Kempe Engineering Pty Ltd Private · ABN 23 118 446 110
Specialist Engineering Services · Aluminium Smelting, Mining & Heavy Industry · Geelong HQ · Global Operations
~A$79M
Revenue (est. 2025)
~60
Employees
1999
Founded
Australia · South Africa · Middle East · Asia
Global Regions
~38,000
SKUs (Master Data)
Your AX2009 has 38,000 SKUs and no automation layer. I've built that layer three times — including on Dynamics AX.

I have 27 years building centralised planning systems on top of legacy ERPs. At Pacific Brands I unified 6 ERPs into one engine with mandatory CEO sign-off governance. At Australian Defence Apparel I worked directly in Dynamics AX — your platform — managing complex military bills of material and expensive defence-grade inventory where precision is not optional. At Saputo Dairy Australia I built a 7-layer architecture on SAP ECC: 10,000+ lines of VBA, 231 Power BI visuals, daily dashboards across Saputo's Australian operation. The Saputo DOT system — the most recent of the three — was delivered using AI-augmented methodology with Anthropic's Claude as a genuine build partner, not Copilot autocomplete. That AI-as-partner methodology is what I'd bring to your AX2009 environment, applied on top of 27 years of structural planning-system experience.

Your ad describes a funded strategy, autonomy to deliver, and a team in place. Below is what I found when I researched Kempe, paired with exactly how I'd approach each challenge. I built this entire site — analysis, industry report, and the platform delivering it — as a working demonstration.

Robert M. Clark · robert@robertmclark.dev · 0447 515 074 · Warrnambool, VIC · Hybrid: remote + on-site Geelong
COMPANY ANALYSIS & MY APPROACH
Industry Pulse — Heavy Engineering & Aluminium
KEMPE'S SITUATION
Legacy ERP — Dynamics AX2009

Kempe runs Microsoft Dynamics AX2009, end-of-mainstream-support since October 2018. Extended support ended 2021. Rather than a $2–10M+ migration to D365, Kempe is taking the pragmatic route: layering RPA and AI automation on top of the existing platform.

This is a greenfield build reporting directly to the Operations Director. Strategy is funded. Team is in place.

MY APPROACH
I've Automated on Top of AX Before

At Australian Defence Apparel I worked directly in Dynamics AX — military bills of material, defence-grade inventory, zero tolerance for error. At Pacific Brands I built automation across 6 different ERPs without replacing any of them. At Saputo I built the DOT system on top of SAP ECC while the ERP stayed untouched.

Week 1–4: Map every data flow in and out of AX2009. Identify the extraction points, the manual workarounds, the reports people are building in Excel because AX can't do it. Month 2–3: Build the first automation layer — VBA/Python harvesting AX data into a structured reporting environment. Month 4+: AI overlay for search, dashboards for management, governance framework locked in.

KEMPE'S SITUATION
38,000 SKUs — Master Data Challenge

The ad explicitly calls out master data cleansing and enrichment across ~38,000 SKUs to automation-ready standards. In engineering services — manufactured products, spares, hydraulics, custom fabrications — SKU proliferation and data quality are chronic. Duplicates, inconsistent naming, missing attributes, orphaned records.

This is the foundation. Nothing automates cleanly until the data underneath it is right.

MY APPROACH
I've Cleaned Master Data at This Scale

At Saputo I managed product hierarchies across 14 manufacturing sites with shelf-life constraints, production scheduling, and demand-supply matching. The DOT "Gold Chain" was specifically a master data enrichment pipeline. At Pacific Brands, multi-division product catalogues across workwear divisions.

Approach: AI-assisted classification and deduplication. Export the full SKU table from AX, run it through an AI model to flag duplicates, standardise naming conventions, and enrich missing attributes. Human review on the exceptions. Target: 15–25% SKU rationalisation, which directly reduces working capital and procurement complexity. This is month 1–2 work — it runs in parallel with the ERP mapping.

KEMPE'S SITUATION
AI Overlay for Search, Reporting & Dashboards

Kempe wants a structured environment with an AI overlay for search, reporting, and management dashboards. The business data and IP need consolidating into a system that operations and leadership can actually query.

This isn't R&D. This is production tooling for daily decision-making across a multi-branch, multi-geography operation with clients across the Gulf, South Africa, and Australia.

MY APPROACH
The DOT System at Saputo IS This

At Saputo I built exactly this: a structured reporting layer with 231 Power BI visuals, daily operational dashboards, and an AI-augmented development process that produced the entire system. It was consumed daily by planners, managers, and executives across the business.

For Kempe: Power BI dashboards connected to AX2009 data via automated extraction. Branch-level inventory visibility. Procurement analytics. Client contract performance. AI-powered natural language search across your consolidated data — so an engineer can ask "show me all hydraulic cylinder orders for Qatalum in the last 12 months" and get an answer in seconds, not hours of spreadsheet work.

KEMPE'S SITUATION
Automation Standards & Governance

The ad calls for establishing automation standards, documentation, and governance frameworks. Plus mentoring and building capability within the AI Systems team through hands-on delivery and code review.

Kempe is a 25-year-old engineering company building its first digital automation programme. The standards set now will define how this scales.

MY APPROACH
I Build Systems That Survive After I Leave

At Pacific Brands I built S&OP governance with mandatory CEO sign-off on branded dashboards per division. Physical sign-off sheets. No dashboard, no meeting, no decision without the system. At Saputo the DOT system included version control, documentation, and operational governance for daily planning.

For Kempe: Every automation gets documented as it's built — not after. Code review is built into the development cycle. The team learns by building alongside me, not by reading a manual after I've gone. Standards aren't a phase at the end. They're how the work gets done from day one.

KEMPE'S SITUATION
Multi-Branch Global Supply Chain

Branches in Geelong, Melbourne (Campbellfield), Perth (Canning Vale), QLD (Hemmant/Berrinba/Gladstone), Portland VIC, plus South Africa and Middle East. Warehouse in Wuxi, China. Centralised order management from Geelong HQ.

Major smelter clients: Sohar (Oman), Qatalum (Qatar), EMAL/Dubal (UAE). Long-term maintenance contracts requiring precise spare parts management and predictive ordering.

MY APPROACH
Multi-Site Is My Entire Career

At Pacific Brands: multi-site national operations across VIC/NSW/NZ, consolidated reporting to HQ, supplier management across geographies. At Saputo: 14 manufacturing sites with different product profiles, all feeding into one planning system. Branch consolidation isn't a new problem for me — it's an architectural pattern I've repeated.

For Kempe: Unified branch reporting — every location's inventory, orders, and contract status visible in one dashboard. Automated purchase order generation based on consumption patterns and lead times. Predictive reorder points for high-value spares. Your engineers in Gladstone and your procurement team in Geelong looking at the same numbers, in real time.

🎯 Skills Match — AI Project Manager (ERP Automation)

INTELLIGENCE & RECENT ACTIVITY
OPERATIONS & TECHNOLOGY NOTES
2026-04-14 09:42 AEST|Kempe AI Project Manager ad identified (Seek #91387733)
09:44|Ad parsed — AX2009, 38K SKUs, RPA/AI, dashboards, governance
09:47|Research — kempe.com.au, LinkedIn, RocketReach, D&B, ZoomInfo
09:55|Skills match: 9 DIRECT, 1 ADJACENT. AX via Australian Defence Apparel
10:01|Dashboard generated — Kempe blue (#6BA3D6) applied
10:35|Industry dashboard — peers, charts, digital maturity
11:45|Paired layout, sticky nav, cover letter, build log, i18n
Total: ~2 hours|One person. One AI build partner. One session.