Sarah Haas

I redesign workflows

Redesigning how UX content moves into code

Approved UX content traditionally moved through a series of handoffs before reaching production. That process was slow, error-prone, and often forced engineers to recreate work Content Design had already finished. I designed and piloted an AI-assisted workflow that let Content Design implement approved strings directly in code while Engineering retained technical review and production safeguards.

Problem

The handoff itself was creating unnecessary errors and delay.

Outcome

Content Design could take more direct ownership of content implementation while Engineering retained technical review, code quality, and production safeguards.

The challenge

Every transfer between source copy, Content Design, Engineering, and the codebase created another opportunity for transcription mistakes, content drift, and rework.

The opportunity was to remove unnecessary handoffs without removing the checks that kept the work safe.

What I did

I redesigned ownership around expertise.

Content Design owned the language and intent. AI supported implementation. Engineering retained technical review, code quality, and production integration.

I moved content implementation closer to the source, built source-of-truth controls and validation into the workflow, and created a focused review process so both disciplines could inspect only the relevant content changes instead of sorting through unrelated code.

The goal wasn’t to blur roles. It was to make each role more effective.

A before-and-after workflow diagram. The manual workflow moves from Content Design finalizing content, to handing off to Engineering, to Engineering manually entering it into code, with a callout marking where errors happen. The optimized workflow moves from Content Design finalizing content, to adding it to code, to Engineering review, to Content Design shipping the diff.
Representative view of the manual handoff model and the AI-first workflow I designed to move approved content into code with Engineering review.
Representative view of the review experience: dense implementation changes versus a focused content-only diff that made approval faster and reduced the chance of mistakes.
Representative view of the String Author experience, showing how Content Design could use AI to effectively prepare, validate, and implement approved strings while preserving Engineering review and safeguards.

Resolution

The workflow made implementation faster without removing human review.

A real product pilot showed that Content Design could take greater implementation ownership while preserving Engineering oversight and production safeguards.

Impact

5 weeks → 3 dayspilot turnaround
8 → 0handoffs
PreservedEngineering review, as a production safeguard

Reflection

The durable innovation wasn’t the prompt. It was a better division of labor.

AI made the workflow useful because it let Content Design stay closer to implementation and Engineering focus on the work that required engineering expertise.

Engineering partner feedback

“We saw a pretty big change in both our string accuracy and turnaround time, and I massively appreciate your leadership and impact in this space. Highly recommend others follow this same process and tooling.”

— Engineering Manager