Skip to content
Back to work
Flagship Enterprise Case Study·Case study· Sanitized case study

From Nightly Batch to Near Real-Time Retail Integration

Redesigned a long-standing nightly retail-closing process into continuous, near real-time integration between multiple stores and the central Dynamics AX ERP — cutting weekly electronic-invoicing errors by roughly 98%.

Role

Problem owner / Architecture / Implementation lead

Technology

.NETREST APIsDynamics AXAIFSQL ServerWindows ServicesClaude Code
Multiplestores integrated
Thousandsof daily sales processed
Hours → Secondsnightly batch to real-time
~98%fewer weekly invoicing errors

Context

For many years, a dedicated person processed store closures every night before sales became visible in the central ERP. I identified the opportunity myself, redesigned the operating model around continuous, demand-driven processing, and owned the architecture, implementation and rollout end-to-end.

Architecture

Store POS databases
Store synchronization services
Central API
Asynchronous .NET workers
Dynamics AX (AIF)
Sales order → Invoice → Payment journal → Settlement
Central ERP, inventory & accounting

Challenge

Sales from multiple stores processing thousands of transactions a day were not visible in the central ERP until a nightly batch job closed each store — a process that took hours on a normal night and considerably longer on a difficult one, and had defined how the business operated for many years.

Before → After

Before

One person processed all store closures every night — a multi-hour process on a normal day, considerably longer on a difficult one — a role that had existed for many years.

After

Each store synchronizes continuously; sales become enterprise transactions within seconds, and the dedicated nightly-processing role is no longer needed.

Approach

01

Identified the opportunity myself — this wasn't a specification handed down, but a process I chose to challenge after understanding its real business cost.

02

Redesigned the operating model around continuous, demand-driven processing instead of one large nightly batch.

03

Built store-level synchronization services and a central API backed by asynchronous .NET workers so stores are never blocked waiting on the ERP.

04

Integrated with Dynamics AX AIF to generate sales orders, invoices, payment journals and settlement automatically.

05

Handled the full technical complexity of moving to continuous processing: inventory availability, lot numbers, kits and component products, transaction validation and failure handling.

06

Used Claude Code alongside my existing Dynamics AX, business and infrastructure knowledge to accelerate implementation — the modernized platform was built in about a month.

Key insight

This wasn't a specification I received and implemented. I identified the opportunity, redesigned the operating model, and owned it end-to-end — from proposal to production validation.

What this demonstrates

Removed a long-standing dependency on dedicated nightly processing, shifted more operational ownership to store managers, gave the business far more current inventory visibility, and centralized electronic invoicing — cutting weekly errors by roughly 98%.