Operational systems we've designed and deployed in production. Each case study breaks down the system before engineering, the architecture after, and the measurable transformation between. Not testimonials. Engineering postmortems.
These are not success stories. Each study is a technical breakdown of an operation before engineering, the architecture that replaced it, and the measurable delta between them. Three rules hold for every entry:
Every study starts with the actual operational graph — fragmentation, manual handoffs, observability gaps — drawn in the same vocabulary used on an Audit.
The replacement system is shown as a five-layer model: trigger, decision, execution, data, observability. Reproducible, not bespoke.
Every claim is a number that came out of the production system — not a testimonial. Time, throughput, errors, coverage, cost.
A mid-market Swiss investment firm sourced deals from Flatfox, Homegate, and ImmoScout24 — via inbox alerts and spreadsheets. We replaced the copy-paste layer with an orchestrated pipeline: intake, portal-specific extraction, structured normalisation, contact enrichment, and a single sheet the team could trust.
A regional 3PL with 400+ daily shipments ran dispatch through a spreadsheet, a call rota, and the judgement of three senior planners. We replaced the judgement with an engineered decision layer.
A DTC health & beauty brand scaled from 50 to 500 orders a day using humans as the integration layer. We rebuilt the spine so the next 5× happened without adding headcount.
Three dimensions: which industry the system lives in, what kind of system it is, and what the engineered delta moved.
A mid-market Swiss investment firm sourced deals from Flatfox, Homegate, and ImmoScout24 — via inbox alerts and spreadsheets. We replaced the copy-paste layer with an orchestrated pipeline: intake, portal-specific extraction, structured normalisation, contact enrichment, and a single sheet the team could trust.
A regional 3PL with 400+ daily shipments ran dispatch through a spreadsheet, a call rota, and the judgement of three senior planners. We replaced the judgement with an engineered decision layer.
A DTC health & beauty brand scaled from 50 to 500 orders a day using humans as the integration layer. We rebuilt the spine so the next 5× happened without adding headcount.
A 40-person creative agency in London was losing 1.5 FTE of project-management time to coordination that wasn't coordination. We rebuilt the intake-to-invoice spine.
Every case on this page reflects the same principle. The shift isn't in the tools — the tools are almost always the ones the client already had. The shift is in how the work is modelled. Three ideas do most of the heavy lifting.
Every operation is a directed graph of nodes and edges. The moment you can draw it, you can reason about it. The moment you can query it, you can improve it. Most ops teams are running graphs they can’t see.
The slowest edge in the graph determines the throughput of the whole system. Adding staff anywhere else adds cost without adding capacity. Engineering means finding the constraint and moving it, not parallelising the easy work.
Gluing Zaps together is tooling. Designing an event spine, a decision layer, and an observability layer is architecture. The same business outcome on one of those approaches and not the other is the category gap.
Every case here started as an audit. If one of these studies described a system that looks like yours, the fastest next step is a formal operations diagnostic.
