Launching Australia’s National Sustainability Data Platform Redesigning-Australias-Business-Digital-Identity-Experience Redefining-Mobile-Banking-Interaction Accelerating AI Insight Without Losing Depth Fueling Growth with Customer Insight Steering Smarter Digital Investment Across Government National Health Records From Friction to Function Connecting Officers to Real-Time AI Situational Awareness Redesigning Authorisation for Inclusive Access From Legacy Quoting to a Live Jobs Pipeline in Eight Weeks

From Legacy Quoting to a Live Jobs Pipeline in Eight Weeks

AYVA - Forward-Deployed Engineering / Manufacturing

AYVA Packaging  supplies to brands such as Selleys, Dulux and other manufacturing and warehouse operations. Embedded as the sole engineer to replace a legacy, ERP-adjacent quoting and job-management workflow with a working, production-facing system; inside an eight-week contract.

Detail: The quoting-to-jobs pipeline in daily use — sales quotes flow through backlog, quoting and in-progress stages to the shop floor, with urgency and overdue flags visible at a glance.

THE CHALLENGE

Critical business rules lived in people's heads, not in the system

AYVA's quoting and job management ran on a fragile workflow sitting beside the ERP. The rules that turned a quote into a job - item codes, BOMs, artwork approvals, supplier communication — were tribal knowledge. The CEO had solo-built internal tools using AI coding assistants, but the estate had no CI/CD, no tests, no security audit and no shared identity layer. The brief: audit what exists, then replace it with something production-grade — without disrupting live quoting.

MY APPROACH

Audit first, then a working proof of concept in five days

began with a security and architecture audit — inventorying apps, repos, databases and credentials, and delivering a severity-ranked findings report. In parallel, I reverse-engineered requirements directly from the operations team, treating their workarounds as the real specification. By day five I had a working proof of concept querying on-premises SAP through SQL Server, generating live quotes and jobs.

Detail: High-level concepts framed to spark discussion, validate assumptions, and seed future-facing ideas during multi-stakeholder co-design workshops.

From there, productionisation; delivered solo:

  • CI/CD, code tests, and a documented discovery and epic backlog on a wiki
  • A custom CRM on SQLite with a ledger-based, append-only architecture; every state change recorded, nothing overwritten - laying the foundation for a future cognitive/AI layer
  • A Control Plane shell authenticating staff against Active Directory, providing authorisation across apps (and designed for agent identities as well), with cross-app logging and history
  • A current-and-future architecture roadmap, so the build outlives the contract

Constraints drove the design: with no server budget early on, apps shipped as local executables; with on-prem servers provisioned but no GPU, the roadmap includes small local inference agents via Tailscale and Unsloth.

THE IMPACT

A production-facing pipeline inside five weeks

By week five, the majority of the quoting-to-jobs pipeline was built and in use — hundreds of quotes and jobs moving from sales to the shop floor, with per-item setup checklists, artwork proofing stages and supplier messaging built in. Business rules that previously existed only in conversation now run in the system, and the ledger architecture makes every quote, job and approval auditable. The remaining weeks are hardening, handover documentation, and preparing the AI layer the architecture was designed for.

Detail: Per-item setup workflow - artwork receipt, supplier proofing, packaging, manufacturing and BOM synthesis tracked as auditable stages for every SKU.