dynamic Content Strategy modules in the Advanced UX Content course — and in initiatives adopted into enterprise AI programs. The method now walks around in several hundred pairs of shoes.
Your organization could be next.
Every case above started the same way: one email about a real problem. Thirty minutes, no deck.