Enterprise ERP Migration

BUSINESS PROBLEM

A major medical equipment manufacturer based in Oregon was migrating from a legacy Microsoft Dynamics AX ERP system to Microsoft Dynamics 365 Finance & Operations. Two large data-engineering challenges came with the initiative, and HERA was brought in to solve the following:

1. Years of legacy AX historical data and live D365 data had to be merged into a single unified history for each business entity so reporting wouldn't fall off a cliff at go-live.

2. The enterprise data warehouse consumption layer (sales, production, inventory, financial dimensions, BOMs, product, vendor) had to be rebuilt to sit on top of D365 schemas while preserving the legacy history.

OUR APPROACH

After speaking with key stakeholders, end users, and other impacted parties, HERA utilized our skills in Snowflake, Coalesce.io, Microsoft Dynamics 365 Finance & Operations, Microsoft Dynamics AX (legacy), Azure DevOps, Streamlit, Python, SQL to deliver:

· AX-to-D365 merge pattern for multiple entities (Inventory Location, Financial Dimension, Sales Line, and others) using a multi-branch selector node design — backloads legacy history once, then streams D365 going forward, transparent to downstream consumers.

· Standardized view-stage / history / output node patterns applied consistently across 15+ data subgraphs.

· D365 timezone-handling fix pattern rolled out across every date-cast in the warehouse — D365 stores dates as UTC midnight with a time zone offset, which silently produced off-by-one-day errors in the standard SQL date-cast patterns. I identified the root cause and applied the fix across all affected entities.

· Forecast pipeline cleanup: identified a duplicate-row defect in a supplier forecast feed, fixed via aggregation in the staging view while preserving every quantity.

· Data Health Monitor: built a Streamlit operations dashboard backed by a Snowflake stored procedure + error table + scheduled task to give the data team a

single-pane view of which transform jobs are failing and why. Includes a written operations manual covering token rotation, troubleshooting, and setup.

· Production deployments: ~15 deploy tickets closed in a single cutover session — ran jobs, validated row counts in Snowflake, posted result snippets, handed off to the client's scheduling engineer.

Project/Change Management

SERVICES PERFORMED

Technology Assessment

Requirements Gathering

Custom Integration

ERP Migration

VALUE DELIVERED

· ERP go-live unblocked. All assigned merge entities delivered to QA / Production on schedule.

· ~15 production tickets closed in a single deploy cycle.

· Recurring data-quality issue eliminated: the time zone-offset pattern fix prevented an entire class of off-by-one-day errors across the warehouse — would have been near-impossible to debug after the fact in downstream reporting.

· Operations dashboard delivered so the client's internal team can monitor pipeline health themselves rather than depending on engineering support.

· Several reusable patterns documented (legacy-to-modern merge pattern, time zone fix, hash-column / SHA256 migration approach) that the client's team has continued to apply to entities I didn't personally touch.

15+

Production Tickets Closed

ERP

Go-Live Unblocked

Ops

Dashboard Delivered