W022025–26Delivery / Content & DAMHealthcare, national16-person team

Outcome

Migration governance fixed at the source, not patched after the fact.

16distributed content specialists led across nearshore and offshore delivery

Multi-marketLumApps → SitecoreAI content migration across consumer-facing markets

Governance-firstcorrective re-export and normalization process established before build resumed

Finding the data problem before it shipped

01 / Context

A national hospital system’s enterprise intranet modernization, supporting consumer-facing digital experiences across multiple markets — the core of it is the content migration from LumApps to SitecoreAI.

02 / Problem

Platform exports looked clean until someone actually tried to use them.

Systemic data quality issues surfaced in the exports themselves — not a content problem, an export, normalization, and ownership problem — and it needed governance controls to fix without introducing new ones.

EXPORT VALIDATION FLOW — DIAGRAM

FIG 01 — Platform export to corrected, governed content.

03 / Approach

Find the real problem in the data before building anything on top of it.

01Identify the systemic data quality issues in the platform exports.

02Implement corrective processes — re-exports, normalization, and ownership reconciliation — under defined governance controls.

03Establish and lead validation, analysis, and build phases, deploying multiple AI agents to accelerate stabilization of highly ambiguous inputs across audits, exports, and legacy systems.

04 / What I did

Set up the migration framework itself and managed the data validation, mapping, build, and QA teams while aligning content, UX, engineering, and platform teams.

Led a distributed team of 16 content specialists, and ran recurring working sessions and office hours specifically to converge on taxonomy, media library structure, and component usage rules — the decisions that are expensive to get wrong after migration.

Role: Sr. Migration Project Manager · National hospital system · Nov 2025–Apr 2026

05 / Outcome

The corrective governance process fixed the export issues at the source instead of patching content after the fact, and the taxonomy, media library, and component decisions locked in during the working sessions became the standard the rest of the migration built on.

TAXONOMY STRUCTURE — DIAGRAM

FIG 02 — Taxonomy and component rules locked in working sessions.

06 / What I would do differently

I’d push for an export-quality audit before validation and build start, not during them — by the time the systemic issues surfaced, some downstream work had already been built on bad data and had to be redone.

← All work7 case studies Next / W03 →Governing AI Inside a $750K Scope Change