AI · IT · Ops
Jay Carpenter — the AI–IT–Ops guy.
Wherever they intersect, or any one of them alone.
C-suite level excellence engineering
Project ownership from data · floor to board · change only for improvement
I am a data guy with project ownership. I take systematic operational and business data, couple it with real consultancy practice — speaking with people, watching how they use systems, understanding what utopia would look like for their role — and I translate that into cross-departmental, cross-role, cross-level work: systems → operations → business need, from the stack you already have.
That can mean new functionality and features. Automation. Process redesign. It can mean a smart VBA module that finally stops the spreadsheet task that keeps breaking because customer service cannot get order details into the right fields — and the impact of that mess is downstream on the floor. Small enabling wins open doors. I prioritise them on purpose.
Available for project and contract engagements (AU-based; remote-friendly). Operators. Logistics & warehouse startups. One edge of the triad, or the join.
Two markets — same join
I work with two kinds of buyer. Both live in warehousing, logistics, and high-volume ops. They buy different things. The work sits in the same place.
Operators run the network, the DC, the marketplace, the P&L. They need excellence engineering from the inside: data, floor truth, systems that hold under load, AI and IT that serve the service promise.
Logistics and warehousing startups — and scale-ups — sell AI and IT into those same environments. They sell product. They need implementation, pilot, and roll-out support that makes the product true on the floor: process, people, data, shift handovers, exception paths, and the political reality of a live site.
Operators & enterprise ops
You run the network, the DC, the marketplace, the P&L. You need AI, IT, and ops to stop arguing. I bring project ownership from data, floor truth, and delivery that holds the number.
Logistics / warehouse tech startups
You sell AI, WMS add-ons, robotics, visibility, optimisation, labour tools. I sell operational capability to make it work — pilot design, site readiness, roll-out, training that sticks, feedback into your product so “it demoed” becomes “it runs on shift three.”
They are not the same buyer. They buy different contracts. The work sits in the same place: the join between what the system claims and what the floor can actually do. I have lived both sides of that join — hyper-growth fulfilment, multi-site ops, vendor and integration pressure, and the culture gap between sales, product, and the people who pick and pack.
Startups: I am not competing with your product. I am the capability that stops your pilot dying in the lunch room. Operators: I am not a vendor salesperson. I am the person who makes vendors and internal systems earn their keep under real load.
How the work actually runs
- Gather what is true. Operational data, business data, system exports, and the data of experience: walking the process, doing the work where needed, interviewing the people who live in it.
- See the whole chain, not the symptom. I do not only see an inefficient picking method that allows errors. I see the process and the systems that let a pick error become a customer impact, a safety risk, re-work, delay, and a P&L cost.
- Clean data first; make it meaningful. Small enabling wins: fields that work, reports that tell the truth, cost of inefficiency tied to a real line on the P&L — not vanity metrics.
- Roadmap with realistic change capacity. Clear targets, sequenced so the organisation can absorb them. Cross-department ownership. Utopia as a north star; today’s stack as the starting constraint.
- Improve the stack in place. Functionality, features, automation, redesign. Fix the broken spreadsheet task if that unblocks the floor. Hold the number until it sticks.
I have lived every layer
I have used AI to develop real systems in real operational environments — and I have lived through every role that those systems are supposed to serve.
I have been a warehouse picker and a packer. I have unloaded containers in a yard on a forty-degree day in Western Sydney. I have dropped pallets on a forklift. I have sat in the lunch rooms and the pubs with the people who run the floor. I understand the cultural divide between sales and ops — and that the shape of that divide is unique to every business.
I have also sat in the meeting room at 10 pm when a major operation is held up — senior client leadership on site, significant IT issues driving operational delay, the question aimed at the room: why — and been the person who has to own the answer. I have been in the private boxes at the football and the concerts. The range matters: floor culture, executive pressure, commercial hospitality. Different rooms, same obligation to tell the truth about the system.
When things break, I step up. Acknowledge what is broken. Determine the impact now and the impact if we do nothing. Determine immediate actions that start resolution — not a post-mortem that waits for calm. I am strong in a crisis and when something is critical. Ownership is not a title; it is the sequence.
That is not colour. That is method. Roadmaps that ignore lunch-room truth, boardroom pressure, or the 10 pm room do not survive contact with Monday morning.
Shapes of work (anonymised)
No client names. The patterns are the product of data, floor time, and ownership of the outcome.
Outbound scaled from tens of people to well over a hundred. Performance systems that roughly doubled pick rates and lifted pack rates ~75%. On-time despatch into >99.9%. Speed products (same-day interstate, Saturday delivery), live ERP/WMS dashboards, multi-site fulfilment launch — owned as a measurable chain, not a slide.
Purchasing through import, warehouse, freight, installation, and customer experience as one system. Multi-year people / process / systems strategy with owners. Change only when service predictability moves.
IT product, customer service, and contractor supply side under aggressive growth. National multi-city expansion, apps both sides of the marketplace, commercial booking, fulfilment >99.85%, response times cut ~70%. Under NDA with a grant-backed data partner: churn-propensity signals → national location heatmap overlaid with where field contractors actually lived → expansion decisions from that map. Same pattern as any two-sided field marketplace. Short runway; acquisition followed. I am confident that twelve-month footprint changed what the business was worth to a buyer.
Procurement-to-delivery rebuild in a growth sprint. JIT forecasting to manufacturing and freight lead times. Capacity without a new building. WMS selection; location-driven barcode ops. National transport held under scale.
The join where most programmes fail. Continuity under commercial pressure; systems that match the floor; executive targets that survive the lunch room.
Long-arc innovation and optimisation: warehousing, logistics, transport, IT, CX. Automation, relocation, procurement, KPI control, commercial warehouse agreements. Where drift hides when no one is looking.
Where AI enters
AI is only beginning. I already use it to build working systems inside real operations — not demos. I also read large architectures and codebases (order of ~10k-repo class analysis: breakpoints, structure, where soft middles hide), and I design safeguards so machines that act do not get a free pass.
The strategy is not “AI everywhere.” It is AI that brings security as much as speed, people development as much as process development, and that can carry operational baggage so people focus on the work with measurable impact on how the business and its operations arm deliver end-to-end — doing things right now so they stay better forever.
Same rule as everything else on this page: change for improvement, never for the sake of change.
Public artefact of how I think about fail-closed action: Terminal · Architecture. SECS is not mandatory on every engagement; process and policy often come first.
Intersect or standalone
AI alone
Capability vs need. Architecture reading. Admission rules, identity boundaries, override, audit before automation touches work or money.
IT alone
WMS/ERP truth, integrations, product, tenders, go-lives, the spreadsheet that is secretly a system of record.
Ops alone
Throughput, labour, DIFOT, cost-to-serve, multi-site, marketplace supply, expansion. Hold the number.
Where they meet
Default home. AI on real workflows. Systems that only pay if ops changes. Ops excellence that needs IT and governed AI.
Good fit / bad fit
Good fit
- Operator: someone owns the P&L or the service promise
- Startup: pilot / roll-out that must survive a real warehouse
- Willing to start with data and small enabling wins
- AI, IT, and ops misaligned — or one edge on fire
- Culture and the floor are allowed into the room
- Change for improvement, not for announcement
Bad fit
- Body-shop coding with no operational outcome
- Strategy theatre with no owner of the number
- Vendor who only wants a logo slide, not a live pilot
- AI for the press release only
- Unwilling to clean data or touch the broken process
- Soft middle required when machines act
Next to the rest of this site
Founder is the personal trail. The journal is the honest record. This page is how I work when you hire the AI–IT–Ops guy.